text
stringlengths 8
5.77M
|
|---|
# Tests for PERFORMANCE_SCHEMA
#setup
let $select_column = COUNT_STAR;
let $table = performance_schema.events_statements_summary_by_program;
let $olddb = `select database()`;
create database explain_test_db;
use explain_test_db;
create table explain_test_table(c int);
delimiter //;
create procedure explain_test_proc()
BEGIN
select * from explain_test_table;
END //
delimiter ;//
truncate table performance_schema.events_statements_summary_by_program;
call explain_test_proc();
###########################################################################
# Test index on OBJECT_TYPE, OBJECT_SCHEMA, OBJECT_NAME
###########################################################################
let $column_count = 3;
let $col1 = OBJECT_TYPE;
let $col2 = OBJECT_SCHEMA;
let $col3 = OBJECT_NAME;
let $col1_act = "PROCEDURE";
let $col2_act = "explain_test_db";
let $col3_act = "explain_test_proc";
--source ../include/idx_explain_test.inc
# Cleanup
drop procedure explain_test_proc;
drop table explain_test_db.explain_test_table;
drop database explain_test_db;
|
Characteristics of macrophage activation by gamma interferon for tumor cytotoxicity in peritoneal macrophages and macrophage cell line J774.1.
We investigated the characteristics of macrophage-mediated tumor cytotoxicity (MTC) against Meth A target, H2O2 generation and release of effector molecule(s) for MTC, by comparing with those of peritoneal macrophages (PMP) and macrophage cell line J774.1 during stimulation with recombinant gamma interferon (IFN-gamma). In PMP, MTC was demonstrated when they were stimulated with IFN-gamma for 12 hr (short-term stimulation) and was abrogated when they were stimulated for 48 hr (long-term stimulation). Enhanced H2O2 generation was observed in PMP activated by long-term stimulation followed by triggering with PMA, but not observed by triggering with Meth A cells. By contrast, whereas non-treated J774.1 cells have already attained a definite level of MTC, a higher MTC level was demonstrated both by short- and long-term stimulations. Conversely, J774.1 cells were unable to generate H2O2 at any stage of IFN-gamma stimulation followed by triggering both with PMA or Meth A cells. The time course for stimulation of PMP by IFN-gamma for release of cytotoxic factor (CF) corresponded to that for MTC by PMP, and activities of the CF released from both activated PMP and J774.1 cells also closely corresponded to those of MTC by both cells. The serological and physicochemical characteristics of CF released from both activated PMP and J774.1 cells were determined to be closely related to those of tumor necrosis factor (TNF). These results indicate that in contrast to PMP, the J774.1 cell line is free from suppression stage for MTC and CF release during stimulation with IFN-gamma. The results suggest that TNF-like CF plays a crucial role for MTC against Meth A target, and that H2O2 is irrelevant for MTC against Meth A.
|
"'Today in Moscow, President Bush and President Putin met 'for their economic summit." "'$20 billion of aid to the Russian Federation 'has been granted by the International Monetary Fund." "'Our correspondent, Geoff Meade, is in Moscow." "'Geoff, $20 billion." "A huge amount of money?" "'" "'Kay, this is a massive vote of confidence in the Russian Federation." "'And it's very much down to pressure from Bush that this aid has been given.'" "Patience." "(Shouts in Russian)" "Ah!" "You are a clerk at the Ministry Financo." "You have a password to an account." "I can't access it unless you enter the password." "Please do so now." "What's the matter?" "It's only American money." "Yes." "Thank you." "(Gunshot)" "(All) Na zdorovye." "(Distant siren)" "Do we really have to involve them?" "Can't Ml6 do the job?" "They want to keep a safe distance and the last thing we want is the Fraud Squad." "Dear God, no!" "So we're stuck with these people." "Ah, Harry." "Come in." "Sir John." "This is Tom Quinn." "Sir John Barry, Governor of the Bank of England." " Pleased to meet you." " Tom." "This is Sir Richard Bowman, head of the bank, Bowman and Company." " Pleased to meet you." " How are you, Dicky?" "Oh, not quite in the pink, but well." " Dicky and I go back a very long way." " Right." "So, shall we?" "Actually, someone is joining us." "Here she is." "Running a little late." "Hope no one's inconvenienced." "Please, sit down, everyone." "This is Amanda Roke from the Chancellor of the Exchequer's office." "Ms Roke, this is Harry Pearce and Tom Quinn from Ml5." "Pleased to meet you, Ms Roke." "You're a Treasury civil servant?" "I'm the Chancellor's political counsel." "Ah." "Perhaps Sir Richard should..." "Last night, one of our employees stole $1 billion from us." "This is our thief" " John Lightwood." "A young man I would've trusted with my life." "The Chancellor is deeply concerned about this matter." "Bowman and Co is more than an old family bank." "It is a British institution." "A scandal like this could send it to the wall, trigger a stock exchange crash, even lead to a major economic depression." "We cannot allow that to happen." "That is why we suggested Ml5 be called in to find John Lightwood and recover the money he stole." "Surely this is a job for Scotland Yard?" "There's a sensitivity here." "Police would mean court cases" " and leaks to the press." " What sensitivity?" "Look, all you have to do is do what you're told." " I beg your pardon?" " It's all right, Tom." "The stock exchange may tremble but banks have failed before and Ml5 hasn't been called in." "What's so special about this case?" "Very well." "Bowman and Company handle certain government accounts." " Ah." " So you're the government's dirty bank?" "I don't think this is helpful." "This is very adventurous of you, Dicky." "Thank you, Harry." "As far as Ml5 is concerned, this operation is strictly on a need-to-know basis." "All you have to do is find John Lightwood and the money." "This is his file." "Right." "Well, you'd better get on with playing at policeman." " You will liaise with my office." " Absolutely." "(Siren)" "What'll they want next?" "Officers as traffic wardens?" "It stinks." "I agree." "I've known Dicky for years." "I've never seen him so worried." "We could put someone in the bank undercover." " Good idea." " Really?" "Why not?" "We are spies." "We're doing all we can to find John Lightwood." "Mobile phone usage, credit cards, the works." "Probably in Barbados now." "I had Special Branch turn over his flat and the bank have identified this man as Tim Prachett." "He's also a trader at Bowman's." "Seems like those two were best friends." "Do I get the impression you all hate doing this?" "Well, I'm with you." "Enough playing at PC Plods." "We're putting an officer into this bank, undercover." "No one pulls the wool over our eyes and no one will be told, certainly not Sir Richard." "So who's going in?" "Me?" "Undercover in a bank?" "You've a history of swindling credit cards." "I thought you'd be peffect." "Oh, yeah." "Thanks." "They want to replace Lightwood." "We'll make sure you get the job." "Malcolm will give you a legend and a great CV and Ruth'll teach you share trading." " Don't worry, it's all about cheating people." " Oh, right." "Zoe, I want you in there as well." "Go through records, files." "We want records of Lightwood transferring this money." "I want to know what they're not telling us about this bank." "We can see who can make the most money." "You'll make zilch." "You'll be a cleaner." " Buy half a yard at 75." " OK, done." "Zurich silver." "Selling half a yard at 82." " Too high, Danny." " Look at Singapore." "Cripes!" "OK, done." " Ruth, what just happened?" " Danny just made 50 grand or he would've if we'd been really trading." "How do you know all about this stuff?" "I was in love with a big swinging dick." "That's slang for a trader." "How's Danny measuring up?" " Oh, he's a huge talent." " I was afraid of that." "So what news from our cousins over the water?" "Washington CIA are hyperventilating about money." "$20 billion of aid has gone missing in Moscow." "American money." "20...?" "It's goodwill to Russia, to get them in step on the Middle East." "The money could come here." "If you hear anything at all, you'll tell us?" "You'd gain much love from the heart of America." "How could I refuse?" "Tim, this is Joshua lkoli." " Hi." " Hi." "Joshua will do John Lightwood's accounts." "He's very hot stuff." "Blazing references." " Have to watch ourselves, then, won't we?" " Indeed." "Tim will help you if you're not familiar with the systems." "All right, Tim?" "Yeah." "I look forward to your accounts at the end of the day then, Joshua." " The disgraced man's shoes." " I don't get you." "They say Johnny Lightwood stole money." "Success or theft?" "It's a thin line in this game." " You could say that." " Did you know John Lightwood well?" "Don't be crass." "None of us know each other that well round here." "Why the hell should we?" "Jazzer, Josh at Bowman's." "Yeah, great." "Socking it to the natives over here." "GCHQ says the German government is manipulating the cost of German steel." "'They want the price up five points.'" "No trade." " First-day nerves?" " Never piss on your own shoes." "True, true." "This is Joshua lkoli, new at Bowman's." "Yeah, hi." "Düsseldoff Steel - buy a hundred at 67." "Thanks." "Did you just squeeze the price on those options and sell them on?" "Yep." "Christ, you just made 250 grand, minimum." "Yep." "This is Joshua lkoli, new at Bowman's." "Ah, Zoe." "You have delivered me shredding." "Delicious." " (Zoe) 'How long will it take?" "'" " An artist cannot be hurried." " Oh, yes, he can, Malcolm." " (Beeps)" "(Police radio, indistinct)" " Tom." " Hi." "It's definitely your man." "He's in a right mess." "(Camera clicks)" "They crucified him." "Looks like." "I want Special Branch on to all the media." "No news, no press, absolutely none." "OK." "Someone's sending a warning." ""Mess with us, we'll crucify you."" "Look, I've got an order to buy at 52 and 100 million." "Can you improve on that?" "OK." "No sale, no trade." " I'm Maxi Baxter." " Josh lkoli." " Nigerian?" " My dad." "I was born in Peckham." " That'll be good." " Go for it." " Yeah!" "OK!" " Yes!" "Welcome to the animal room, Peckham boy." "Singapore." "Hi, Lei Fing, it's Maxi." "I want 50 Hang Seng at market price." "(Man) Deal!" "...work for you." "I really can." " 48, 49." " Come on, come on." " Yep." " But that's what I said to you this morning." "(Bell rings, low cheering)" "Bitch day!" "Bitch market!" "I wrenched my testicles just to keep even." "How about you?" "Up 1.5 mil." "(Applause, whooping)" "Animal room trader of the day." " A tradition." " And drinkies are on you." " Whoo!" " Oh, great!" " Great." "A greasy evening?" " Hm." "All right." "Little girls'." "She is a bad woman." "You want to find out how bad, sleep with her." "Johnny Lightwood did." "Noted." "What's the rest of the squeeze like around here?" "I'm off." "Girly club." " They lick you." " I'll have some of that." "Good man." "(Liberty X:" "Just A Little)" "Get hot just a little" "Meet me in the middle" "Let go just a little bit more" "Just a little bit" "Give me just a little bit more" "Just a little" "Oh baby won't you just" "Work it a little" "Get hot just a little" "Meet me in the middle" "Let go just a little bit more" "Just a little bit" "Give me just a little bit more" "Just a little bit more" "Work it a little" "Get hot just a little" "Let go just a little bit more" "Just a little bit" "Give me just a little bit more" "(Laughing)" "To pleasure." "Yeah." "I mean, abusing your mind and your body." "It's sort of an art form." "John Lightwood good at that, was he?" "Lightwood?" "Ooh!" " An artiste of the good times." " (Laughing)" "He couldn't say no to nothing." "He'd a big thing going." "What thing?" "Don't know if I ought to say." "Try me." "Johnny Lightwood was helping dirty money through the bank." " Tasty." " Very." "Money laundering these days." "If you've got the bottle, why not?" "You know?" "I mean, this is the mother of bear markets, right?" "But everything's in pieces." "A butteffly flutters in Tokyo, millions are wiped off the London, New York exchanges." "In times like these, if you can't make money the right side of the line, go over it." "And John was well over the line?" " All the way to Lugano." " Lugano, Switzerland?" "Is that where the money was going?" " To the Banco Co-operativo." " (Belches)" "Whose name was the account in?" "No idea." " You're not as pissed as me." " (Laughs)" "Come on!" " I wouldn't say that." " You love it." " Come on!" " Ooh!" "Sorry about the early morning debrief, Danny!" "Needs must." "Things are becoming murderous." "John Lightwood was tortured to death." "We must know why." "But, Danny, you and Zoe watch your backs." "Hear me?" " Sure." " Right." " Money laundering." "Do we understand it?" " You wash money clean." " See laundrettes." " Right." "It's not laundrettes, it's banks." "You move the money from one bank to another so fast, no one knows where it came from." " And Bowman's is?" " One bank on the way." "It can be there then not there in a second." "Don't think of cash, coins and notes." "It's..." "It's like a cloud." "It can be blown around the world, split into little clouds and can gather in a big storm in one place." "It's a very beautiful thing, money." "Do we have anything on this bank in Lugano?" "It's a Swiss bank." "God Almighty himself wouldn't get a statement and its computers are firewalled." " You can't even get a staff list." " Do they have a London office?" "Yes, they have a suite permanently booked at the Royal Paramount Hotel, Mayfair." " A hotel suite?" " To entertain clients discreetly." "Maxi Baxter." "I think she slept with Lightwood." "Can we get everything on her?" "Yes." "Ruth, you look into this woman." "Let's put an obbo on the hotel." "Are the security circuits in the rooms turned off?" "Don't worry, I checked." "Here's a copy of the Official Secrets Act." "Sign it for your own peace of mind." "My guests must not know." "They won't." "No one sees us, we go through walls." "Just pop through the wall, get us a cup of tea would you, Mal?" "Sorry." "Got a bit carried away there." "Right." "Let's bug this place and go." "Tom, it was Russian." " What?" " The way they crucified him." "Look." "Here's a 16th century Russian icon." "See?" "Arms straight, feet not crossed." " A Russian did this." " Russia." "It's right in the tradition of the Russian mafia." "This was your father's room, wasn't it?" "I should never have got involved with you." "I approached you." "Remember?" "You were horrified but then you said yes." "For a handsome commission." "This boy, Lightwood." " You had no idea what he was doing?" " I swear to you, no." "Greed must have overcome him." "But in the name of God, why leave him on Hampstead Heath like that?" "It was a warning... to you, you genitally-shrivelled, English pant-pisser." "That is so much better in Russian." "If only you spoke my language." " You're insane." " No." "Just free to do what I want." "The huge sums of money make anything possible." "Listen, English banker." "The transfer of $1 billion was a trial." "Now we have to take the risk of transferring the rest of the money out of Moscow." " $19 billion." " Oh, God!" "The CIA, the Russian Federation security services, are looking high and low for that money." "It must go in the next 48 hours." "Nothing will go wrong this time." "When the money comes into your bank, you will personally make sure it gets to Switzerland." "Understand?" "What is that?" "It's beautiful." "I had no idea your father owned an icon by Dionysius." "15th century." "It's incredibly rare." "Do you know what Dostoevsky said?" ""Beauty will redeem the world."" "Please, take it as a gift." "Why, thank you, Dicky." "I'll see myself out." "(Phone)" " (Beeps)" " Speak." "Shvitkoy wants to launder the rest of the money." " When does he want to do it?" " In 48 hours' time." " Well, terrific." " I don't see anything terrific about it." " This time, it'll work." " It's too dangerous." "Same as before." "Who'll handle the money this time?" "I don't know." "I can't think." " All right." "I will." " No." "You must be deniable in this." "I am very unhappy." "I want to talk to the Chancellor." "This is too sensitive for the Chancellor." " But if it goes wrong again..." " Sir Richard!" "We are all aware of the risk you are taking for the government." "I'm not taking it for the government." "I'm taking it for the country." "Yes, of course." "Just find someone to do it." "Sir Richard?" "So, what do we have?" "An amazing thing." "Richard Bowman opened an account for John Lightwood with a balance of £500,000, the day the $1 billion went missing." " A bonus?" " Nope." "They're paid at the end of the year." "This was done by Sir Richard with a copy sent to Lightwood and someone destroyed any record of it." "Or thought they had." "Until our Leonardo of the dustbins got hold of it." "Oh." "Thank you, Harry." "So what does this mean?" "John Lightwood was paid to steal from the bank?" "We need to know more." "Tell Danny to get friendly with the Maxi girl." "As friendly as it takes." "(Maxi) A carrot and coriander cocktail." "They say back in the '80s, young city bloods like us drank two bottles of champagne for lunch." "Heroic days." "But where are all the young bloods now?" "In alkie clinics." "Though some get away with it still." "Like John Lightwood." "Really?" "And what is it you think he got away with?" "The jungle drums say he nicked a billion dollars." "Now there's a true modern hero." "He was a friend of yours, wasn't he?" "Maybe." "What are you saying?" "Do you want to try what he did?" "Well, you know this male thing we have - recklessness." "I don't think that's just a male thing." "Tim said you were a bad woman." "Do you think he's right?" "He said there's a way I could find out." "Are you going to try it?" "Would I be allowed?" "That depends on the strength of your market penetration." "Of course." "(Harry) Does Downing Street want to know what's going on?" "It does not want you accusing Sir Richard of laundering Russian mafia money!" " Close the door." " You exceeded your remit." " Remit?" " You put a mole in the bank." "We must know what's been going on at Bowman's." "You're inteffering with matters that aren't your concern." " You are to withdraw your officer." " Are you giving us orders?" " Absolutely!" " And who are you to do that?" "Are you from the Joint Intelligence Committee?" "No." "Are you a member of government?" "No." "Did someone elect you?" "No." "Are you a real civil servant?" "No." "So what are you?" "Take this seriously, gentlemen." "Very seriously." "How come our gremlin in Downing Street knows Danny's undercover?" "If there's a leak here, we must plug it now." "Colin." "I'm gonna have to trust you." "I just have to take that leap in the dark." "What do you mean?" "Someone has been passing on the details of the bank operation." "I want you to monitor all the computers on the grid." "You want me to snitch?" "I don't want to do that, Tom." "This job asks difficult things of us sometimes, Colin." "(Typing)" " I suppose you have evidence?" " Computer traffic from your station, yes." " I coded it good." " Colin uncoded it good." "You betrayed us." "Ml5 is a government department." "I only told Downing Street what you were doing." "The government can't betray itself to itself." "Don't be naive, Ruth." "You know exactly what's going on here." "GCHQ planted you on us." "Tom, I so much wanted to join Ml5." "To be a real spy." "They said I could have the transfer if... very occasionally, I reported what you were doing." "This is the first time I've done it." "Reported directly to Amanda Roke?" "Well, now you're a classic double agent." "How does real spying feel?" "The horrible thing is, it's rather exciting." "Ruth, I've got two officers in the field at high risk." "Do you want to get Danny or Zoe crucified on Hampstead Heath?" "Don't." "I'll have to tell Harry." "I'm so bloody good at this job." "You know I am." "Well... that's that, then." "Of course, double agents can be turned." "Oh, Tom." "You're on probation." "Thank you." "No, it's not a thing I should have done." "As a banker, I am ashamed of myself." " Why?" "What have you done?" " I've booked a hotel room." "Well, er..." "You don't mind?" "No, I love it." " Is it...?" " Around the corner, Park Lane." "Oh, you are not thinking "greasy Italian man"?" "Oh, yes, I am." "Shall we...?" "It's the Royal Paramount." " Mm, posh." " Of course." "But if you don't like..." " Oh, no, no, no, no." "I like." " Ah." "Oh, no." " (Zoe) 'This feels so decadent.' - 'I'm glad you like it.'" "'Oh, and there is this.'" "No, no." "How dare you put an obbo team on me." " We rigged that room to see what we'd get." " I hope you all had a good gawp." "Oh, Zoe." "Who was on obbo duty?" "Colin." " You didn't sign in you were seeing him." " I'm not the first to do that." "He took you to rooms used by a Swiss bank, the same bank the laundered money was destined for." "Oh, please God, no." "You know what's going to have to happen next?" "What?" "Are you going to tell me I can carry on sleeping with him if I inform on him?" "I'm sorry, Zoe, I have to tell you." "He's married." " No, he's not." " Her name's Maria." "She's the daughter of the Italian finance minister, Antonio Gustino." "Your man married very advantageously." "For God's sake, Zoe, you didn't even look him up." "Mm?" "I, erm," "I just wanted a little bit of space, just for my own life, you know?" "I will get Colin to lose that section of the tape and I will shut him up but we need to know if this man is involved." "Yeah." "OK?" "Yeah." "Tom, something wondefful." "Death certificate." "Maxime Anne Baxter." "Died in Victoria Gardens, Farnham, Surrey, 31 March 1979, aged 3 months." "Birth certificate." "Maxime Anne Baxter." "Born Victoria Gardens, Farnham, Surrey, 31 December 1978." "This birth certificate used to get a minor's British passport in 1990." "My God." "What about, er, schools on her CV from the bank?" "Enrolled at Salehill Boarding School, aged 13." "Good place for the oddball girl." "Before that, she was meant to have gone to, Farnham Park, has no record of her." "And there are no records at all - doctors, anything - before 1990." "She suddenly appeared in London with a complete legend?" "It's a puzzle, isn't it?" " Ruth, this is great work." " Tom, I'm so glad you think so." "Come with me." " A wild guess." " Try me." "Daughter of a powefful man." "Wants her safe in England, has the clout to get her a new identity in 1990," " when the Soviet Union's collapsing." " Ex-KGB, turned businessman." "I'm beginning to see him in the shadows." "Aren't you?" "So?" "Can we set up a trap for this girl?" "All right." "But the budget won't run to the Docklands shag-pad of Danny's dreams." "His flat will have to do." "Get a team to erase Danny's identity and tart it up." "So, er, we walk round to the hotel?" "My lunch time's almost over, actually." "Better get back." "But lunch time was your idea." "Emma, have I offended you?" "I don't want to see you any more." "What?" "It's because of work." "Work?" "You're a banker and, er, we're doing some work with the bank." "So?" "So there's a conflict of interest." "Dear God, you're talking like some middle-aged bureaucrat." "I can't have my work compromised." "I'm very ambitious." "I love my work." "It's, er..." "It's just this thing I've got." "Hey, no, don't do this." "Mr Carlo Franceschini." " Who are you?" " We work for a government department." "Why should I believe you?" "You could be common muggers." "Your father, Alberto, was born in Milan." "Your mother Margherita, maiden name Pollini, is Italian Swiss." "What is this about?" "$1 billion was stolen from a British bank." "The money was placed in an account in the bank you work for." " We want the name of its user." " Hm." "How about the secret name of God as well?" "The men who stole this money are major criminals." "They kill." "Look, for a bank, money is morally neutral." "Mr Franceschini, think very carefully about the situation you're in here." "The situation is, I am sitting as a free man in a pub in a free country." "Do you wish your wife to receive copies in the post?" "Or we could e-mail them to her and to her father." "You realise, for the Swiss, a banker's confidence is as binding as that of a doctor or a priest?" "The account is held in the name of Victor Shvitkoy." "Why did you allow us to come into this without knowing that one of the most powefful psychopaths in the Russian mafia was involved?" "Did you know he has an agent working in your bank?" "Oh, no." "Not Maxime Baxter." "She's one of our best." "Birth certificates, passports - everything's false." "We believe she's Victor Shvitkoy's daughter." "Dear God." "(Tom) I have two officers in the field." "One of them is going to meet this woman in two hours' time." "So we must know, why is Bowman's Bank involved with this money?" "Governor." "Explain." "I have Downing Street's authorisation?" "When the $20 billion of aid went missing in Moscow," "Victor Shvitkoy approached Sir Richard." "Would he be interested in handling a large sum of money very quickly?" "When I reported Shvitkoy's approach to Ml6, they asked me to do what Shvitkoy wanted." "Sting him." " Yes." " Steal the money from him?" "I instructed John Lightwood to move the $1 billion into Shvitkoy's Swiss account." "That was a trial run." "Unfortunately, Mr Lightwood stole it." "And Shvitkoy found him." " Yes." " What about the other 19 billion?" "Shvitkoy is still going to try to launder it." "He believes I am his creature." "He thinks I can be used." "You realise governmental proximity cannot be acknowledged?" "The Treasury can't be seen near laundered money?" "Yes." "This will be yellow flagged "top secret" all the way, have no fear." "An incredibly brave and dangerous game, Dicky." "Nature intended me to sit behind a desk balancing books, not go double-crossing Russian gangsters." "All the more credit to you." "(Phone)" "Excuse me." "I must... concentrate." "The idea was... and is..." " to steal the money back." " For the Americans?" "(Russian man) 'Sir Richard." "'Dicky." "'I like calling you by your familiar name." "'We are close to doing a famous thing, a famous theft." "'You have no choice but to move the money." " 'Do you understand?" "'" " Yes." "I understand." "Is there something you're still not telling us?" " I can assure you..." " Governor!" " (Harry) Dicky." " (Moaning)" " (Governor) Richard." "...EC1." "(Struggling) Shvitkoy." "He's transferred the rest of the money" " from Moscow." " He's had a stroke." "Account number... 5, 9," "9, 8, 7," "4, 9, 5." "The money will only be at the bank for a few minutes." "Your man must divert it at once." "I know exactly what my man must do and at what risk." "(Phone)" "Yep?" " Right." " You look like a man getting a tip." "Something like that." "Give me the numbers." "5, 9, 9, 8, 7, 4, 9, 5." " Want me to repeat that?" " Don't bother." " What did you do just then?" " How do you mean?" "Were you transferring something in the main foreign accounts?" "No, just taking a look." "Taking five." " There's a problem." " (Russian man) 'Speak.'" " Has the money reached Lugano?" " 'I've not been advised.'" "God, no." "It's happened again." "'All right." "Keep calm." "Calm!" "'Listen to me." "Do you know who is responsible?" "'" "I think so." "Yes." "'Right, stay with him, do anything he wants to do." "'But get him to tell you where the money is." "(Fervently) 'Maxime, anything." "'Do what I say!" "'" " I have just nicked $19 billion." " Danny!" " What an amazing experience." " Danny, can you concentrate?" " What?" " Are you with me?" "Your lady friend is the daughter of a Russian hood called Victor Shvitkoy." "He killed John Lightwood." "What are Tom's orders?" "Our lovely flat now has the trappings of a luxury banker's pad." "Can't wait." "Whatever you do, stay with Maxi." "Do whatever you have to to find out where her father is and..." "And be careful, Danny." "Hope there's good booze in the fridge." "I'll do it tonight." " Ready?" " Dark side of the moon?" "Ooh, yes, please." " They're leaving." " Good, Zoe." "Go after them." " They're on their way." "Black London cab." " (Colin) I've got a signal." " Are we ready with the flat?" " Surveillance systems are operational." "Colin, when the word "yes" will do, use it, right?" " Yes." " Let's try and stay on top of this, shall we?" "Come on, Danny." "Thought you'd be more upscale than this." "It does me." "I'm never here." "Drinkies?" "Please." "Ah, shampoo." "Here we go." "What are we celebrating?" "I, er, don't know if I should tell you." "I've got some unbelievably hot info." "What?" "'You wouldn't believe it.'" " She's..." " We're OK." "We're OK." "Try me." "Good old Sir Richard Bowman's family bank is laundering dirty money." "Really?" "And, er, I've nicked it." " You've what?" " It was going from Moscow to Lugano through Bowman's, so I pushed it into a little account of mine." "Well, a very big account now." " How much?" " 19 billion smacks." " This is very risky." " He's trying to hook her." "'Lsn't that what John Lightwood did?" "'" "Lightwood?" "Disappeared, hasn't he?" "I don't know if you're a brilliant thief or a fool or what." "Brilliant thief and in a celebrating mood." "Where's the bedroom?" "(Maxi laughs) 'I do love a corny man.'" "I mean, I'm not actually stealing the money." "No." "I want to return it to its rightful owner." "For a little commission?" "Just two mil." "I mean, I'm not greedy." "No." "Little puritan, aren't you?" "Trouble is, I have to find the rightful owner." "Yes, you do, don't you?" "Can you help me there?" "What makes you ask that?" "Instinct." "What happened to that champagne?" "I'll be right back." " She's..." " The cleaners didn't do their job." "Hold." "Everyone, hold." " I'm going." " What?" "I've left my inhaler back home." "Sorry, I'll have to go." " Shall we lift her?" " Wait." " I'll come with you." " Will you, now?" " What do we do?" " We let them leave." " Danny will be at serious risk." " Yes." "Thank you." "(Distant car alarm)" "Done the place up, have you?" "It's a real bolt hole." "I got it for a song." "Great." "Don't stand in the cold." "Come in." "Ow!" "Bastards!" "So," "Mr Daniel Hunter." "My daughter tells me she thinks you are very beautiful." "I don't think so." "No?" "Well, if I put your eyes out, you wouldn't have to look at your face in the mirror." " Look, I'll do you a deal." " Excellent." "This is the deal I want." "Listen." "You have some money of mine in an account at a bank called Bowman's and Company." "Kindly tell me the codes, the number of the account, so I can enter the details into this computer and transfer the money." "Huh?" "Listen." "My colleague is holding a knitting needle." "And he will now put it straight through the cornea of your eye, into the optic nerve." "I found this in his clothes." " It's toothpick." " (Maxi) That's what I thought." "You see, Victor, tracker equipment's come on a lot since you were in the KGB." "(Slow, heavy breathing)" "(Officer talks, indistinct)" "(Sighs)" "The Chancellor of the Exchequer, too, has asked me to thank you." "It was a highly satisfactory outcome." "Many thanks." "I do hope your hearing's better." "Sorry, what?" "(Laughs) Sorry." "Well, I shall leave you to your internal debrief." "Certainly." "After the excitement, the paperwork." "Written statements, then Tom and I will do the personal debriefings." "Excuse me." " What about the money?" " What?" "The $19 billion." "All of, what, £13 billion?" "What happens to it?" "The right thing." "(Christine) Funny, you were talking about money." "(Tom) No, you did." "(Christine) Our Embassy economics analyst has gone insane." " That's nasty for him." " A very strong rumour is freaking him out." "Rumour being...?" "That the British Treasury have found a huge sum to improve the NHS - £13 billion." "There's a report from Moscow - blood-letting amongst the mafia." "$20 billion has disappeared in London." "World Trade Organisation money, American money." " Rumours, reports." " Yeah." "Take a brave thief to steal from the Russian mob." " What will the American government do?" " Give them another 20 billion." "I mean, we can't be seen to be losing money like that." "No, no." "You did a great service to the country." "You confronted a most dangerous man and did so with great courage." "We have recovered this." "We believe it belongs to you." "If we don't go to Egypt, I'll break this window." "Don't be ridiculous." " (Man) Major incident alarm!" " Treat this as real." " (Tom) What threat did they make?" " To bomb ten British cities." " VX." " Not the gas." " (Danny) What if it's not a drill?" " Is anyone feeling unwell?"
|
/*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
/**
* \file ConfigData.cpp
* \author Ratnadip Choudhury
* \copyright Copyright (c) 2011, Robert Bosch Engineering and Business Solutions. All rights reserved.
*/
#include "StdAfx.h"
#include "ProjectConfiguration/ProjectConfiguration_extern.h"
#include "ConfigData.h"
//#include "GettextBusmaster.h"
#include "Utility\MultiLanguageSupport.h"
#define DEFAULT_PROJECT_NAME "BUSMASTER"
#define BUSMASTER_UNIQUE_ID 133146
CConfigData CConfigData::m_ouConfigDetails;
CConfigData::CConfigData(void)
{
m_strCurrProjName = DEFAULT_PROJECT_NAME;
m_omStrCurrFileName = "";
}
CConfigData::~CConfigData(void)
{
}
CConfigData& CConfigData::ouGetConfigDetailsObject()
{
return m_ouConfigDetails;
}
void CConfigData::vReadConfigFile()
{
if (DoDatastorageOperation(OPEN) == 0)
{
PROJECTDATA sTempData;
GetProjectData(m_strCurrProjName, sTempData);
if (sTempData.m_dwAppUniqueId != BUSMASTER_UNIQUE_ID)
{
AfxMessageBox (_("The Configuration file is not Compatible"),
MB_OK | MB_ICONERROR);
DoDatastorageOperation(CLOSE);
}
}
else
{
AfxMessageBox (_("The Configuration file is not Valid"),
MB_OK | MB_ICONERROR);
DoDatastorageOperation(CLOSE);
}
}
void CConfigData::vSaveConfigFile()
{
DoDatastorageOperation(SAVE);
//bSetCheckSum(m_omStrCurrFileName , &pCheckSum);
}
void CConfigData::vCloseConfigFile()
{
DoDatastorageOperation(CLOSE);
m_omStrCurrFileName.Empty();
}
int CConfigData::SetConfigDatastorage(DATASTORAGEINFO* StorageInfo)
{
m_omStrCurrFileName = StorageInfo->FSInfo->m_FilePath;
return SetDatastorageConfig(StorageInfo);
}
int CConfigData::GetConfigDatastorage(DATASTORAGEINFO* StorageInfo)
{
return GetDatastorageConfig(StorageInfo);
}
void CConfigData::vSetCurrProjName(std::string strCurrProjName)
{
m_strCurrProjName = strCurrProjName;
}
std::string CConfigData::GetCurrProjName()
{
return m_strCurrProjName;
}
BOOL CConfigData::bSetCurrProjInfo(PROJECTDATA* ProjData)
{
ProjData->m_ProjectName = m_strCurrProjName;
return AddModifyProjectTable(m_strCurrProjName , ProjData);
}
BOOL CConfigData::bGetCurrProjInfo(PROJECTDATA* ProjData)
{
return GetProjectData(m_strCurrProjName, *ProjData);
}
BOOL CConfigData::bSetData(LPVOID lpVoid, int nStreamLength, std::string strSectionName)
{
BOOL bRetVal = TRUE;
// check input param
if(lpVoid != nullptr)
{
//if(m_bConfigInfoLoaded == TRUE)
//{
SECTIONDATA tempSecData;
tempSecData.m_omSectionName = strSectionName;
tempSecData.m_nBLOBLen = nStreamLength;
tempSecData.m_bBLOB = new BYTE[nStreamLength];
memcpy(tempSecData.m_bBLOB,lpVoid, nStreamLength);
AddModifySectionData(m_strCurrProjName,
tempSecData.m_omSectionName, &tempSecData);
//delete []tempSecData.m_bBLOB;
//}
}
return bRetVal;
}
BOOL CConfigData::bGetData(void*& lpData, int& nStreamLength, std::string strSectionName)
{
BOOL bRetVal = FALSE;
// is the configuration loaded?
// if((m_bConfigInfoLoaded == TRUE) && (bRetVal == TRUE))
{
SECTIONDATA tempSecData;
tempSecData.m_omSectionName = strSectionName;
bRetVal = GetSectionData(m_strCurrProjName,
tempSecData.m_omSectionName, tempSecData);
if (bRetVal != FALSE)
{
bRetVal = FALSE;
BYTE* pbNewDat = new BYTE[tempSecData.m_nBLOBLen];
if (pbNewDat != nullptr)
{
memcpy(pbNewDat, tempSecData.m_bBLOB, tempSecData.m_nBLOBLen);
nStreamLength = tempSecData.m_nBLOBLen;
lpData = (void*) pbNewDat;
bRetVal = TRUE;
}
else
{
}
}
}
return bRetVal;
}
|
Malaysians could see a possible extension of the movement control order (MCO) period, according to a report from The Star.
The number of infected cases within the country have still been rising, with the number hovering around 100 – 200 new reported cases daily.
Although a possible extension, a government source said that discussions to open up more economic activities to allow industries to operate are currently ongoing.
News company Bernama reported yesterday that the government could be announcing the status of the MCO to the nation “within the next 72 hours”, which is tomorrow.
|
Today's map shows where the United States' legal immigrants are most likely to hail from (besides Mexico). Because Mexican immigrants would dominate the map, they've been excluded to allow a more diverse picture. Creator Giorgio Cavaggion used data from the Department of Homeland Security (DHS) to make the map. The DHS defines a legal permanent immigrant on their website:
… foreign nationals who have been granted the right to reside permanently in the United States. LPRs are often referred to simply as “immigrants,” but they are also known as “permanent resident aliens” and “green card holders.”
The Afternoon Map is a semi-regular feature in which we post maps and infographics. In the afternoon. Semi-regularly. We spotted this one on Brilliant Maps: Making Sense of the World, One Map at a Time.
|
News
Penske Racing STP 400 Preview
April 18, 2013
Photo courtesy of Autostock
STP 400 at Kansas Speedway
The No. 2 Miller Lite Ford FusionDriven By Brad Keselowski• FORMER WINNER: Keselowski picked up his first NASCAR Sprint Cup Series (NSCS) win as a member of Penske Racing in the 2011 STP 400. In addition to his NSCS win, Keselowski also dominated the field in the NASCAR Nationwide Series event at Kansas in 2011, leading a total of 173 of the 200 laps run. • MR. KESELOWKSI GOES TO WASHINGTON: On Tuesday afternoon, Keselowski and members of Penske Racing were honored by President Barack Obama for winning the 2012 NSCS championship. The event took place on the south lawn of the White House.• TEAMMATE, TEAM OWNER: For the second-consecutive week, Keselowski will not only be a teammate to Joey Logano, he will be his team owner. Logano will pilot the No. 19 Ford F-Series in Saturday's NASCAR Camping World Truck Series race.
BRAD KESELOWSKI, NO. 2 MILLER LITE FORDWHAT WAS THE BEST PART ABOUT YOUR TRIP TO THE WHITE HOUSE? "Getting to spend some one-on-one time with the leader of the free world is something I never thought I would do in my life. To have the focus just on Penske Racing was very cool. My goal is to make him tired of seeing my face year after year." HOW MUCH DO YOU THINK ABOUT YOUR WIN AT KANSAS TWO YEARS AGO, YOUR FIRST CUP WIN AT PENSKE RACING? "As a racecar driver, it's really hard to try to reflect back on past successes as you are trying to continue to move forward and progress. I've always said that past success, or failure for that matter, doesn't guarantee anything for the future. Paul (Wolfe, crew chief) and I knew we could get it done at this level and that win at Kansas kind of validated that for us. For the reason it will always be a fond memory of mine."
The No. 12 SKF Ford FusionDriven By Sam Hornish Jr.• FIRST GO AROUND: Hornish has 129 NSCS starts with five of those coming at Kansas ,but this will be his first start in the Gen 6 Ford Fusion. • ADVANTAGE ERWIN: Greg Erwin likes Kansas Speedway and he has two NSCS victories at the venue (2007 & 2010).• ONE OF TWO: This weekend is the first of two NSCS events that Hornish and the No. 12 SKF Ford are scheduled to compete in this year.• LEADERS OFTEN WIN: As the leader of the Nationwide Series standings, Hornish relishes the opportunity to show just how strong the No. 12 Penske Ford team is this year.• PIT CREW SHUFFLE: Due to members of the No. 12 NNS Penske pit crew having other NSCS roles this weekend the pit crew of the No. 22 NNS Penske Ford will pit the No. 12 SKF Ford with the lone exception being Fueler Kellen Mills who is a team regular.• LOCAL TALENT: Mechanic Ryan Tschudi hails from nearby Creighton, Mo. and Tire Carrier Dylan Dowell is from Fowler, Kan.
SAM HORNISH JR. NO. 12 SKF FORDWHAT IS YOUR PRIMARY GOAL THIS WEEKEND? "Our initial focus is on qualifying. We have to get our No. 12 SKF Ford in the race. Once we have accomplished that we will go to work on the race set-up. After that we must execute during the race in order to contend and we plan on contending."WHAT IS UNIQUE ABOUT KANSAS? "From a racing standpoint, the resurfaced track makes it interesting. On a personal note, my daughter, Addison loves the Wizard of Oz and she thinks we should find her a Dorothy dress, so I have my work cut out for me on and off the track."
CREW CHIEF GREG ERWINDOES YOUR APPROACH CHANGE FOR THIS RACE?"Since we have to qualify for this one, we will place more emphasis on that than we do for a NNS race. Other than that it isn't much different. The pace of the weekend is a bit different and we have more tires, which is nice. I like this track and based on what I've seen from Sam and this team I think we may surprise a few people this weekend."
The No. 22 AAA Insurance Ford Fusion Driven ByJoey Logano• COMEBACK KID: Joey Logano might have a new nickname, that of the "Comeback Kid." Last weekend, Logano had to come from the back of the pack to earn a top-five finish in the Sprint Cup Series race at Texas Motor Speedway. Then on Sunday, racing in just his second NASCAR Camping World Series Truck race at Rockingham Speedway, Logano came back from a lap down to challenge for the lead on a final green-white-checkered finish, eventually finishing second. • TRAVELING MAN: This will not be a week for Logano to just kick back and get some rest. The 22-year-old will be on a whirlwind appearance schedule, spending time in New York City and Indianapolis making appearances for sponsors before heading back to North Carolina just in time to leave for Kansas.• DOUBLE DUTY AGAIN: Coming off of two top-five finishes last weekend, Logano will once again be a double-duty driver this weekend at Kansas, competing in the Truck Series again for BKR and the Cup Series event. • BACK HOME AGAIN: Shell-Pennzoil crew member Eric "Stretch" Baily will return home this weekend. Baily hails from Olathe, Kan., just miles away from Kansas Speedway. Baily, who is one of the mechanics on the No. 22, is an avid runner and hiker as well. JOEY LOGANO, NO. 22 AAA FORDHOW BIG WAS LAST WEEKEND'S TOP-FIVE FOR YOUR TEAM?"It was huge for our team. When you are thrown some adversity, you let it define you. You can either fold under the pressure or excel because of it. And that is what we did. We've had a lot of things thrown at us this year and every time, we've been able to come out of it on a positive note. That builds a team's moral and character and it's really brought our team together. That's a pretty awesome feeling. I think in the long run, it will make us even stronger and give us a lot of momentum heading into this long stretch of races."HOW'S IT FEEL TO BE NINTH IN POINTS? "It is nice, but you always want to be better and I feel like we could be even better than ninth. But we just have to work on maintaining that position. If we can go out there and have some consistent runs, grab some top-five and top-10s and a few wins along the way, the points will take care of themselves. The biggest thing is to make the most out of bad days, which is what we did in Texas. If we can keep that up, then things will work out just fine."
Team Penske is one of the most successful teams in the history of professional sports. Cars owned and prepared by Team Penske have produced more than 500 major race wins, over 570 pole positions and 33 Championships across open-wheel, stock car and sports car racing competition. Over the course of its 52-year history, the team has also earned 17 Indianapolis 500 victories, two Daytona 500 Championships, a Formula 1 win and overall victories in the 24 Hours of Daytona and the 12 Hours of Sebring. For 2018, Team Penske will compete in the Verizon IndyCar Series, the Monster Energy NASCAR Cup Series, the NASCAR XFINITY Series and the IMSA WeatherTech SportsCar Championship. The team also races in the Virgin Australia Supercars Championship, in a partnership with Dick Johnson Racing, as DJR Team Penske.
|
Q:
Creating an array from properties in an object in Drupal 7, php
I'm creating this form with checkboxes that I want to populate from a database. The problem is, how do I create an array of key=>value pairs when the data I need to create the pairs is in an object? I'm not sure I'm explaining myself properly, heres' the code and hopefully it will be clearer:
function myform_form($form, &$form_state) {
$options_query = db_query('SELECT name, mname FROM event_type');
$options = array();
foreach($options_query as $o) {
$options(($o->mname) => ($o->name)); //This is where I get the error unexpected T_DOUBLE_ARROW
}
$form['options'] = array(
'#type' => 'checkboxes',
'#title' => t('Search options'),
'#options' => $options,
'#description' => t('Choose what you want.'),
);
Is there a way to do this?
A:
Try changing:
$options(($o->mname) => ($o->name));
to
$options[$o->mname] = $o->name;
|
A word from Tim, the 'King Kong and the Empire State Building' costume creator:
Always one of my favorite movies! Building was made from three cardboard boxes taped together with packing tape. The boxes were then spray painted and holes were cut for my head and arms. Gorilla costume was found on eBay for $3.00! Thanks for looking!
|
The ticket correctly matched all six winning numbers, 09-20-25-26-46-49, to win a jackpot of $2,200,000, less 25 percent federal withholding.
The retailer will receive a $10,000 bonus for selling the jackpot-winning ticket.
The jackpot had been growing since one ticket matched all six winning numbers in the Nov. 26 drawing.
Lottery officials cannot confirm the identity of the jackpot winner until the prize is claimed and the ticket is validated. Match 6 winners have one year from the drawing date to claim prizes.
The Lottery encourages the holder of the winning ticket to sign the back of the ticket, call the Lottery at 717-702-8146 and file a claim at any of Lottery's seven area offices or at Lottery headquarters in Middletown, Dauphin County.
Claims may be filed at Lottery headquarters Monday through Friday, from 8 a.m. to 4 p.m., and from 8:30 a.m. to 4:30 p.m. at area offices.
In the 2011-12 fiscal year, in addition to awarding more than $63.4 million in prizes to York County winners, the Pennsylvania Lottery contributed nearly $27.2 million to programs serving the county's older residents.
How to Play Match 6 Lotto: Match 6 tickets cost $2 each. Players can choose six numbers from 1 to 49 or have numbers randomly selected by computer. The computer will then randomly select an additional two sets of six numbers, for three lines of six numbers each or 18 numbers total. Players can win up to four times on each ticket – once on each of three lines (base play) and by combining all 18 numbers (combined play).
Match 6 Lotto drawings are held on Monday and Thursday evenings.
About the Pennsylvania Lottery: The Pennsylvania Lottery remains the only state lottery that designates all its proceeds to programs that benefit older residents. Since its inception in 1971, the Pennsylvania Lottery has contributed nearly $22.6 billion to programs that include property tax and rent rebates; free transit and reduced-fare shared rides; the low-cost prescription drug programs PACE and PACENET; long-term living services; and the 52 Area Agencies on Aging, including full- and part-time senior centers throughout the state.
The Pennsylvania Lottery reminds its players to play responsibly. Players must be 18 or older.
|
Italian Minister of Labor and Industry Luigi Di Maio speaks at the Italian Business Association Confcommercio meeting in Rome, Italy, June 7, 2018. REUTERS/Tony Gentile
ROME (Reuters) - Italy’s Deputy Prime Minister Luigi Di Maio said on Thursday he still expects French President Emmanuel Macron to apologize for critical comments he made about Italian immigration policy two days ago.
Macron said Rome had acted with “cynicism and irresponsibility” by closing its ports to a migrant ship earlier this week, setting off a bitter diplomatic spat between the two countries, with Italy’s new Prime Minister Giuseppe Conte considering putting off a meeting with Macron due on Friday.
“We’re waiting for an apology. If we get one, we can start down a new path,” Di Maio said in a radio interview. “There’s still time to take a step back, apologize, and then start over.”
Macron called Conte on Thursday, a source in the prime minister’s office said, without elaborating.
|
import mkdirpTop from "mkdirp";
import path from "path";
import { promisify } from "util";
import store from "./store";
const mkdirp = promisify(mkdirpTop);
export default async function fsCreateProfile(profileName) {
await mkdirp(path.join(this.mediaDirectoryPath, profileName));
const promises = [
...store.getters["readHandlers/folders"].map(folder =>
mkdirp(path.join(this.mediaDirectoryPath, profileName, folder))
)
];
return Promise.all(promises);
// try {
// mkdirp.sync(path.join(this.mediaDirectoryPath, profileName));
// } catch (e) {
// throw new Error(`Could not make "${profileName}" profile directory ${e}`);
// }
// try {
// mkdirp.sync(path.join(this.mediaDirectoryPath, profileName, "image"));
// } catch (e) {
// throw new Error(
// `Could not make "${profileName}" profile image directory ${e}`
// );
// }
// try {
// mkdirp.sync(path.join(this.mediaDirectoryPath, profileName, "video"));
// } catch (e) {
// throw new Error(
// `Could not make "${profileName}" profile video directory ${e}`
// );
// }
// try {
// mkdirp.sync(path.join(this.mediaDirectoryPath, profileName, "palette"));
// } catch (e) {
// throw new Error(
// `Could not make "${profileName}" profile palette directory ${e}`
// );
// }
// try {
// mkdirp.sync(path.join(this.mediaDirectoryPath, profileName, "preset"));
// } catch (e) {
// throw new Error(
// `Could not make "${profileName}" profile preset directory ${e}`
// );
// }
// try {
// mkdirp.sync(path.join(this.mediaDirectoryPath, profileName, "module"));
// } catch (e) {
// throw new Error(
// `Could not make "${profileName}" profile module directory ${e}`
// );
// }
// try {
// mkdirp.sync(path.join(this.mediaDirectoryPath, profileName, "plugin"));
// } catch (e) {
// throw new Error(
// `Could not make "${profileName}" profile plugin directory ${e}`
// );
// }
// try {
// mkdirp.sync(path.join(this.mediaDirectoryPath, profileName, "plugin_data"));
// } catch (e) {
// throw new Error(
// `Could not make "${profileName}" profile plug_data directory ${e}`
// );
// }
}
|
18:37
Advocates have been pleading for weeks that San Francisco move people out of homeless shelters and into hotel rooms, given that the conditions inside these facilities are often unsanitary and crowded, making it easy for a virus to quickly spread.
Now, there has been a major coronavirus outbreak inside a homeless shelter in San Francisco: nearly 70 residents at MSC South have tested positive, which is roughly half of all the people who were tested.
San Francisco’s mayor, London Breed, said the city is now moving the rest of the MSC South residents who have not tested positive into hotel rooms and quarantining those with Covid-19 inside the shelter.
Jennifer Friedenbach, the director for the Coalition on Homelessness in San Francisco, said these kinds of outbreaks will happen at every shelter if the city doesn’t immediately move all residents to hotels. Some have suggested the coronavirus fatality rate for unhoused people could be significantly higher than the general population, with a 10-20% dying and 30% hospitalized, she noted.
“This was preventable and predictable. You can’t leave people in congregate settings, you just can’t,” said Friedenbach, adding that the city needs to test residents of all shelters given their potential exposure at this point. “They’ve waited so long.”
Those who remain in other shelters are now also facing strict lockdowns, where they are even more at risk of contracting the virus, said Leah Simon-Weisberg, an attorney with the Eviction Defense Collaborative.
Sam Levin (@SamTLevin) After 70 residents of a San Francisco homeless shelter test positive for Covid-19, advocates say others are trapped in lockdown in shelters with restrictions. "Telling them they cannot leave is not protecting them"-@Leahfsw. "People need to be moved to hotels, not just locked up" pic.twitter.com/YTfxVaMIGg
“Locking them in a congregate shelter and packing hundreds of people into one place is only endangering their lives,” she said.
|
All the highlights in the world of games, lovingly delivered 2-3 times/week.
You're Good to Go!
You are now subscribed
Subscribe for the latest gaming news
Last week, publisher Ubisoft revealed that its recent game releases have sold very well. So well, in fact, that the company's sales in the last three months have risen 374 percent over the same time period in 2013. But one of the big questions during Ubisoft's earnings call was how those sales were distributed over the platforms. On the day of the call, the company also published a breakdown of the software sales by system, and it turns out that games sold best on the PlayStation 4 by a wide margin.
Driven by a very strong performance by March's release of Watch Dogs, over 50 percent of Ubisoft's software revenue came from the Xbox One and PlayStation 4 during the past three months. But Ubisoft games on the PS4 sold much better than on Xbox One. Sales on the PS4 made up 36 percent of Ubisoft's revenue, more than double the One's 17 percent. It's interesting to note that PS4 revenue also more than doubled sales on PC, which came in at 14 percent of total sales.
You can check out the entire console breakdown over the last three months below. Software sales for the Wii U, Vita, and 3DS were negligible over this time period.
|
Honesty up front ensures hassle-free insurance claims
For cash-strapped consumers forced to tighten their belts in the current climate, it may be tempting to try and get cheaper long-term insurance cover by omitting vital information during the application process.
But Dr Maritha van der Walt, Convenor of the Medical Underwriting Standing Committee of the Association for Savings and Investment South Africa (ASISA), cautions that this short-sighted approach is likely to leave clients or their families without cover when a devastating life event occurs.
The risk of non-disclosure
Dr van der Walt explains that a life insurance company is fully within its rights to reject a claim and declare a client’s policy void if it finds the client was dishonest or didn’t disclose important details when they took out the policy.
Life insurers assess all claims objectively and fairly and a decision to reject a claim or to void a policy is never taken lightly, but clients can avoid problems at claim stage by fully disclosing all required information when taking out a policy.
When in doubt, rather disclose
‘If you’re in doubt about whether information is important or not, rather disclose it and let the underwriters decide whether it’s material,’ Dr van der Walt advises.
Depending on the information disclosed, a life insurer may request a medical assessment to determine the severity of a condition. For example, if a client mentions that they suffer from chest pain, the insurer will require an assessment to determine the client’s risk of suffering a heart attack. If the client discloses that they’re scheduled for any medical investigation within a couple of weeks, the company may delay issuing a policy until the results are known.
If the client’s medical condition or lifestyle choices are considered a material risk, the life insurer may offer the cover on different terms, such as a higher premium or exclusions on disability cover or critical illness benefits. In a small percentage of cases, an application is declined.
While it may therefore seem tempting to keep quiet about ailments or pending medical investigations, a claim for a condition that was subsequently diagnosed as a result of the investigations might be declined. ‘Rather be honest – if your cover is declined then you can at least plan ahead for the financial impact should something happen,’ Dr van der Walt says.
Even if the client doesn’t qualify for comprehensive life cover, for example, they can apply for accidental death cover, which only pays out in the event of death as a result of an accident.
A growing range of products are also available to people living with the human immunodeficiency virus (HIV), including life, disability and critical illness cover. South Africa was the first country to introduce life cover for people living with HIV in 2001.
|
Despite disagreement over prediction methods, some professors were successful in predicting last Tuesday's election outcome
Receive the latest national-international updates in your inbox
Allan Lichtman of American University in Washington, D.C. testifies during a Senate Redistricting Committee hearing on the proposed Illinois redistricting map, at the Illinois State Capitol in Springfield, Illinois, Tuesday, May 24, 2011. Lichtman is one of three professors who correctly predicted the outcome of the 2016 presidential election.
Donald Trump is not the only one who can say, "I told you so."
A few college professors stood up to a tidal wave of prevailing thought to separately predict for months that the New York billionaire and reality TV star would pull off one of the biggest shockers in American political history.
A political scientist at New York's Stony Brook University based his prediction on a formula using primary results and his "pendulum of change" theory. A Yale professor tied his pick to economic factors. And a history professor at American University relied on a formula he developed in the 1980s.
"I had a lot of people saying, 'This isn't going to work. You're going to fall on your face.' I got emails berating me for being an idiot and irresponsible," says Helmut Norpoth, a longtime political science professor at Stony Brook.
Alabama Candidates Cast Their Ballots
Republican Roy Moore, facing numerous allegations of sexual misconduct with teenage girls, and Democrat Doug Jones cast their ballots in the vote that will send one of them to the U.S. Senate. NBC's Chris Pollone reports.
(Published Tuesday, Dec. 12, 2017)
Norpoth has successfully predicted every presidential winner since developing the formula for the 1996 presidential race. He also used his formula to review every presidential election since 1912, and found the indicators would have accurately predicted the outcome every time except 1960.
He acknowledged he harbored some private doubts as this year's contentious campaign raged on, he never wavered publicly, even doubling down on the pick after Trump's explosive "Access Hollywood" video came out.
"I put up a good face," he says. "But deep down I got a little uneasy, I can't deny that."
Norpoth says his model is based, in part, on the idea that candidates who excel in primaries tend to do better in the general election. He says Trump's victory and Clinton's defeat at the hands of Sen. Bernie Sanders in the New Hampshire primary was a key indicator for him of the Republican's strength.
He also factored in that Barack Obama had held the White House for two terms and estimated the pendulum could swing Republican. The only time in the past 30 years that either party won three straight presidential elections was in 1988, when Republican George H.W. Bush defeated Democrat Michael Dukakis.
Also successfully predicting a Republican victory was Yale University professor Ray Fair. He used economic factors — GDP growth and inflation — and not the vagaries of either candidate's personality in his prognostication.
American University professor Alan Lichtman said he uses 13 factors to assess the performance of the party in power to determine whether it will keep the White House. He examined presidential elections from 1860 to 1980 to establish the factors to develop his study. He said the factors this year narrowly showed a Republican would prevail.
Lichtman's success comes with an asterisk. He has accurately calculated the winner of the popular vote in every election since 1984, but Al Gore actually lost in the electoral college in 2000, as it appears has happened to Clinton this year.
"I feel vindicated in a sense, but I had issued a qualification because in Donald Trump you had a history-smashing candidate," Lichtman says.
|
---
abstract: 'Machine Learning has proven useful in the recent years as a way to achieve failure prediction for industrial systems. However, the high computational resources necessary to run learning algorithms are an obstacle to its widespread application. The sub-field of Distributed Learning offers a solution to this problem by enabling the use of remote resources but at the expense of introducing communication costs in the application that are not always acceptable. In this paper, we propose a distributed learning approach able to optimize the use of computational and communication resources to achieve excellent learning model performances through a centralized architecture. To achieve this, we present a new centralized distributed learning algorithm that relies on the learning paradigms of Active Learning and Federated Learning to offer a communication-efficient method that offers guarantees of model precision on both the clients and the central server. We evaluate this method on a public benchmark and show that its performances in terms of precision are very close to state-of-the-art performance level of non-distributed learning despite additional constraints.'
author:
-
bibliography:
- 'biblio.bib'
title: 'Combining Federated and Active Learning for Communication-efficient Distributed Failure Prediction in Aeronautics'
---
Distributed Learning, Failure Prediction, Active Learning, Federated Learning, Machine Learning
Introduction
============
General problem
---------------
In the recent years, the efficiency of Machine Learning for automated processing of large volumes of data has been widely demonstrated. This has been of particular interest for industrial applications that commonly generate large datasets. One such application is failure prediction as the ability to anticipate on a system failure can lead to improvement in its reliability and to more optimized maintenance processes. However, a known limitation for the use of Machine Learning is its high computational costs relative to more traditional threshold-based system monitoring. In some application contexts, a simple solution to this can be found by either deploying additional resources at the location where they are needed or by offloading the data to a remote location following the cloud computing paradigm. There are however situations where neither of those options are realistic.
If we consider the application case of vehicular systems in general and, more specifically, aeronautical systems, they would greatly benefit from increased reliability and optimized maintenance processes and the latest generations of aircraft do have both on-board computational capabilities and data links with the ground. However, aeronautical systems generate too much data to be able to handle failure prediction on the aircraft for lack of computational resources and deploying additional on-board resources is very complex and costly. Moreover, the data links available between the aircraft and the ground are too expensive and not fast enough to transfer all the data. Therefore, we cannot directly apply a cloud computing paradigm which would consist in offloading all the data to a remote location such as a data centre with sufficient computational resources and collecting the results.
The solution for this would be to distribute the workload between the aircraft and a ground station in a client/server architecture in order to use their different capabilities in the most efficient way. In this situation, we can identify requirements that the solution will need to demonstrate to function in a distributed manner. Firstly, it will need to be able to classify data from normal readings, i.e. situations where there are no signs of imminent failures, using a minimal amount of computation power. We will refer to this requirement as $R_1$. This is key to be able to process the large volume of data generated by the aeronautical systems without overloading the data link. Secondly, the solution for failure prediction will need a mechanism to identify interesting data, i.e. data that could be indicative of an imminent failure with regards to a pre-defined threshold, and transfer this data from the aircraft to a central server on the ground without spending too much communication budget. We will refer to it as $R_2$. This is a direct consequence of the observation that the amount of computational power on board is too limited to give an accurate prediction of failure risks. Thirdly, we need a way to train a prediction model on the ground without accessing the data from the aircraft. We will refer to it as $R_3$. This is a direct consequence of the previous two requirements as only a small amount data can be sent to the central server. But, as accurate predictions are expected from it, we need a mechanism to enable centralized learning without accessing the raw data.
Different solutions already exist for each of the requirements we have identified taken individually. The main concern here is to figure out how to fulfil all three requirements simultaneously.
Distributed Learning paradigms
------------------------------
The field of Machine Learning dealing with distributed algorithms is Distributed Learning. There are two sub-fields in particular that provide interesting elements of answer for our application case, Active Learning [@cohn1996active] and Federated Learning [@mcmahan2016communication]. They have been designed with slightly different use cases in mind which obviously leads to slightly different requirements but that nevertheless matches our use case quite well.
The Active Learning paradigm offers a way to model the envisioned relationship between the aircraft and the central server on the ground. Its premise is that the model being trained has access to a third-party source of information, called an oracle, to reveal labels when necessary and its goal is to maximize the performance of the trained model under a request budget. We can immediately draw a parallel with the requirement $R_2$ of the solution we identified, identify uncertain samples and transmit them under a communication budget. With the aircraft playing the role of the active learning model in training and the central server playing the role of the oracle albeit an imperfect one (an unusual situation but already studied, for example in [@miller2014adversarial]), the active learning paradigm offers an approach to select the hard-to-classify data and balance the amount of requests with the communication budget. Regarding requirement $R_1$, enabling computation-efficient classification of normal samples, however, further investigations are required as a number of works in Active Learning favour a greedy approach incompatible with the idea that normal readings need to be processed quickly.
The Federated Learning paradigm is directly related to requirement $R_3$, i.e. enabling centralized learning without accessing raw data. The communication budget and the computational resources of the hosts are also concerns that are considered in this approach. However, it is purely driven by the central server and does not provide any mechanism for the clients to assess their performance and request clarifications from the central server.
The goal of this contribution is to propose a new approach that fulfils all three requirements by combining Active Learning and Federated Learning. The performances are evaluated on a public benchmark and compared to existing solutions.
The contribution is organized as follows. First, we study existing works in Active Learning and Federated Learning in section \[related\]. Next, we detail our solution in section \[algorithm\]. We present our results and interpret them in section \[results\] and section \[conclusion\] concludes the contribution.
Related Work {#related}
============
In this section, we review state-of-the-art works related to our problem focussing on the applicability of the techniques reviewed with regards to the three identified requirements.
On the matter of distributed learning publications that are of particular interest to us we have to mention [@valerio2017communication]. In this article, an Internet of Things (IoT) distributed learning framework for classification is presented and evaluated on network overhead and model accuracy. The authors present a solution called Hypothesis Transfer Learning based on an implementation of Greedy Transfer Learning found in [@10.1007/978-3-319-23231-7_1]. In this decentralized approach, every host computes a local model based on the data accessible to it, sends its model to all its neighbours in a synchronization phase and collects their models. A second round of training is done, using Greedy Transfer Learning from neighbour models to improve the local model. The synchronization phase is then repeated until the host receives the updated local model of every host that has contributed to the distributed learning. Finally, the models are aggregated in an ensemble model using majority voting, that is to say the result of the classification is the most frequent prediction among the aggregated models. This work considers several hypothesis that are relevant for aeronautical systems; namely, it avoids data transfer between hosts, provides an explicit way to trade-off network overhead and models accuracy by tuning how many nodes take part to the Greedy Transfer Learning training and it considers the case of class imbalance (though with a much less drastic ratio). The most significant difference with the prediction in aeronautical systems, beside the order of magnitude change in class imbalance ratio, is that it is a decentralized algorithm with communication between hosts, which is not a realistic hypothesis for an aircraft. Another problem that it does not address is the fact that the performance of individual models is not controlled meaning that it is not lower bounded and that there are no mechanisms to correct a potential drift of a local model. It makes sense in the situation considered in [@valerio2017communication] but it is a requirement that needs to be addressed in the aeronautical use case.
Note that for aeronautics applications, we would like to use Random Forest (RF) because of its resilience to noise and class imbalance as seen in [@AusselJGPFC17] and that it can be proved reliable in a range of safety scenarios. An issue remains in that RF is easy to distribute but there is no guarantee on the performance of individual trees.
Federated Learning
------------------
The terminology of Federated Learning was coined recently in [@mcmahan2016communication]. It describes a Distributed Learning situation where the following observations apply:
- There is a massive number of clients with limited computational resources each holding a unique fraction of the dataset. Their availability is subject to changes. This matches well the aeronautical use case where individual aircraft can be seen as clients. Their availability is not perfect and depends on the satellite connection.
- The data are not independent and identically distributed between clients. In the original use case because it is assumed that different users will interact differently with their mobile device and type differently. In the aeronautical use case, this holds true as different aircraft can face different flight conditions depending on their route and the way their systems are used by the crew and the passengers.
- Communications are allowed between the clients and a central server but they are limited by a communication budget and, for privacy concerns, no exchange of raw data is allowed. In the aeronautical use case, the privacy requirement between the client and the server is relaxed but the communication constraint is certainly relevant.
The original solution proposed is to initialize randomly a deep learning model at the server level and to gather incremental Stochastic Gradient Descent (SGD) updates from the clients while exchanging only the model parameters with them.
Since the publication of the original article, several contributions have followed and improved different aspects of this method. More specifically, [@konevcny2016federated] proposes a way to improve the communication efficiency and [@DBLP:journals/corr/BonawitzIKMMPRS16] and [@geyer2017differentially] offer new insights as how to better guarantee that the privacy concerns are respected.
The Federated Learning approach is of high interest for us as it describes a framework for massively distributed and communication-efficient learning in a context very close to ours. In particular, Federated Learning would be a great answer to our requirement $R_3$, learning in a centralized model without downloading data from the clients. There are however limitations. We have so far identified RF as our most promising choice for a base learning model. Also, Federated Learning is developed for Deep Neural Networks (DNNs) and uses incremental SGD updates from the clients. However, RF does not use SGD so we cannot apply Federated Learning directly. Also, Federated Learning does not offer any solution for our second identified requirement $R_2$, identify and uncertain samples and transmit them under a communication budget. For that, we turn to Active Learning.
Active Learning
---------------
Active Learning is a sub-field of semi-supervised learning where a third-party, called an oracle, can provide missing labels on request. It is traditionally applied in situations where data labelling requires the intervention of a human operator that is both costly and slow. The goal is then to learn the most accurate model possible under a given request budget. The general principle of Active Learning is interesting for our problem because it allows us to dynamically balance uncertainty on client-side and communication budget which would offer us a solution to fulfil the requirement $R_2$, identify uncertain samples and transmit them under a communication budget.
A summary of the principles of Active Learning from a statistical perspective can be found in [@cohn1996active].The most common approach for concrete applications is called pool-based active learning and described in [@10.1007/978-1-4471-2099-5_1]. The assumption in this approach is that there is a large pool of unlabelled data and a smaller pool of labelled data available. The labelled data is used to bootstrap a tentative model and a metrics to measure the amount of information gain one can expect from a sample is defined. Some examples of metrics can be a distance metrics between the sample and the decision boundary of the model, the density of labelled data in the vicinity of the sample or the difference in the learned model with regards to the possible labels. Once the metrics is defined, the unlabelled samples that are expected to maximize the information gain are queried from the oracle then the tentative model is updated and the process is repeated until the expected information gain falls below a pre-defined threshold, meaning that no new label is expected to significantly change the model, or the query budget is spent.
However, pool-based approaches are not a very good fit for our aeronautical application case as they often use a greedy sampling approach meant to minimize the intervention of the oracle, often a human operator, with the limitation of being prohibitively expensive in terms of computation; a human operator is indeed still orders of magnitude slower and more expensive than a complex computation. Examples of this can be found in [@Georgala:2014:SFA:2611040.2611059], [@5206627] and [@LI2007459] illustrating the versatility of pool-based Active Learning with three applications on very different use cases of respectively spam detection, image classification and network intrusion detection. In our case, however, we envision the oracle as another more powerful model faster and less expensive than a human operator and we assume that complex computations are not possible on the client-side.
There are other works on online active learning that adopt a stream-based approach. In that approach, the decision to request a label from the oracle must be done immediately and leads to more efficient decision from a computation point of view. Examples of this approach can be found in [@bouguelia2013stream] for an application to document classification and in [@smailovic2014stream] for an application to sentiment analysis. [@de2017confidence] is an article of particular interest to this article as it proposes a new way to train decision trees using an active learning approach that focuses on minimizing the risk of selecting a sub-optimal split when a new leaf is added to the tree by computing confidence intervals. To put it differently, when a new leaf is added to the tree, the comparison is made between the ideal decision tree that could have been made if the true distribution of the samples were known and the expected performance of the current tree. If the difference is then found to exceed a certain threshold, labels are requested to minimize the risk of selecting a sub-optimal split for the new leaf. The point that is the most relevant for this article is that, by moving away from a greedy approach, we can ensure that the computation cost remains manageable, making this algorithm suitable for an on-board implementation. Moreover, the notion of approaching the uncertainty in terms of bounded risks for the model to be sub-optimal instead of considering sample misclassification makes it possible to directly manage the trade-off between the communication budget and the quality of the model. The only remaining limitation is that this model is not distributed. Therefore, we still need to implement a mechanism to leverage remote data sources to create a centralized model in order to fulfil requirement $R_3$, enable centralized learning without accessing client raw data.
Proposed Algorithm {#algorithm}
==================
The approach we propose to combine Active Learning and Federated Learning is to use the confidence based active online Decision Tree (DT) from [@de2017confidence] as a base model that is learned on the client side and sent to the server. The server aggregates the client DTs in an ensemble model similar to a RF to learn an accurate model without accessing the raw data. Finally, the individual DTs use the ensemble model from the server as an oracle for label requests. The figure \[fig:illustration\_federated\_active\] provides a high-level illustration of the approach envisioned.
To get more into details, each client has access to its local dataset and trains a confidence based active online DT locally. Regularly, the central server sends an update request to the clients. The clients reply with their current DT. The server creates an ensemble model by collecting all the DTs. The ensemble model relies on majority voting to classify new samples, that is to say a new sample is run through each DT and its classification is the most frequent result among all DTs. A first parameter to control the communication budget of the approach will therefore be how often the central server requests updates from the clients.
From the clients perspective, they have at their disposal a stream of observations to train their local DT. As described in [@de2017confidence], whenever the uncertainty surrounding a new split in the DT exceeds a certain confidence threshold, a request is made to reveal the true label of the samples needed to decide on the right split and a part of the request budget is spent. Here the requests are sent to the server and the labels are decided using the ensemble model. There is a second parameter to control the communication budget of the approach in the request budget of each client.
There is a last point that needs to be clarified in the initialization of the model. Indeed, the necessary labels for the training of the DT are provided by the ensemble model. However, the ensemble model can only provide labels after it has received base models from the clients. In order to get around this constraint, we assume that some amount of labelled historical data is already available and can be used in the first round of update to initialize the DTs. This assumption is perfectly justified in the aeronautical use case where there is plenty of historical data available and the problem of label availability only concerns recent data where the true status of a system has not been clarified through a maintenance operation yet.
The proposed process is described in algorithm \[Federated\_active\_alg\].
$F_{0} \gets \Call{SeedForest}{C}$ $T_{k,t} \gets \Call{ClientUpdate}{(T_{k,t-1})}$ $F_{t} \gets \bigcup\limits_{\forall k}^{}(T_{k,t})$
$T_{k,0} \gets \Call{SeedTree}$ $F_{0} \gets \bigcup\limits_{\forall k}^{}(T_{k,0})$
$B_{updated} \gets B_{k,t-1} + \Call{UpdateBudget}{sizeof(n_{k,t})}$ $T_{k,t}, B_{k_t} \gets \Call{ConfidenceDecisionTree}{T_{k,t-1}, n_{k,t} B_{updated}}$ ($T_{k,t}$)
$T_{k,0}, B_{k,0} \gets \Call{ConfidenceDecisionTree}{Root\_Tree, n_{k,0}, B_{max}}$ ($T_{k,0}$)
\[Federated\_active\_alg\]
Experimental Results {#results}
====================
Experimental settings
---------------------
In order to enable a meaningful comparison with other works, we have selected for our experiments the MOA airlines dataset[^1], one of the public benchmarks that was used in [@de2017confidence]. Though it is related to the air transport industry, it is not strictly speaking a dataset from an aeronautical system but it presents the advantages of being public and enabling comparisons with other solutions. This dataset consists of over $500,000$ samples each with a label and seven features. Each sample corresponds to a real flight and the task is to predict if the flight has taken off on schedule or if it was delayed. The seven features available are summarized in Table \[tab\_MOA\_features\]. The imbalance ratio in this dataset is quite even at $0.80$ in favour of the flights that are not delayed. This means that some of the techniques such as Support Vector Machine (SVM) or Logistic Regression (LR) could have acceptable performances on this particular dataset. However, since the target application for aeronautical systems would display a much higher level of class imbalance due to their reliability and since our proposed method is DT based, we do not consider them further.
Feature Description Format
---------------- ------------------------------------------------------- --------------
Airline Unique identifier of the airline operating the flight alphanumeric
Flight numeric Unique flight ID numeric
AirportFrom Origin airport unique IATA code string
AirportTo Destination airport unique IATA code string
DayOfWeek Day of the week numeric
Time Timestamp (scale not disclosed) numeric
Length Length of the flight (min) numeric
To evaluate the performance of our algorithm, we report the precision, recall and F-score as in the previous contributions. Given the relatively balanced nature of this dataset and to provide as many elements of comparison as possible with other works done on this benchmark, we also report the accuracy. Defining the delayed flights as positives and the flights on schedule as negatives, as the labels in the dataset suggest us to do, and writing true positives, false positives, true negatives and false negatives respectively TP, FP, TN and FN, we define: $$accuracy = \frac{TP + TN}{TP + TN + FP + FN}$$ $$precision = \frac{TP}{TP + FP}$$ $$recall = \frac{TP}{TP + FN}$$ $$F-score = \frac{2 \times precision \times recall}{precision + recall}$$
In order to ensure the validity of our results, we apply a $2$-fold cross-validation by randomly splitting out the dataset into two sets, one for training and one for testing, measuring the performances a first time and repeating every measurement a second time after inverting the training and testing sets. Regarding the parameters of this experiment, we examine more specifically the performance of our method with regards to the number of hosts among which the samples will be split, the maximum request budget available for the active learning process and the number of communication rounds for the federated learning process. When studying the variations of a parameter, others were kept at constant with $20$ communication rounds, $5$ clients and a request budget of $10\%$. The numbers of communication rounds and clients were chosen so that the constraints of distributed learning would be visible and the request budget was chosen so that comparisons with other methods are facilitated.
Results and discussion
----------------------
The results are grouped by variable parameter and displayed in Figure \[federated\_active\_hosts\] for the number of hosts, in Figure \[federated\_active\_comm\] for the number of communication rounds and in Figure \[federated\_active\_req\] for the request budget. As additional information, Figure \[federated\_active\_req\_per\_host\_acc\] provides insights on the average performances of individual clients for varying number request budgets.
[ ![Performance with regards to number of clients[]{data-label="federated_active_hosts"}](federated_active_hosts_acc.png "fig:") \[federated\_active\_hosts\_acc\] ]{}
[ ![Performance with regards to number of clients[]{data-label="federated_active_hosts"}](federated_active_hosts_pr.png "fig:") \[federated\_active\_hosts\_pr\] ]{}
[ ![Performance with regards to number of clients[]{data-label="federated_active_hosts"}](federated_active_hosts_r.png "fig:") \[federated\_active\_hosts\_r\] ]{}
[ ![Performance with regards to number of clients[]{data-label="federated_active_hosts"}](federated_active_hosts_fs.png "fig:") \[federated\_active\_hosts\_fs\] ]{}
[ ![Performance with regards to the number of communication rounds[]{data-label="federated_active_comm"}](federated_active_comm_acc.png "fig:") \[federated\_active\_comm\_acc\] ]{}
[ ![Performance with regards to the number of communication rounds[]{data-label="federated_active_comm"}](federated_active_comm_pr.png "fig:") \[federated\_active\_comm\_pr\] ]{}
[ ![Performance with regards to the number of communication rounds[]{data-label="federated_active_comm"}](federated_active_comm_r.png "fig:") \[federated\_active\_comm\_r\] ]{}
[ ![Performance with regards to the number of communication rounds[]{data-label="federated_active_comm"}](federated_active_comm_fs.png "fig:") \[federated\_active\_comm\_fs\] ]{}
[ ![Performance with regards to the request budget[]{data-label="federated_active_req"}](federated_active_req_acc.png "fig:") \[federated\_active\_req\_acc\] ]{}
[ ![Performance with regards to the request budget[]{data-label="federated_active_req"}](federated_active_req_pr.png "fig:") \[federated\_active\_req\_pr\] ]{}
[ ![Performance with regards to the request budget[]{data-label="federated_active_req"}](federated_active_req_r.png "fig:") \[federated\_active\_req\_r\] ]{}
[ ![Performance with regards to the request budget[]{data-label="federated_active_req"}](federated_active_req_fs.png "fig:") \[federated\_active\_req\_fs\] ]{}
![Average recall per client with regards to the budget request[]{data-label="federated_active_req_per_host_acc"}](federated_active_req_per_host_r.png)
In Figure \[federated\_active\_hosts\], the first observation that we can make is that the results seem slightly unstable, especially considering that cross-validation has already been applied to them so one could expect smoother curves. In fact, the relative instability is a side effect of the majority voting employed in the ensemble model. When the number of clients is even, there is a limit case where each classification outcome gets the same number of votes. In such a case, our model was configured to favour the delay prediction, resulting in an increase in recall and a decrease in precision. This limit case is more frequent for small numbers of clients. Beside this, we can observe that the accuracy is stable at about $61\%$ which is an excellent level of performance comparable to state-of-the-art results in the non-distributed case with [@de2017confidence] reporting about $62\%$ of accuracy with a similar request budget. This validates the potential of the approach we propose for communication-efficient distributed learning. For the other metrics, we can observe that they decrease slightly for a high number of clients which illustrates the fact that the distributed learning constraint does have an impact on the performance even if it is relatively small with the F-score decreasing from about $49.75\%$ with a single host to $41.3\%$ with $10$ hosts.
Concerning Figure \[federated\_active\_comm\], the results are more straightforward. We observe that as the number of communication rounds increase, the model accuracy is quite stable but it hides the fact that precision is increasing at the expense of recall leading to an overall decreasing F-score from $51\%$ to $30\%$. The interpretation to this behaviour is that the ensemble model is being updated too frequently. The base DT model makes use of a grace period parameter which defines a minimum amount of samples that need to be collected before a request can be made. The default value used was $100$, the same as in the original publication. However, as the dataset is split between multiple clients and further down between multiple communication rounds, it is possible when the number of communication rounds is too high that between consecutive updates of the ensemble model, no updates of the local DTs have been made and consequently no change is made on the ensemble model leading to “skipped” communication rounds that eventually lead to lower the overall performances. We tested this interpretation by lowering the grace period to $30$ and observed that performances for a high number of communication rounds were increased.
Regarding Figure \[federated\_active\_req\], it is worth noting that it is using a different scale than the other figures as the request budget is presented on a logarithmic scale. The reason for that is to highlight the extreme stability of the performances with a very slight increase of the accuracy for high request budget from $56\%$ to $59\%$. The interpretation for the very low values is largely due to the initialization step that relies on fully labelled data. Given $5$ clients, $20$ rounds and a $50/50$ training/test split, the amount of data used in the initialization represents $2.5\%$ of the total dataset so even for an extremely low request budget of $0.1\%$, it is necessary to label $2.51\%$ of the dataset. Still, given that typical range for request budget in similar studies such as [@de2017confidence] starts at $10\%$ this is a very promising result, showing that our approach is efficient for extremely low request budgets thus fulfilling one of the requirements for applicability to aeronautical systems. Finally, we have included Figure \[federated\_active\_req\_per\_host\_acc\] to give insight as to what changes in the model for different request budget. Indeed, if the performance of the ensemble model is very stable, it is hiding the fact that the performance of the local DTs. Selecting recall as the best illustration of this, we see that the average recall of individual DTs increases from about $30\%$ to $37\%$ when the request budget increases. This is significant as local DTs would act in a real-world implementation as back-up in situations where the data link between the aircraft and the ground would be unavailable, as such, being able to guarantee a certain level of performance of the local model is important.
Overall, it can be noted that the performances of the Federated Active Learning approach that we propose are very stable and very close to state-of-the-art level of performance for non-distributed learning despite the additional constraints.
Conclusion
==========
In this paper, we propose a new applied Machine Learning framework to enable real-time distributed learning for aeronautical systems. To do so, we identify three requirements for a solution, namely $R_1$ the ability to classify normal samples with minimal computations, $R_2$ a mechanism to identify and transmit uncertain data under a communication budget and $R_3$ the ability to train a centralized model without accessing remote data. Then, we review existing solutions for each of the three requirements and propose a new method to fulfil all of them at the same time. Our method is a combination of Active Learning and Federated Learning. It relies on training confidence decision trees at client level and aggregating them at server level to create an ensemble model to act as an oracle. We provide a detailed performance analysis of our method on a public benchmark and compare our results to state-of-the-art classifiers.
Our approach achieves an accuracy of up to $61\%$ very close to state-of-the-art levels of performance of $62\%$ despite the additional constraint of Distributed Learning that implies that each client has to make decisions with partial information on the dataset and the constraint on limited computation power on the clients. Our method also achieves very consistent levels of performance with label budgets as low as $3\%$ thanks to the initialization step and the use of the ensemble model.
In general, with regards to the objective of this article of proposing a new communication-efficient and computationally cheap distributed learning framework suitable for the application case of real-time failure prediction for aeronautical systems, we have determined a new method that fulfils all three requirements we identified by combining existing works in an original way.
[^1]: https://moa.cms.waikato.ac.nz/datasets/
|
How Less Can Be More: Limited Palette Oil Painting
When you walk into the paint section of an art store, you are presented with every color of the rainbow, and some you’ve probably never even heard of. I know that I have spent a good chunk of my life (and my income) trying to get a tube of every possible color. Yet sometimes the most effective paintings are the ones that do a lot with only a few colors. These are called limited palette paintings, and you might be surprised at just how many successful paintings use only a minimal palette. Let’s take a look at some examples.
Analogous Color Palettes
Analogous color palettes are groups of colors that are found close together on the color wheel. For instance, a painting using only orange, red-orange and red would be said to have an analogous color palette.
The analogous color range doesn’t have to include only three colors, but the wider the range, the less likely the palette can be described as analogous. The values may range from light to dark, and even include pure blacks and whites, but the limited color range is most important. Thomas Eakins’ Sailing from 1775 is an example of an analogous color palette, using a range consisting of almost entirely desaturated warm yellows and oranges in different values.
As demonstrated above, an analogous palette can be especially effective at establishing a mood or tone. Eakins’ palette adds drama and atmosphere to an otherwise unremarkable and common painting subject.
In Picasso’s Mother and Child from 1902, the viewer is almost overwhelmed by soft blues and blue-greens. Picasso’s intention with this deep, cool analogous palette was almost certainly to create a feeling of quiet peacefulness.
Complementary Color Palettes
By drawing only from opposite ends of the color wheel and not the colors between, you are using a complementary color palette. Artists have understood for centuries that, for whatever reason, the relationship between complementary colors can create a strong visual dynamic. Blue and orange, red and green, yellow and violet are all striking combinations that draw the viewer’s eye, even when the colors are muted rather than saturated or intense. Degas’ A Roman Beggar Woman from 1857 is an excellent example.
If not for the cool blues and grays in the woman’s kerchief and scarf this would simply be an analogous palette. But the blue range serves to emphasize the warmth of the rest of the woman’s clothes and the background, just as those elements set off the blues. This visual dialogue makes Degas’ painting even richer and more powerful.
Paul Gauguin takes this same palette and intensifies it further in his Self-Portrait from 1894. Here the colors are highly saturated and almost fight with each other for the viewer’s attention, yet are balanced enough that the result is still pleasing to the eye, if not as harmonious as Degas’ painting.
Color Plus Neutrals
In this color relationship, there are typically one or two predominant saturated colors in the palette surrounded by desaturated neutral colors that can range in value. Manet’s The Fifer from 1866 is an excellent example of this type of limited palette. Manet’s nondescript background of mostly cool grays set off the rich saturated reds and yellows. The sharp colorless black of the jacket, hat and shoes serves as a perfect counterpoint to the red pants, which instantly become the most memorable element in the painting.
Despite the name, there are countless examples of limited palettes, and there are an infinite number of color relationships to explore even in a small corner of the color wheel.
What are some of your favorite color combinations? What kind of limited palette would you like to try next?
|
Israel Eliraz
Israel Eliraz (Hebrew: ישראל אלירז; born Israel Rothstein on 23 March 1936 [Hebrew: ישראל רוטשטיין]; died on 22 March 2016) was an Israeli poet who won the Bialik Prize (2008), the Brenner Prize (2013), the ACUM lifetime achievement award (2003), the Nathan Alterman Award (2002), the Jerusalem Foundation-Jerusalem Municipality’s Belles-Lettres Award (1992 and 1999), the Award (1963 and 1965), the (2009), and the (1994, 2008, and 2009).
Biography
Born Israel Rothstein on 23 March 1936 to a religious family in the Knesset Yisrael neighborhood of Jerusalem his parents were Ya’akov and Shifra Rothstein. He attended Alliance Israélite Universelle and Hebrew University of Jerusalem where he specialized in Jewish literature and philosophy, received his master’s degree from Tel Aviv University where he specialized in comparative literary studies, started writing poetry in 1980, studied theatre at University of Paris between 1995 and 1996 on a stipend from the French government, and worked for a living as a teacher and principal at Gymnasia Rehavia and as an instructor at . Occasionally he would also lecture in France, Belgium, and the United States: for example he was visiting professor of drama at Virginia Commonwealth University in 1980 and was invited to in 1999 and to in 1998. He lived in Jerusalem all his life and was married to Naomi née Brunner with whom he had three children. Eliraz died on 22 March 2016 in Jerusalem, Jerusalem District. His resting place is in Kiryat Anavim, Mateh Yehuda Regional Council, Jerusalem District, Israel.
Publications
אלירז, ישראל. דְּבָרִים דְּחוּפִים: מבחר שירים 1980–2010, עורך: ד״ר דרור בורשטיין, עם ״איך אלירז יכול לשנות את חייך,״ עמ׳ 380–390 מאת ד״ר דרור בורשטיין. תל אביב–יפו: הוצאת הקיבוץ המאוחד, ה׳תש״ע/2010, 392 עמ׳.
Haaretz critic Prof. Dr. wrote about this book that upon reading it he felt as if he “received endless letters from the poet regarding existential wonders" while Iton 77 critic said that it changed his life.
אלירז, ישראל. הֵבִּט: שירה מביטה בציור – בעקבות ציוריו של יהושע (שוקי) בורקובסקי. תל אביב–יפו: קשב לשירה, ה׳תשע״ב/2012, 132 עמ׳.
Haaretz critic Ouzi Zur praised this book's “wonderful poetic introspection regarding the quintessence of seeing.”
אלירז, ישראל. כַּמָּה זְמַן עוֹד נִשְׁאַר אֵינֶנָּה שְׁאֵלָה אֶלָּא דֶּלֶת. תל אביב–יפו ובן־שמן: הליקון ומודן הוצאה לאור, ה׳תשע״ג/2013, 96 עמ׳.
Further reading
פנחס־כהן, חוה. אסכולה של איש אחד: דיאלוג עם ישראל אלירז, סדרת קו אדום – אמנות, עורך: גיורא רוזן. תל אביב–יפו: הוצאת הקיבוץ המאוחד, ה׳תשע״א/2011, 237 עמ׳.
References
Category:1936 births
Category:2016 deaths
Category:20th-century Israeli poets
Category:20th-century Israeli writers
Category:20th-century male writers
Category:21st-century Israeli poets
Category:21st-century Israeli writers
Category:21st-century male writers
Category:Bialik Prize recipients
Category:Brenner Prize recipients
Category:Deaths from cancer in Israel
Category:Hebrew University of Jerusalem alumni
Category:Israeli academics
Category:Israeli expatriates in Belgium
Category:Israeli expatriates in France
Category:Israeli expatriates in the United States
Category:Israeli Jews
Category:Israeli male poets
Category:Israeli people of Ashkenazi descent
Category:Jewish educators
Category:Heads of schools in Israel
Category:Secular Jews
Category:Tel Aviv University alumni
Category:University of Paris alumni
Category:Virginia Commonwealth University faculty
Category:Writers from Jerusalem
Category:20th-century Israeli educators
Category:21st-century Israeli educators
|
Influence of yogurt fermentation and refrigerated storage on the stability of protein toxin contaminants.
Dairy products sold in a ready-to-eat form present the risk that adulterants persisting through manufacturing, storage, and distribution would reach consumers. Pathogenic microbes, including shigatoxigenic strains of Escherichia coli and the toxins they produce, are common food safety hazards associated with dairy products. Ricin and abrin are plant-derived ribosome-inactivating protein toxins related to the shiga-like toxins produced by E. coli. Limited information exists on the effects of manufacturing processes on the stabilities of these heat-resistant ribosome-inactivating proteins in the presence of foods. The goal of this study was to determine how typical yogurt manufacturing and storage processes influence ribosome-inactivating protein toxins. Ricin and abrin were added to skim or whole milk and batch pasteurized. Complete inactivation of both toxins was observed after 30 minutes at 85 °C. If the toxins were added after pasteurization, the levels of ricin and abrin in yogurt and their cytotoxic activities did not change significantly during fermentation or refrigerated storage for 4 weeks. The activities of ricin and abrin were inhibited by skim milk, nonfat yogurt, whole milk, and whole milk yogurt. The results showed minimal effects of the toxins on yogurt pH and %titratable acidity but inhibitory effects of yogurt on toxin activity.
|
Theoretical study of the geometries and decomposition energies of CO₂ on Al₁₂X: doping effect of Al₁₂X.
The adsorption and decomposition of CO₂ molecule on X-centered icosahedronal Al₁₂X clusters (doping atom X=Al, Be, Zn, Fe, Ni, Cu, B, C, Si, P) were investigated by the DFT methods of PW91 and PWC. Adsorption energies, chemisorption energies and energy barriers of physic- and chemisorptions for CO₂ were determined. It was found that the doping atoms and spin states have important influences on the Al₁₂X geometries, electronic properties and energies of the adsorption processes. CO₂ chemisorption on the Al₁₂C cluster is energetically and kinetically unfavorable. CO₂ decomposition on the metallic doping Al₁₂X (X=Fe, Ni, Cu) clusters has relatively low energy barriers. On contrary, the barriers are large when X=B, C, Si and P. The energy barriers for CO₂ chemisorption and decomposition on the Al₁₂Fe cluster are 5.23 kJ/mol and 38.53 kJ/mol, respectively. These values are the lowest among all the clusters being discussed. The adsorption and decomposition of CO₂ on the Al₁₂X cluster can be tuned by X doping.
|
Introduction {#Sec1}
============
The first global report on antibiotic resistance by the World Health Organization shows carbapenem resistance rates above 40% in *Acinetobacter baumannii*. Additionally, this bacterium is a leading cause of nosocomial infections, especially in intensive care units (ICUs), causing outbreaks with high mortality rates \[[@CR1], [@CR2]\]. *A. baumannii* has a natural resistance to some classes of antibiotics and an extraordinary ability to acquire resistance to others, including those regarded as last-line antibiotics, such as carbapenems. The most frequent carbapenemases in *A. baumannii* are the class D serine oxacillinases, which mainly belong to OXA-23. OXA-24, OXA-58, and OXA-143-like types begin mobilizing within plasmids or are inserted directly in chromosomes. In Colombia and other Latin American countries, the OXA-23 group-derived carbapenemases are the most prevalent. Carbapenem resistance has led to the need to use "old" antibiotics, such as fosfomycin and polymyxins (B and E). However, the use of these antibiotics has already caused the emergence of resistant isolates. In this scenario, without new antibiotics, the last option is treatment with combination therapies of two or more antibiotics \[[@CR3]\]. This report describes a case of a patient with bacteremia caused by a pan-resistant *A. baumannii* isolate.
Case presentation {#Sec2}
=================
On March 11, 2016, a 64-year-old mestizo man with a history of benign prostatic hyperplasia and use of an indwelling catheter presented to our emergency department with urinary retention. Cystoscopy revealed intravesical clots and obstructive bilobar prostate. Following the procedure, the patient exhibited signs of systemic inflammatory response syndrome and pathologic urinalysis. Therefore, antibiotic therapy with ampicillin/sulbactam was initiated without improvement in the clinical features. The urine culture report showed the presence of carbapenem-sensitive *Pseudomonas aeruginosa* and *Enterobacter cloacae*. Therefore, the therapy was subsequently escalated to meropenem (1 g every 8 h). The patient presented with clinical deterioration and ventilatory failure and was referred to the ICU for orotracheal intubation. He also developed cardiopulmonary arrest, which required basic and advanced resuscitation techniques for 11 min with subsequent sinus rhythm. On physical examination, the patient was under sedation, tachycardic, and hypothermic with evidence of purulent urethral discharge, and he required vasopressor support and sedoanalgesia. He had multiple organ dysfunctions due to urinary and pulmonary sepsis with the identification of carbapenem-resistant *Klebsiella pneumoniae* in the blood and lower respiratory tract secretions. The antibiotic therapy was adjusted to colistimethate (90,000 IU/kg) divided into three daily doses, doripenem (1 g every 8 h), and fosfomycin (4 g every 6 h). The patient initially progressed toward improvement, but 72 h later, he presented with new signs of inflammatory response. Therefore, a new blood culture was performed, revealing a carbapenem-resistant *A. baumannii* isolate (minimum inhibitory concentrations \[MICs\] ≥ 16, ≥ 16, and ≥ 8 μg/ml for imipenem, meropenem, and doripenem, respectively) that was also resistant to gentamicin (MIC ≥ 16 μg/ml), ciprofloxacin (MIC ≥ 4 μg/ml), and colistin (MIC 16 μg/ml) and sensitive only to tigecycline (MIC 1 μg/ml). Following this test, tigecycline administration was initiated (100 mg every 12 h), and colistimethate was suspended, with fosfomycin and doripenem continued until the completion of 14 days of treatment. Nevertheless, due to the persistent systemic inflammatory response on day 10, rifampicin (600 mg/day) and ertapenem (1 g every 12 h) were initiated for pharmacological synergism until day 14. With the prescribed antibiotic therapy, the patient improved in terms of infection, with an absence of fever and decreased leukocyte count. New blood and urine cultures were done, with negative results.
Molecular typing confirmed the genus and species of the isolate as *A. baumannii* (excluding other species of the *Acinetobacter* genus), as well as the presence of the *bla*~OXA-23~ gene (associated with carbapenem resistance). The isolate belonged to the sequence type (ST) 944, which has also been described in isolates identified in Russia, Italy, and the United States. Studies have identified intrahospital outbreaks of colistin-resistant bacteria with predominant genotype according to multilocus sequence typing was ST-258, ST11, ST273, and ST15 for *K. pneumoniae*; ISAba1 in *P. aeruginosa*, and pmrA1 and pmrB genotype ST94 for *A. baumannii* \[[@CR4]\]*.* There are few reports on the resistance mechanisms associated with resistance to colistin in *A. baumannii*, the most important one being that related to loss in the production of lipid A of lipopolysaccharide. This loss has been associated with mutations in the *pmrAB* and *lpxACD* operons, *lpxK*, or the *IS*Aba11 insertion \[[@CR5], [@CR6]\]. The sequencing of these two operons in the *A. baumannii* isolate identified two nonsynonymous mutations (H31P and I215M) in the *lpxC* gene from the ATCC 19606 strain and another nonsynonymous mutation (A178G) in the *pmrC* gene from the ATCC 17978 strain \[[@CR7]--[@CR9]\]. None of these mutations has previously been reported in colistin-resistant strains. Although they have been found in the *A. baumannii* isolate, their possible association with resistance should be confirmed experimentally.
Discussion {#Sec3}
==========
*A. baumannii* infections are a public health problem worldwide because there are fewer and fewer treatment options available. Multidrug-resistant (MDR) *A. baumannii* infection rates of up to 60% and mortality rates of 7% \[[@CR10], [@CR11]\] have been reported. One of the possible causes of this resistance is the indiscriminate use of antibiotics such as carbapenems, which are used in monotherapy in patients without comorbidities, resulting in increased resistance \[[@CR12]\]. In the particular case of *A. baumannii* infection, the treatment options include the combination of carbapenems with tigecycline, aminoglycosides, and polymyxin B and E (colistin) \[[@CR13]\]. Combination therapy including carbapenems aims for the "suicidal effect of the carbapenem," and although it has been used in *K. pneumoniae,* it is also a treatment possibility for *A. baumannii* \[[@CR14]\].
Regarding the use of polymyxin B or colistin, the former is usually preferred because its dosage is easier and more effective, the time to achieve optimal serum concentration is faster, there is less variability in pharmacokinetics, and no adjustment is required for renal failure. However, in our patient, the use of colistin was preferred due to the need to cover a pulmonary nidus and the possibility of inhaled administration \[[@CR15]--[@CR17]\]. Furthermore, fosfomycin is a broad-spectrum antibiotic that is especially used in combination therapy because it decreases the MIC of MDR strains and its use in monotherapy induces the development of resistance \[[@CR18]\]. Currently, the clinical use of tigecycline in MDR infections is under discussion. Some meta-analyses suggest that it should be the treatment of choice in patients with carbapenemase-producing MDR gram-negative bacteria in combination with other classes of antibiotics. Synergism with polymyxins or carbapenems is not as clear as with other antibiotics such as fosfomycin or aminoglycosides \[[@CR17]\]. Rifampicin was included in our patient's regimen because this antibiotic, although effective in monotherapy, has a good synergistic activity with other antimicrobials, especially colistin, by permeabilizing the bacterial membrane and increasing the treatment effectiveness \[[@CR19]\]. Although cefiderocol---a cephalosporin---is not a new class of antibiotic, it has a novel method of penetrating the tough outer membrane of gram-negative bacteria, including MDR strains. Attached to the drug's main molecule is a siderophore, a compound secreted by bacteria to seek out iron, which bacteria need for survival, and transport it across cell membranes. The results of a phase 2 clinical trial suggest that a new treatment for patients with complicated urinary tract infections caused by MDR pathogens could be on the horizon \[[@CR20]\].
Conclusions {#Sec4}
===========
The emergence of colistin-resistant *A. baumannii* strains in the clinical setting, although still occasional, is increasing due to the growing use of this antibiotic. We have identified a strain of *A. baumannii* with mechanisms of resistance to colistin not previously reported in a patient with bacteremia, required treatment with multiple antibiotic schemes with adequate response. More studies are required to determine other possible genetic mechanisms associated with this resistance in order to develop a better approach to the management of patients with infections caused by this microorganism.
We acknowledge the clinical laboratory of Compensar at Clínica Universidad de La Sabana.
Funding {#FPar1}
=======
This work was financially supported by El Bosque University PCI-2014-63 and Administrative Department of Science, Technology and Innovation/Colciencias grant 1308-71250819.
Availability of data and materials {#FPar2}
==================================
Data sharing was not applicable to this article, because no datasets were generated or analyzed during the current study.
JCGC, HRA and AMCM: conception, patient revision, literature review; ZCR and JEP: bacterial molecular analysis, literature review; JCGG, HRA, AMCM, ZCR and JEP: articule review and final publication. All authors read and approved the final manuscript.
Ethics approval and consent to participate {#FPar3}
==========================================
Approval for publication was obtained from the local ethics committee of Clínica Universidad de La Sabana. A copy of the ethics committee approval is available for review by the Editor-in-Chief of this journal.
Consent for publication {#FPar4}
=======================
Written informed consent was obtained from the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.
Competing interests {#FPar5}
===================
The authors declare that they have no competing interests.
Publisher's Note {#FPar6}
================
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
|
1.. Introduction
================
Among the possible medium access control (MAC) techniques for wireless communication systems, the simplicity of random access schemes (*i.e.*, ALOHA, carrier sense multiple access (CSMA)) make them suited to be implemented in several standards for short range applications \[[@b1-sensors-14-05622],[@b2-sensors-14-05622]\]. Even though mitigation methods can be introduced in random access MAC (i.e., carrier sense multiple access with collision avoidance (CSMA-CA)), collisions are not completely avoidable. Nevertheless, some receivers have the ability to correctly receive a signal despite a significant level of co-channel interference, and collisions do not always lead to a total loss of the collided packets. This co-channel interference tolerance is called capture effect \[[@b3-sensors-14-05622]\]. In the presence of concurrent transmissions at the same carrier frequency (collisions), packet capture may happen even for low values of the signal-to-interference ratio (SIR).
The first papers in the literature about the capture effect mostly consider Frequency Modulation (FM) demodulators \[[@b3-sensors-14-05622]--[@b5-sensors-14-05622]\]. Later, the capture effect has been also studied in a variety of transceivers and MAC schemes, including ALOHA networks \[[@b6-sensors-14-05622]--[@b8-sensors-14-05622]\], IEEE 802.11 devices \[[@b9-sensors-14-05622]--[@b11-sensors-14-05622]\], Bluetooth radios \[[@b12-sensors-14-05622]\] and cellular systems \[[@b13-sensors-14-05622]\]. Focusing on IEEE802.15.4, several papers describing experimental results can be found in the literature \[[@b14-sensors-14-05622]--[@b22-sensors-14-05622]\]. To give some examples, in \[[@b14-sensors-14-05622]\], the packet capture probability of a Chipcom CC1000 transceiver \[[@b15-sensors-14-05622]\] is measured with the aim of exploiting the capture effect for collision detection and recovery. Another study \[[@b16-sensors-14-05622]\], carried out again with CC1000 transceivers, which work in the sub-1 GHz band, obtained a threshold for the capture probability for the case of one interferer, but unstable results for the multiple interferers. This early work seems to suggest that the number of interferers might have an important effect on the capture probability with CC1000 devices. However, in contrast to the previous CC1000 measurements, successive studies, carried out with Chipcon CC2420 transceivers \[[@b17-sensors-14-05622]\], which operate in the 2.4 GHz industrial, scientific and medical (ISM) band, show that the capture effect depends on the total interfering power, but it is independent of the number of interferers \[[@b18-sensors-14-05622],[@b21-sensors-14-05622]\]; therefore, in our mathematical analysis, we assume that the performance is independent of the number of interferers for the 2.4 GHz physical layer (PHY) of IEEE 802.15.4, and as a result, we use only one interfering signal for the mathematical evaluation. A behavior on the packet capture similar to that reported in \[[@b16-sensors-14-05622],[@b18-sensors-14-05622]\] is also observed in \[[@b20-sensors-14-05622]\] with Freescale MC1224 transceivers \[[@b22-sensors-14-05622]\], which, again, operate in 2.4 GHz. The experiments conducted in interferer power-dominant (with respect to the noise) environments in \[[@b16-sensors-14-05622],[@b18-sensors-14-05622]--[@b20-sensors-14-05622]\] show a couple of common behaviors: (i) the receiver starts capturing the useful packet when signal-to-interference-plus-noise ratio (SINR) goes beyond 0 dB; (ii) packet reception rate (PRR) reaches one for values of SINR larger than 4 dB (please see in \[[@b16-sensors-14-05622]\], [Figure 5](#f5-sensors-14-05622){ref-type="fig"} and Figure 16, for CC1000 and CC2420, respectively; in \[[@b18-sensors-14-05622]\], [Figure 4](#f4-sensors-14-05622){ref-type="fig"} for CC2420; in \[[@b19-sensors-14-05622]\], [Figure 7c](#f7-sensors-14-05622){ref-type="fig"} for CC2420; and in \[[@b20-sensors-14-05622]\], [Figure 3](#f3-sensors-14-05622){ref-type="fig"} for MC13192).
Although the experimental results in \[[@b16-sensors-14-05622],[@b18-sensors-14-05622]--[@b20-sensors-14-05622]\] agree with each other, there is no theoretical model, purely based on mathematical analysis, which can be applied to the 2.4 GHz PHY of IEEE 802.15.4 to explain such characteristics. Motivated by this consideration, we propose an analytical framework to investigate the behavior of the IEEE 802.15.4 2.4 GHz PHY layer. A review of low-rate wireless personal area network (LR-WPAN) solutions, including IEEE 802.15.4, can be found in \[[@b23-sensors-14-05622]\].
The impact of interference in wireless sensor networks plays a very important role and can severely degrade the overall performance of the network and the efficiency of the upper layers. In our opinion, this aspect has not been sufficiently addressed in past years. In a dense sensor network deployment, where many nodes are periodically sending data to the sink, concurrent transmissions are highly probable. However, the probability of having a collision because of more than two concurrent transmissions is relatively low \[[@b18-sensors-14-05622],[@b20-sensors-14-05622]\], thanks to CSMA-CA. In such conditions, the performance of the receiver depends on the overall amount of interferer signal energy and does not change with the number of interferers \[[@b18-sensors-14-05622],[@b21-sensors-14-05622]\]. Hence, solely the impact of one interferer on the capture probability will be considered in the mathematical analysis.
On the other hand, we believe this study can also contribute to the identification of signal reception models for network simulators. In particular, as shown in [Table 1](#t1-sensors-14-05622){ref-type="table"}, most common network simulators use signal-to-noise ratio threshold (SNRT)- or bit error rate (BER)-based models in order to decide the correct reception of a packet.
In SNRT-based models, the packet is correctly received if the signal-to-noise ratio (SNR) is larger than a given threshold, whereas, in BER-based models, the packet reception decision is based on the BER, which is determined probabilistically depending on the value of the SNR. These models are rather simple, but have some drawbacks. In particular, SNR-based models do not take into account the impact of interference. This latter effect can be considered, in principle, by BER-based models, but the impact of the waveform of the interferer signals should be carefully considered, as it plays a significant role. Typically, conventional interference models are based on the assumption that the disturbance can be modeled as a Gaussian random variable; unfortunately, this is not the case of IEEE 802.15.4 systems, where only a limited number of strong interferers is present. To counteract this problem, we mathematically analyze the impact of the waveform of the interferer on packet reception and obtain curves that are organized as specific look-up tables. Figures, such as those derived in [Figures 4](#f4-sensors-14-05622){ref-type="fig"} and [6](#f6-sensors-14-05622){ref-type="fig"}, can be used to provide accurate PHY models for network simulators. In that case, the conventional on-off behavior of SNRT-based models can be replaced by a probabilistic model, where the actual value of SIR leads to a given probability of packet loss. In other words, we provide a SIR-based signal reception model for the interference-dominant environments, where noise is not the serious cause of packet loss (i.e., enough transmit power is used or nodes are using the best links to reach the destinations in a dense sensor network deployment). Furthermore, [Figure 7](#f7-sensors-14-05622){ref-type="fig"} shows that, in the case of non-coherent detection in an interferer-dominant environment, an on-off model can be also applied. In any case, behavior changes when thermal noise cannot be neglected. As a conclusion, the results of this paper on chip error rate (CER) and PRR (see [Figures 6](#f6-sensors-14-05622){ref-type="fig"} and [7](#f7-sensors-14-05622){ref-type="fig"}) can be used within network simulators in terms of look-up tables. That allows a fast characterization of the behavior of the PHY layer.
The computational complexity of the model for the coherent detection is O(1) (in big O notation). This makes it usable without intensive computational effort. For the non-coherent case, we show that the performance curve has a step-like behavior with the threshold at 0 dB. This simple model can capture the behavior of the non-coherent case without any computational effort.
The rest of the paper is organized as follows: Section 2 describes CSMA-CA and the 2.4 GHz PHY of the IEEE 802.15.4 Standard. Sections 3 and 4 evaluate CER for coherent and non-coherent offset quadrature phase shift keying (O-QPSK), respectively. Section 5, obtains PRR in concurrent transmission by finding an upper bound, which transforms CER to data symbol error rate (DSER).
2.. System Description
======================
Two different approaches are foreseen in IEEE 802.15.4 to coordinate the data traffic: beacon-enabled mode and non-beacon-enabled mode \[[@b2-sensors-14-05622]\]. In the former, periodic beacons are sent by the coordinator to synchronize the channel access, while in the latter, no synchronization is required. CSMA-CA is used in both modes, but in a slightly different way. In the beacon-enabled mode, every node synchronizes itself to the backoff slots determined by the coordinator; for this reason it is named slotted CSMA-CA. In non-beacon-enabled mode, every node has its own backoff slot timing, so that it is called unslotted CSMA-CA. Default values of the parameters and the number of sensing phases before assessing the channel idle are different in slotted and unslotted versions as a result of synchronization (*i.e.*, two backoff slots in slotted, one backoff slot in unslotted). In both modes, each node waits for a random number of backoff slots, then channel sensing is performed. If the channel is found to be free, the node immediately starts the transmission. If the channel is found to be busy, the node turns back to the backoff state again. There is a maximum number of attempts a node can try to sense the channel. When this maximum is reached, the algorithm ends with a failure. Even though CSMA-CA avoids collisions by sensing the channel before transmitting a packet, a hidden terminal effect or sensing the channel as idle exactly at the same time by more than one node can cause a collision. As discussed in the previous Section, a collision does not always lead to a total loss. One of the packets, most probably the one with the better signal strength, can be successfully received.
IEEE 802.15.4 PHY uses direct sequence spread spectrum (DSSS), and it operates over different ISM bands: the 868 MHz (EU) ISM band, the 915 MHz (US) ISM band and the 2.4 GHz global ISM band. Because of the global availability of the higher frequency band, currently transceiver manufacturers mostly have products working in this band (*i.e.*, Chipcon CC2420, Freescale MC13224, *etc.*). In the paper, we focus on the 2.4-GHz frequency band, which utilizes half-sine shaped O-QPSK modulation. In the PHY layer of the 2.4 GHz band, the signal is first modulated by forming a data symbol of four bits, and then, this symbol is mapped to 32-chip long sequences, obtaining 16 data symbols. The data symbol set modulated at the carrier frequency, *f~c~*, and carrier phase, *ϕ~C~*, can be written as: $$\begin{matrix}
{s_{Ci}(t) = I_{i}(t)\cos(2\pi f_{c}t + \phi_{C}) - Q_{i}(t)\sin(2\pi f_{c}t + \phi_{C})} & {i = 0,1,\ldots,15} \\
\end{matrix}$$
For each chip sequence, even-indexed chips are modulated to the in-phase (I) carrier, and odd-indexed chips are modulated to the quadrature-phase (Q) carrier of O-QPSK. The offset between the I-phase and Q-phase is formed by delaying the Q-phase symbols one chip duration with respect to the I-phase. *I*(*t*) and *Q*(*t*) for each data symbol can be written as in \[[@b25-sensors-14-05622]\]: $$I_{i}(t) = \underset{n = 0}{\overset{15}{\text{∑}}}{c_{2n}^{i}h(t - 2nT_{c})}$$ $$Q_{i}(t) = \underset{n = 0}{\overset{15}{\text{∑}}}{c_{2n + 1}^{i}h(t - (2n + 1)T_{c})}$$where $c_{k}^{i}$ is ±1 with respect to the chip value and *i* is the index of the data symbols. The function, *h*(*t*), is half-sine shaped and can be expressed as: $$h(t) = \begin{cases}
{\cos\left( \frac{\pi t}{2T_{c}} \right)} & {\text{if} - T_{c} \leq t < T_{c}} \\
0 & \text{otherwise} \\
\end{cases}$$Finally, the complex baseband signal is given as: $${\widetilde{s}}_{C_{i}}(t) = I_{i}(t) + jQ_{i}(t);$$the corresponding data symbol set in time domain will be expressed by using the complex envelope as: $$s_{Ci}(t) = \Re\left\lbrack {{\widetilde{s}}_{C_{i}(t)}\exp\left( {j(2\pi f_{c}t + \phi_{C})} \right)} \right\rbrack$$where ℜ\[·\] is the real part of the complex expression.
3.. Probability of Chip Error in Coherent O-QPSK
================================================
In this section, we consider coherent communication and evaluate the chip error probability when there is one interferer with an unknown carrier phase and symbol timing. As the experiments in \[[@b16-sensors-14-05622],[@b18-sensors-14-05622]--[@b20-sensors-14-05622]\] suggest, we assume that the signal of the interferer is additive, and the probability of error is independent of the number of interferers. As a consequence, we will consider one interfering signal. Under such an assumption, the received O-QPSK modulated signal, in the time interval −*T~c~* ≤ *t* \< *T~c~*, can be written as: $$\begin{matrix}
{r(t) = \Re\left\lbrack {{\widetilde{s}}_{C}(t)\exp\left( {j(2\pi f_{c}t + \phi_{C})} \right)} \right\rbrack + \Re\left\lbrack {{\widetilde{s}}_{I}(t)\exp\left( {j(2\pi f_{c}t + \phi_{I})} \right)} \right\rbrack} & {- T_{c} \leq t < T_{c}} \\
\end{matrix}$$where *s͂~C~* (*t*) and *s͂~I~* (*t*) are the complex envelopes of the useful and interferer signals and *ϕ~C~* and *ϕ~I~* are the carrier phases of the useful and interferer signals, respectively. There is no phase difference between the useful transmitter and the receiver, since we are assuming coherent demodulation. Therefore *ϕ~C~* = 0, but the carrier phase of the interferer, *ϕ~i~*, is unknown. After some algebra, [Equation (6)](#FD7){ref-type="disp-formula"} can be expressed as: $$\begin{matrix}
{r(t) = \Re\left\lbrack {\widetilde{r}(t)\exp(j2\pi f_{c}t)} \right\rbrack} & {- T_{c} \leq t < T_{c}} \\
\end{matrix}$$where the complex envelope is: $$\begin{matrix}
{\widetilde{r}(t) = {\widetilde{s}}_{C}(t) + {\widetilde{s}}_{I}(t)\exp(j\phi_{I})} & {- T_{c} \leq t < T_{c}} \\
\end{matrix}$$
In the case of coherent demodulation, the I and Q components should be sampled at the instance corresponding to the maximum signal amplitude; hence for a sequence of chips, the I-phase should be sampled at the instants *t~I~n~~* = 2*nT~c~* (see [Equation (2)](#FD2){ref-type="disp-formula"}) and the Q-phase at the instants *t~Q~n~~* = (2*n* + 1)*T*~c~ (see [Equation (3)](#FD3){ref-type="disp-formula"}), where *n* = 0, 1, 2,.... This results in sampling the I and Q components of the useful signal at the indicated sampling instances in [Figure 1a](#f1-sensors-14-05622){ref-type="fig"} and [Figure 3a](#f3-sensors-14-05622){ref-type="fig"}. Due to the delayed sampling between the I and Q components, we will consider a Q-I constellation plane, where the Q component is delayed by *T~c~* seconds with respect to the I component. We named the resulting constellation plane Q*~d~*-I to emphasize the delay. In the next two subsections, we derive the CER for the received signal in [Equation (8)](#FD9){ref-type="disp-formula"}, as a function of SIR, considering two cases: without and with pulse shaping. Thermal noise is neglected in the analysis, as we want to investigate solely the effect of interference on CER. The effect of noise will be considered in Section 5 to compare the analysis with the experimental results.
3.1.. Without Pulse Shaping
---------------------------
When we take the case of no pulse shaping into account, the shaping function, *h*(*t*), is a non-return to zero (NRZ) rectangular pulse, whose expression is: $$h(t) = \begin{cases}
1 & {- T_{c} \leq t < T_{c}} \\
0 & \text{otherwise.} \\
\end{cases}$$
Accordingly, the complex envelope of the received signal in [Equation (8)](#FD9){ref-type="disp-formula"} on the Q*~d~*-I constellation plane can be written as: $$\begin{matrix}
{\widetilde{r}(t) = \left( {\pm \sqrt{\frac{C}{2}} \pm j\sqrt{\frac{C}{2}}} \right) + \left( {\pm \sqrt{\frac{I}{2}} \pm j\sqrt{\frac{I}{2}}} \right)h(t)\exp(j\phi_{I})} & \begin{matrix}
{- T_{c} \leq t < T_{c},} & {0 \leq \phi_{I} < 2\pi} \\
\end{matrix} \\
\end{matrix}$$where *C* and *I* represent the energies of the useful and interferer signals, respectively. ϕ*~I~* is the carrier phase of the interferer. Since the demodulation is coherent, perfect knowledge of the carrier phase and symbol timing is assumed for the useful signal. These parameters are supposed to be uniformly distributed random variables for the interfering signal. The pulse shaping function is constant throughout a chip duration; thus, random symbol timing of the interferer, *t*, does not give any contribution to the performance. On the contrary, the presence of an interferer carrier phase has a significant impact on the received signal, because *ϕ~I~* rotates the interfering signal on the Q*~d~*-I domain, as indicated in the second addend of [Equation (10)](#FD11){ref-type="disp-formula"}. The geometric representation of [Equation (10)](#FD11){ref-type="disp-formula"} on the Q*~d~*-I constellation plane is depicted in [Figure 1](#f1-sensors-14-05622){ref-type="fig"}.
With respect to the transmitted symbol, the first addend in [Equation (10)](#FD11){ref-type="disp-formula"} will be one of the constellation points on the Q*~d~*-I domain, as shown in Figure1b, while the second addend in [Equation (10)](#FD11){ref-type="disp-formula"} represents the interfering signal, which can be any of the points shown in [Figure 1d](#f1-sensors-14-05622){ref-type="fig"}, but rotated according to the interferer carrier phase, *ϕ~I~*. The final sum in the equation can be interpreted as in [Figure 1e](#f1-sensors-14-05622){ref-type="fig"}; the sampling point of the received signal is located in one of the quadrants in the Q*~d~*-I plane, in agreement with the transmitted useful symbol, but, due to the presence of the interferer, it is shifted to a random location over the circle with radius $\sqrt{I}$ and the center at $\left( {\pm \sqrt{\frac{C}{2}}, \pm j\sqrt{\frac{C}{2}}} \right)$. The amplitude of the useful signal on the Q*~d~*-I plane is $\sqrt{C}$, and the amplitude of the interferer signal is $\sqrt{I}$. In the following parts, *P~c~* and *P~s~* will denote the probability of chip error and the probability of O-QPSK symbol error, respectively If we consider an ideal maximum likelihood demodulator \[[@b26-sensors-14-05622]\], three distinct cases need to be investigated.
### 3.1.1.. Case (a): $\sqrt{I} < \sqrt{C/2}$
When $\sqrt{I}$ is smaller than $\sqrt{C/2}$, as shown in [Figure 2a](#f2-sensors-14-05622){ref-type="fig"}, the received signal never crosses other symbol regions. As a consequence, the demodulator will never make an error, and therefore: $$\begin{array}{ll}
P_{c} & {= 0} \\
P_{s} & {= 0} \\
\end{array}$$
### 3.1.2.. Case (b): $\sqrt{C/2} \leq \sqrt{I} < \sqrt{C}$
When $\sqrt{I}$ is larger than (or equal to) $\sqrt{C/2}$, the demodulator will start deciding erroneously, because of the stronger signal energy of the interfering component, which forces the received signal to be sampled in one of the adjacent quadrants on the constellation plane, as shown in [Figure 2b](#f2-sensors-14-05622){ref-type="fig"}. In such a condition, the probability of symbol error will be given by: $$P_{s} = \frac{S_{0}}{\pi} = 2\frac{2\alpha}{2\pi} = \frac{2\alpha}{\pi}$$where *S*~0~ is the arc of the circle shown in [Figure 2b](#f2-sensors-14-05622){ref-type="fig"}, the angle, *α*, in radians, is equal to *S*~0~/2, and *α* is: $$\alpha = \arctan\sqrt{\frac{I - (C/2)}{\sqrt{C/2}}}$$
The sampling point of the received signal can only pass to the adjacent quadrants, and the receiver makes a one chip error per each transmitted symbol; as a result, *P~c~* will be the half of *P~s~*: $$P_{c} = \frac{4\alpha}{4\pi} = \frac{\alpha}{\pi}$$
### 3.1.3.. Case (c): $\sqrt{C} \leq \sqrt{I}$
For increasing values of $\sqrt{I}$, the circle will move to the other side of the origin (see [Figure 2c](#f2-sensors-14-05622){ref-type="fig"}). *α* can still be written as in [Equation (11)](#FD14){ref-type="disp-formula"}, but now: $$\begin{array}{ll}
S_{0} & {= \alpha} \\
S_{1} & {= \pi/2 - \alpha} \\
S_{2} & {= 2\alpha - \pi/2} \\
\end{array}$$
When the received signal is in *S*~2~, the receiver makes two chip errors per each transmitted symbol. On the other hand, when it is in *S*~0~ or in *S*~1~, the receiver makes one chip error per each transmitted symbol. Regarding the number of chip errors in the different quadrants, *P~s~* and *P~c~* can be written as: $$\begin{array}{ll}
P_{s} & {= \frac{2(S_{0} + S_{1}) + S_{2}}{2\pi} = \frac{\alpha}{\pi} + \frac{1}{4}} \\
P_{c} & {= \frac{2(S_{0} + S_{1})}{2\pi}\frac{1}{2} + \frac{S_{2}}{2\pi} = \frac{\alpha}{\pi}} \\
\end{array}$$
The computational complexity of *P~s~* and *P~c~* is *O*(1) in big O notation, since *C* and *I* are constant values.
3.2.. Half-Sine Pulse Shaping
-----------------------------
Using the pulse shaping given by [Equation (4)](#FD4){ref-type="disp-formula"}, the received signal in [Equation (8)](#FD9){ref-type="disp-formula"} can be written, on the Q*~d~*-I plane, as: $$\begin{matrix}
{\widetilde{r}(t) = \left( {\pm \sqrt{\frac{C}{2}} \pm j\sqrt{\frac{C}{2}}} \right) + \left( {\pm \sqrt{\frac{I}{2}} \pm j\sqrt{\frac{I}{2}}} \right)\cos\left( {\frac{\pi t}{2T_{c}} + \phi_{I}} \right)} & \begin{matrix}
{- T_{c} \leq t < T_{c},} & {0 \leq \phi_{I} < 2\pi} \\
\end{matrix} \\
\end{matrix}$$where the first addend represents the useful symbol and the second addend represents the interferer symbol. In the variable, *t*, in [Equation (12)](#FD18){ref-type="disp-formula"}, is the sampling instance for the interferer symbol, and *ϕ~I~* is the carrier phase of the interferer, which shifts the pulse shaping function on the Q*~d~*-I domain.
When there is perfect knowledge of the useful carrier phase and symbol timing, there is no difference, from the demodulator point of view, between half-sine-shaped or not shaped useful symbols, since the sampling is done at the maximum amplitude instance. As a consequence, the first addends in [Equation (10)](#FD11){ref-type="disp-formula"} and [Equation (12)](#FD18){ref-type="disp-formula"} are equal; this can also be observed by comparing [Figure 1b](#f1-sensors-14-05622){ref-type="fig"} with [Figure 3b](#f3-sensors-14-05622){ref-type="fig"}. When the pulse shaping function, $\cos(\frac{\pi t}{2T_{c}} + \phi_{I})$, of the interferer is included to the scenario, the interferer signal will be positioned on one of the diagonal lines shown in [Figure 3d](#f3-sensors-14-05622){ref-type="fig"}. As a consequence, the received signal plus interferer will be positioned over the dashed lines of [Figure 3e](#f3-sensors-14-05622){ref-type="fig"}. To obtain the error probability, we need to derive the probability distribution of the amplitude of the interferer at the sampling instance.
The random symbol sampling instance, *t*, and carrier phase, *ϕ~I~*, of the interferer can be modeled as uniformly distributed random variables. To simplify the notation in a pulse shaping function, we substitute the random variables as $X = \frac{\pi t}{2T_{c}}$, *Y* = *ϕ~I~* and *Z* = *X* + *Y* and denote the Probability Density Functions (PDFs) of *X* and *Y* as *f~X~*(*x*) and *f~Y~*(*y*); then the PDF of Z, which is our interest, will be the convolution integral \[[@b27-sensors-14-05622]\] given as: $$\begin{array}{ll}
{f_{Z}(z) = {\int_{- \infty}^{\infty}{f_{X}(x)f_{Y}(z - x)dx}}} & {= \begin{cases}
\frac{\pi + 2z}{4\pi^{2}} & {\frac{- \pi}{2} \leq z < \frac{\pi}{2}} \\
\frac{1}{2\pi} & {\frac{\pi}{2} \leq z < \frac{3\pi}{2}} \\
{\frac{5\pi - 2z}{4\pi^{2}},} & {\frac{3\pi}{2} \leq z < \frac{5\pi}{2}} \\
0 & \text{otherwise} \\
\end{cases}} \\
& \begin{array}{ll}
{= \frac{1}{2\pi}} & {0 \leq z < 2\pi} \\
\end{array} \\
\end{array}$$where: $$\begin{matrix}
{f_{X}(x) = \begin{cases}
{\frac{1}{\pi},} & {\frac{- \pi}{2} \leq x < \frac{\pi}{2}} \\
0 & \text{otherwise} \\
\end{cases}} & {f_{Y}(y) = \begin{cases}
{\frac{1}{2\pi},} & {0 \leq y < 2\pi} \\
0 & \text{otherwise} \\
\end{cases}} \\
\end{matrix}$$
The solution of the convolution integral in [Equation (13)](#FD19){ref-type="disp-formula"} indicates that *Z* is a uniformly-distributed random variable in the interval 0 ≤ *z* \< 2*π*. Thus, the sampled point of the interferer signal is $( \pm \sqrt{\frac{I}{2}} \pm j\sqrt{\frac{I}{2}})\cos(z)$, and it always stays over the dashed lines in [Figure 3d](#f3-sensors-14-05622){ref-type="fig"}. If the interferer is transmitting *m*~3~ or *m*~0~, the slope of the diagonal line is +1; on the other hand, if the interferer is transmitting *m*~1~ or *m*~2~ then the slope is − 1. The probability distribution of the positions of the interferer sampling point on the dashed lines are symmetric on the opposite sides of the origin, because of the cosine function. The amplitude in a quadrant is: $$\begin{matrix}
{i = \sqrt{I}\cos(z)} & {- \pi/2 \leq z < \pi/2} \\
\end{matrix}$$
Using [Equation (14)](#FD21){ref-type="disp-formula"}, the PDF of the amplitude is written as: $$\begin{matrix}
{f_{i}(i) = \frac{2}{\pi\sqrt{1 - {({i/\sqrt{I}})}^{2}}}} & {0 \leq i \leq \sqrt{I}} \\
\end{matrix}.$$Then, we can also obtain the probability that the interferer amplitude lies in an interval (*a, b*\] using [Equation (15)](#FD22){ref-type="disp-formula"}: $$\text{Prob}(a < i \leq b) = F_{i}(b) - F_{i}(a) = {\int_{\frac{a}{\sqrt{I}}}^{\frac{b}{\sqrt{I}}}{f_{i}(i)di = \frac{2}{\pi}(\sin^{- 1}(\frac{b}{\sqrt{I}}}}) - \sin^{- 1}(\frac{a}{\sqrt{I}}))$$where *F~i~*(*i*) is the Cumulative Distribution Function (CDF) of *i*.
Without loss of generality, as we did in Section 3.1, it can be assumed that the transmitter is transmitting m~3~. Therefore, to obtain *P~c~* and *P~s~*, we should consider the following two cases: (a) $\sqrt{I} < \sqrt{C}$, and (b) $\sqrt{I} \geq \sqrt{C}$
### 3.2.1.. Case (a): $\sqrt{I} < \sqrt{C}$
When $\sqrt{I}$ is smaller than $\sqrt{C}$, the demodulator never makes an error, since the received signal never crosses the other quadrants. Thus, *P~c~* and *P~s~* are equal to zero: $$\begin{array}{ll}
P_{s} & {= 0} \\
P_{c} & {= 0} \\
\end{array}$$
### 3.2.2.. Case (b): $\sqrt{I} \geq \sqrt{C}$
When *i*, which refers to the amplitude of the interferer signal at the sampling instance, is larger than or equal to $\sqrt{C/2}$, the receiver can decide for the wrong symbol with a probability of $\frac{3}{4}$. Regarding *P~c~*, a transition towards *m*~2~ or *m*~1~ will result in one chip error per symbol, while a transition towards *m*~0~ will result in two chip errors per symbol. Therefore, the error rates when $i > \sqrt{C/2}$ are: $$\begin{array}{ll}
P_{s} & {= \frac{3}{4}} \\
P_{c} & {= \frac{1}{4} \cdot \frac{1}{2} + \frac{1}{4} \cdot \frac{1}{2} + \frac{1}{4} = \frac{1}{2}} \\
\end{array}$$
In general, *P~s~* and *P~c~* can be written as a function of *i* as follows: $$\begin{array}{ll}
{P_{s}(i)} & {= \begin{cases}
{0,} & {i < \sqrt{C}} \\
{\frac{3}{4},} & {\sqrt{C} \leq i \leq \sqrt{I}} \\
\end{cases}} \\
{P_{c}(i)} & {= \begin{cases}
{0,} & {i < \sqrt{C}} \\
{\frac{1}{2},} & {\sqrt{C} \leq i \leq \sqrt{I}} \\
\end{cases}} \\
\end{array}$$Finally, we can derive the error probabilities when $\sqrt{I}$ is greater than $\sqrt{C}$ by evaluating the expected values of *P~s~*(*i*) and *P~c~*(*i*): $$\begin{array}{ll}
P_{s} & {= E\lbrack P_{s}(i)\rbrack = {\int_{\sqrt{C}}^{\sqrt{I}}{\frac{3}{4}f_{i}(i)di = \frac{3}{4} \cdot F_{i}(\sqrt{C} < i \leq \sqrt{I}) = \frac{3}{4}\frac{2}{\pi}(\frac{\pi}{2} - \sin^{- 1}(\sqrt{\frac{C}{I}})) = \frac{3}{4} - \frac{6}{4\pi}\sin^{- 1}(\sqrt{\frac{C}{I}})}}} \\
P_{c} & {= E\lbrack P_{c}(i)\rbrack = {\int_{\sqrt{C}}^{\sqrt{I}}{\frac{1}{2}f(i)di = \frac{1}{2} \cdot F_{i}(\sqrt{C} < i \leq \sqrt{I}) = \frac{1}{2}\frac{2}{\pi}(\frac{\pi}{2} - \sin^{- 1}(\sqrt{\frac{C}{I}})) = \frac{1}{2} - \frac{1}{\pi}\sin^{- 1}(\sqrt{\frac{C}{I}})}}} \\
\end{array}$$where *f~i~*(*i*) is given in [Equation (15)](#FD22){ref-type="disp-formula"} and *F~i~*(*i*) is given in [Equation (16)](#FD23){ref-type="disp-formula"}.
The computational complexity of *P~s~* and *P~c~* are *O*(1) in big *O* notation, since *C* and *I* are constant values.
3.3.. Validation of the Analytical Model through Simulations
------------------------------------------------------------
The validity of our analytical framework has been tested using Monte-Carlo simulations. The simulation tool has been implemented in MATLAB (The MathWorks Inc., Natick, MA, USA) using the complex envelopes of the signals. Asynchronous symbols are obtained by shifting the complex envelope in the time axis, while the random carrier phase is obtained through rotating the complex envelope on the complex plane. SIR values with a step of 0.3 dB from −10 dB to 5 dB are simulated. At each specific SIR value, 10, 000 random O-QPSK symbols are generated for both the interferer and useful transmitter. The complex baseband representation of the O-QPSK symbol is generated by using 100 points; therefore, the resolution of the time for the asynchronous O-QPSK symbols was 10 ns. On the other hand, the resolution for the carrier phase was 2*π*/100. Random asynchronous interfering symbols with a random carrier phase are generated with uniformly-distributed parameters. The excellent agreement between simulation points and analytical curves is shown in [Figure 4](#f4-sensors-14-05622){ref-type="fig"}. Since the analysis in Sections 3.1 and 3.2 is exact (no approximation is used), we expect a perfect overlapping between simulations and the analytical model. Asymptotically, there is a 3-dB difference between the results of half-sine shaped and non-shaped symbols. In half-sine pulse shaping, *P~c~* goes to zero at 0 dB, while the threshold is 3 dB in the absence of pulse shaping.
4.. Probability of Chip Error in Non-Coherent O-QPSK
====================================================
Normally, O-QPSK requires coherent detection \[[@b26-sensors-14-05622]\]; however with the half-sine pulse shaping in 2.4 GHz PHY of IEEE 802.15.4, the information-bearing part of the signal is not only the carrier phase, but also the complex-envelope of the signal. Thus, without recovering the carrier phase, just observing the phase changes of the complex envelope, transmitted information can be extracted. In 2.4 GHz PHY of IEEE 802.15.4, the phase of the complex envelope increases or decreases by an amount of *π*/2 every *T~c~* seconds. In this sense, the behavior of the modulation is similar to that of minimum shift keying (MSK). The main difference between the two modulations is that they use different symbol mappings. In particular, symbol mapping in IEEE 802.15.4 is shown in [Figure 5](#f5-sensors-14-05622){ref-type="fig"}.
Chip intervals in [Figure 5a](#f5-sensors-14-05622){ref-type="fig"} are marked as *T~C~*~1~,*T~C~*~2~, .... In each even chip interval, *T~C~*~2~,*T~C~*~4~,..., a symbol (*m*~0~, *m*~1~, *m*~2~, *m*~3~) is transmitted, and in the odd chip intervals, *T~C~*~3~,*T~C5~*, ..., there are phase changes (*t*~4~, *t*~5~, *t*~6~, *t*~7~), which can be considered as the transitions between two sequential O-QPSK symbols. For instance, in [Figure 5a](#f5-sensors-14-05622){ref-type="fig"}, a sequence of phase changes are shown. At the beginning of the *T~C~*~2~ chip interval (*i.e., ωt* = 0), the phase of the signal is equal to the carrier phase, *ϕ~C~*; then at the end of the same chip interval, the phase reaches *ϕ~C~* + *π*/2, which is indicated in [Figure 5b](#f5-sensors-14-05622){ref-type="fig"} by the arc, *m*~3~. In the next chip interval, *T~C~*~3~, the phase reaches the *ϕ~C~* + *π* value, which is represented by the arc, *t*~5~. The sequential chip intervals send the phase of the complex envelope, *π*/2, further or back with respect to the transmitted symbols. All possible phase transitions are shown in [Figure 5b](#f5-sensors-14-05622){ref-type="fig"}. This relationship between phase transition and transmitted symbol suggests the idea of constructing a non-coherent detector, which identifies the transmitted chip based on the phase transitions of the complex envelope. Using such a detector, *P~c~* will refer to the probability of the erroneous detection of phase transitions during a chip interval. In the absence of interference, the complex envelope of the received chip through a chip duration in the Q-I plane can be expressed as: $$\begin{matrix}
{R(\omega t) = \sqrt{C}e^{j( \pm \omega t + \phi_{C})},} & {\omega t\in\lbrack 0,\pi/2\rbrack,} & {\phi_{c}\in\lbrack 0,2\pi\rbrack} \\
\end{matrix}$$where *C* is the energy of the signal and *ϕ~C~* is the uniformly distributed carrier phase. The definition domain of *ωt* is from zero to *π*/2, because the signal can only move *π*/2 further on the Q-I plane during a chip interval. The instantaneous phase of *R*(*ωt*) in [Equation (17)](#FD28){ref-type="disp-formula"} is: $$\Theta_{R}(\omega t) = \pm \omega t + \phi_{C}$$and the phase of the received signal at the beginning and at the end of the chip interval is: $$\begin{aligned}
{\Theta_{R}(0)} & {= \phi_{C}} \\
{\Theta_{R}(\pi/2)} & {= \pm \pi/2 + \phi_{C}.} \\
\end{aligned}$$Since the carrier phase can be eliminated by subtracting Θ*~R~*(*π*/2) and Θ*~R~*(0), the following discriminator demodulator can be applied \[[@b28-sensors-14-05622]\]: $$\Theta_{R}(\pi/2)\underset{- j}{\overset{+ j}{\gtrless}}\Theta_{R}(0)$$where the decision is based on the comparison of the phases of the complex envelope of the received signal at the beginning and at the end of the chip interval. The demodulator decides for a positive phase change, +*j*, if the phase at the end of the chip interval is higher than the beginning; otherwise, it decides for a negative phase change, −*j*. Firstly, we will analyze the case where the symbols of the interferer are synchronous (according to the chip interval) with the useful symbols. This assumption is not realistic, but provides the mathematical basis for the framework that will be used for the more realistic case, where the interferer and useful symbols are asynchronous.
4.1.. Interferer and Useful Symbols are Synchronous
---------------------------------------------------
When useful and interfering signals are synchronous, the received signal can be written as: $$R(\omega t) = \sqrt{C}e^{j(b_{C}\omega t + \phi_{C})} + \sqrt{I}e^{j(b_{I}\omega t + \phi_{I})}$$where *ωt* = \[0, *π*/2\], *ϕ~C~* = \[0, 2*π*\] and *ϕ~I~* = \[0, 2*π*\]. *b~C~* and *b~I~* represent the turning directions of the phase of the complex envelope of the useful and the interferer chips on the complex domain in [Figure 5b](#f5-sensors-14-05622){ref-type="fig"}: *bC, bI* = +1 means a clockwise direction, whereas *b~C~, b~I~* = −1 means the counter-clockwise direction. We denote *P~eq~* as the probability of erroneous detection when the phases of useful and interferer signals are turning to the same direction (*b~C~* = *b~I~*), while *P~dif~* denotes the probability of erroneous detection when phase transitions are toward the different directions (*b~C~* ≠ *b~I~*). There are two cases; accordingly, *P~c~* is: $$P_{c} = \frac{1}{2}P_{eq} + \frac{1}{2}P_{dif}$$
Note that if the phase of the interferer signal turns toward the direction of the useful signal, the received signal in the Q-I plane will turn toward the direction of the useful signal, as well. As a consequence, *P~eq~* will always be zero. To obtain *P~dif~*, we can consider the case where *b~C~* = +1 and *b~I~* = −1; in fact, this is mathematically identical to the case of *b~C~* = −1 and *b~I~* = +1. Motivated by this consideration, we can define the following success/error rule for the demodulator, based on the phase values at the beginning and at the end of the chip interval: $$\Theta_{R}(\pi/2)\underset{e}{\overset{s}{\gtrless}}\Theta_{R}(0)$$where *s* means success and *e* means error. When *b~C~* = +1 and *b~I~* = −1, the received signal is: $$R(\omega t) = \sqrt{C}e^{j(\omega t + \phi_{C})} + \sqrt{I}e^{j( - \omega t + \phi_{I})}$$To obtain an analytical expression to Θ*~R~*(*ωt*), we assume that during a chip interval, *R*(*ωt*) always stays on the definition domain of arctan (*i.e.*, \[−*π*/2 *π*/2\]). Starting from [Equation (23)](#FD35){ref-type="disp-formula"}, the instantaneous phase function can be written as: $$\Theta_{R}(\omega t) = \arctan\left( \frac{\sqrt{C}\sin(\omega t + \phi_{C}) + \sqrt{I}\sin( - \omega t + \phi_{I})}{\sqrt{C}\cos(\omega t + \phi_{C}) + \sqrt{I}\cos( - \omega t + \phi_{I})} \right)$$
With no loss of generality, we can assume *ϕ~c~* = 0. Now, after some trigonometric simplifications, the phases of the received signal at the beginning and at the end of the chip interval can be written as: $$\begin{array}{ll}
{\Theta_{R}(0)} & {= \arctan\left( \frac{\sqrt{I}\sin(\phi_{I})}{\sqrt{C} + \sqrt{I}\cos(\phi_{I})} \right)} \\
{\Theta_{R}(\pi/2)} & {= \arctan\left( \frac{\sqrt{C} - \sqrt{(}I)\cos(\phi_{I})}{\sqrt{I}\sin(\phi_{I})} \right)} \\
\end{array}$$After replacing Θ*~R~*(0) and Θ*~R~*(*π*/2) in [Equation (22)](#FD34){ref-type="disp-formula"}, the decision rule becomes: $$\arctan\left( \frac{\sqrt{C} - \sqrt{I}\cos(\phi_{I})}{\sqrt{I}\sin(\phi_{I})} \right)\underset{e}{\overset{s}{\gtrless}}\arctan\left( \frac{\sqrt{I}\sin(\phi_{I})}{\sqrt{C} + \sqrt{I}\cos(\phi_{I})} \right)$$by recalling that arctan(·) is monotonically increasing in \[−*π*/1 *π*/2\], [Equation (25)](#FD38){ref-type="disp-formula"} becomes: $$\frac{\sqrt{C} - \sqrt{I}\cos(\phi_{I})}{\sqrt{I}\sin(\phi_{I})}\underset{e}{\overset{s}{\gtrless}}\frac{\sqrt{I}\sin(\phi_{I})}{\sqrt{C} + \sqrt{I}\cos(\phi_{I})}$$to keep the trajectory of *R*(*ωt*) in the definition domain of arctan, we can restrict the definition domain of *ϕ~I~* as \[0,*π*/2\]. With this assumption, $\sqrt{C} + \sqrt{I}\cos(\phi_{I})$ and $\sqrt{I}\sin(\phi_{I})$ will be always positive. After some algebra, [Equation (26)](#FD39){ref-type="disp-formula"} becomes: $$C - I\cos^{2}(\phi_{I})\underset{e}{\overset{s}{\gtrless}}I\sin^{2}(\phi_{I})$$which gives: $$C\underset{e}{\overset{s}{\gtrless}}I$$
It can be showed that [(27)](#FD41){ref-type="disp-formula"} holds for the other possible domain definitions of *R*(*ωt*); therefore, *P~dif~* is: $$\begin{matrix}
{P_{dif} = 1} & {C < I} \\
\end{matrix}$$ $$\begin{matrix}
{P_{dif} = 0} & {C > I} \\
\end{matrix}$$By replacing [Equation (28)](#FD42){ref-type="disp-formula"} in [Equation (21)](#FD33){ref-type="disp-formula"}, the chip error probability becomes: $$\begin{array}{lll}
P_{c} & {= 1/2} & {C < I} \\
P_{c} & {= 0} & {C > I} \\
\end{array}$$
4.2.. Useful and Interfering Symbols are Asynchronous
-----------------------------------------------------
The condition in which interfering symbols can be asynchronous with respect to useful symbols increases the number of possible cases to be considered for the evaluation of Θ*~R~*(*ωt*). In particular, the interferer signal can change its turning direction at a random instance, *t~t~*, and the corresponding random angle is *ϕ~t~* = *ωt~t~*. Consequently, the received signal during a chip interval can be written as: $$R(\omega t) = \begin{cases}
{\sqrt{C}e^{j(b_{C}\omega t + \phi_{C})} + \sqrt{I}e^{j(b_{I_{1}}\omega t + \phi_{I})}} & {\omega t = \lbrack 0,\phi_{t}\rbrack} \\
{\sqrt{C}e^{j(b_{C}\omega t + \phi_{C})} + \sqrt{I}e^{j(b_{I_{2}}\omega t + d)}} & {\omega t = (\phi_{t},\pi/2\rbrack} \\
\end{cases}$$where *b~C~* defines the turning direction of the useful chip, while *b*~*I*~1~~ and *b*~*I*~2~~ are the turning directions of the sequential chips of the interferer. In [Equation (30)](#FD45){ref-type="disp-formula"}, at the angle *ϕ~t~*, the interferer starts transmitting another chip. We denote this asynchronous interferer case probability as *P~mix~*. The occurrences of *P~eq~, P~dif~* (which are given in Section 4.1) and *P~mix~* for the different combinations of *b~C~*, *b*~*I*~1~~ and *b*~*I*~2~~ are shown in [Table 2](#t2-sensors-14-05622){ref-type="table"}.
By considering these occurrences, the chip error probability can be evaluated as: $$P_{c} = 1/4P_{eq} + 1/4P_{dif} + 1/2P_{mix}$$
To obtain *P~mix~*, we consider the case of *b~C~* = +1, *b~I~*~1~ = +1, *b~I~*~2~ = −1, and *d* = *ϕ~I~* + 2*ϕ~t~*, which is mathematically identical to the all other *P~mix~* cases in [Table 2](#t2-sensors-14-05622){ref-type="table"}. Following the same approach followed in Section 4.1, we can write the decision rule as: $$\frac{\sqrt{C} - \sqrt{I}\cos(\phi_{I} + 2\phi_{t})}{\sqrt{I}\sin(\phi_{I} + 2\phi_{t})}\underset{e}{\overset{s}{\gtrless}}\frac{\sqrt{I}\sin(\phi_{I})}{\sqrt{C} + \sqrt{I}\cos(\phi_{I})}$$which can be simplified to: $$C + \sqrt{CI}(\cos(\phi_{I}) - \cos(\phi_{I} + 2\phi_{t})) - I\cos(2\phi_{t})\underset{e/s}{\overset{s/e}{\gtrless}}0$$where the sign in the inequality depends on the sign of denominators in [Equation (32)](#FD47){ref-type="disp-formula"}. [Equation (33)](#FD48){ref-type="disp-formula"} can be rearranged as: $$C/I + \sqrt{C/I}(\cos(\phi_{I}) - \cos(\phi_{I} + 2\phi_{t})) - \cos(2\phi_{t})\underset{e/s}{\overset{s/e}{\gtrless}}0$$
*P~mix~* is obtained by using the test in [Equation (34)](#FD49){ref-type="disp-formula"} considering the different signs of the denominators in [Equation (32)](#FD47){ref-type="disp-formula"} and the definition domain of the arctan function.
4.3.. Validation of the Analytical Model through Simulations
------------------------------------------------------------
Simulation results for the cases of synchronous and asynchronous useful and interfering symbols are shown in [Figure 6](#f6-sensors-14-05622){ref-type="fig"}. Since the analysis in Section 4 is exact, we expect a perfect agreement between simulations and analytical model. The results of [Figure 6](#f6-sensors-14-05622){ref-type="fig"} confirm this intuition.
5.. Probability of Data Error in IEEE 802.15.4
==============================================
The received path of the commercial 802.15.4 compatible transceivers, such as TICC2420 \[[@b29-sensors-14-05622]\] and Freescale MC13224 \[[@b30-sensors-14-05622]\], is composed of two main blocks: an O-QPSK demodulator and a de-spreader to obtain the data symbols from the received chips. In the previous section, we have obtained the chip error rate, *P~c~*, by assuming both coherent and non-coherent detection; this error probability is used here to derive bounds for the symbol error probability, *P~d~*, and for the corresponding packet reception rate, *P~r~*.
5.1.. Chip Error Rate to Data Symbol Error Rate
-----------------------------------------------
As discussed previously, in the 2.4 GHz PHY of IEEE 802.15.4, data symbols are mapped to the chips after the demodulation. Because of the symmetry in the data symbol set used by the 802.15.4 2.4 GHz PHY layer, the error probability of the data symbol, *S*~0~, is equal to the symbol error probability, *P~d~* \[[@b31-sensors-14-05622]\]: $$\begin{matrix}
{P_{d} = P(e \mid S_{0}) = P(e \mid S_{i})} & {\text{for}\ i = 1,\ldots,15} \\
\end{matrix}$$Hence, with no loss of generality to obtain an analytical expression for *P~d~*, we can find the union upper bound of *P*(*e*\|*S*~0~) expressed as: $$P(e \mid S_{0}) \leq \underset{i = 1}{\overset{15}{\text{∑}}}{P(S_{i} \mid S_{0})}$$where *P*(*S~i~*\|*S*~0~) represents the probability of deciding for the symbol, *S~i~*, once symbol *S*~0~ has been transmitted. Each probability in the summation of [Equation (36)](#FD51){ref-type="disp-formula"} can be formulated as a function of Hamming distance, *h~i~*, between *S*~0~ and *S~i~*, and the chip error rate, *P~c~*, as: $$P(S_{i} \mid S_{0}) = P(h_{i},P_{c}) = \begin{cases}
{\frac{1}{2}\left( \begin{array}{l}
h_{i} \\
{h_{i}/2} \\
\end{array} \right)P_{c}^{h_{i}/2}\left( {1 - P_{c}} \right)^{h_{i}/2} + \text{∑}_{i = \frac{h_{i} + 2}{2}}^{h_{i}}{\left( \begin{array}{l}
h_{i} \\
i \\
\end{array} \right)P_{c}^{i}\left( {1 - P_{c}} \right)^{h_{i}/i}}} & {\text{if}\ h_{i}\ \text{is\ even}} \\
{\text{∑}_{i = \frac{h_{i} + 1}{2}}^{h_{i}}{\left( \begin{array}{l}
h_{i} \\
i \\
\end{array} \right)P_{c}^{i}\left( {1 - P_{c}} \right)^{h_{i} - i}}} & {\text{if}\ h_{i}\ \text{is\ odd}} \\
\end{cases}$$
By substituting [Equation (37)](#FD52){ref-type="disp-formula"} in [Equation (36)](#FD51){ref-type="disp-formula"}, we obtain an upper bound that can be applied to any symmetric symbol set. The bound will be used in the next sub-sections to obtain expressions for the data symbol error probability of both coherent and non-coherent detection.
### 5.1.1.. Coherent Detection
In the case of coherent detection, we can calculate the Hamming distance of each chip sequence. In particular, the recurrences of the Hamming distances of the data symbols is given in [Table 3](#t3-sensors-14-05622){ref-type="table"}.
Using these recurrences, the symbol error probability can be bounded as follows: $$P_{d} \leq \underset{i = 1}{\overset{N}{\text{∑}}}{r_{i}P(h_{i},P_{c}) = 2P(12,P_{c}) + 2P(14,P_{c}) + 3P(16,P_{c}) + 2P(18,P_{c}) + 6P(20,P_{c})}$$where *P*(*h~i~, P~c~*) is taken from [Equation (37)](#FD52){ref-type="disp-formula"} and *h~i~* and *r~i~* are given in [Table 3](#t3-sensors-14-05622){ref-type="table"}. *i* is the index for columns, and *N* = 5 is the number of columns in the table.
Note that, for a fixed value of *P~c~*, the computational complexity of the previous upper bound is *O*(1).
### 5.1.2.. Non-Coherent Detection
In the non-coherent case, we can only identify the phase transitions during the chip intervals. For a 32-chip long sequence, there will be 31 phase transitions, as shown in [Table 4](#t4-sensors-14-05622){ref-type="table"}.
Where + means a *π*/2 increase in the phase and − means a *π*/2 decrease in the phase. We can think of the phase transition sequence of a data symbol in [Table 4](#t4-sensors-14-05622){ref-type="table"} as a vector in Hamming space and calculate the Hamming distances between these vectors. The recurrences of Hamming distances from a data symbol to the other data symbols in the set are given in [Table 5](#t5-sensors-14-05622){ref-type="table"}.
The union upper bound is obtained as: $$P_{d} \leq \underset{i = 1}{\overset{N}{\text{∑}}}{r_{i}P(h_{i},P_{c}) = 2P(13,P_{c}) + 2P(14,P_{c}) + 3P(15,P_{c}) + 3P(16,P_{c}) + 2P(17,P_{c}) + 2P(18,P_{c}) + P(31,P_{c})}$$where *h~i~* and *r~i~* are given in [Table 5](#t5-sensors-14-05622){ref-type="table"}. *i* is the index for columns, and *N* = 7 is the number of columns in the table.
Again, the computational complexity of the upper bound is *O*(1).
5.2.. Packet Reception Rate
---------------------------
In 802.15.4 2.4 GHz PHY, four data bits are mapped to a data symbol \[[@b2-sensors-14-05622]\]. When a packet of *b*-bytes is transmitted, the probability of successfully receiving the packet, *P~r~*, is calculated as: $$P_{r} = {({1 - P_{d}})}^{2b}$$
It is worth noting that the upper bounds on *P~d~* in [Equation (38)](#FD53){ref-type="disp-formula"} and [Equation (39)](#FD54){ref-type="disp-formula"} will become lower bounds in [Equation (40)](#FD55){ref-type="disp-formula"}. *P~r~* as a function of SIR is shown in [Figure 7](#f7-sensors-14-05622){ref-type="fig"} for the coherent and non-coherent demodulation schemes considered in Sections 3 and 4. Comparison between analysis and simulations is also shown. Results for the case of no thermal noise (cross-points for coherent and filled circles for non-coherent) have been obtained for a PHY service data unit (PSDU) of 14 bytes. The figure shows that the agreement between analysis and simulations is very good.
Experimental results shown in [Figure 7](#f7-sensors-14-05622){ref-type="fig"} of \[[@b19-sensors-14-05622]\] suggest 7 dB for the threshold of SNR, where the performance of the system, in terms of PRR, completely switches from zero to one. In this sense, the interval from *SNR* = 7 dB to *SNR* = ∞ should be adequate for representing a great variety of different SNR operation conditions, where the noise is not dominant, but the performance of the system is mostly determined by the interferer. This interval results in a considerably narrow region of PRR (the shaded area in [Figure 7](#f7-sensors-14-05622){ref-type="fig"}), and the experiment results in \[[@b16-sensors-14-05622],[@b18-sensors-14-05622]--[@b20-sensors-14-05622]\] fall into this region; therefore, it would not be wrong to conclude that the mathematical analysis presented here captures the essential features of the real-world performance, where the dominant factor of performance degradation is interference. Two different values of PSDU are considered in [Figure 7](#f7-sensors-14-05622){ref-type="fig"}: 14 bytes (experiments in \[[@b20-sensors-14-05622]\]) and 127 bytes (the maximum allowable, which is used in experiments in \[[@b16-sensors-14-05622],[@b18-sensors-14-05622],[@b19-sensors-14-05622]\]). Finally, It is also worth noting that 0 dB is the absolute minimum SIR necessary to successfully detect the useful packets in non-coherent demodulation.
6.. Conclusion
==============
The effect of concurrent transmission on the performance, in terms of chip, symbol and packet error probability, of IEEE 802.15.4 systems was investigated. To this aim, an analytical model, which starts from the PHY characterization of the IEEE 802.15.4 standard, was proposed, and the results were validated using simulations. Since the system performance does not depend on the number of interfering devices, but rather, on the overall amount of interfering power \[[@b18-sensors-14-05622],[@b20-sensors-14-05622]\], the case of one single interferer was considered. Both non-coherent and coherent demodulation were analyzed in the absence of thermal noise. The joint effect of thermal noise and interference was taken into account by means of simulations. For the case of non-coherent demodulation, we found that 0 dB is the absolute minimum necessary value of SIR for the successful detection of the useful packets. We have also shown that the proposed analysis is in agreement with the measurement results of the literature \[[@b16-sensors-14-05622],[@b18-sensors-14-05622]--[@b20-sensors-14-05622]\] under realistic working conditions. We conclude that the mathematical framework can be used to provide analytical explanations to the experimental results shown in \[[@b16-sensors-14-05622],[@b18-sensors-14-05622]--[@b20-sensors-14-05622]\].
The authors declare no conflicts of interest.
{#f1-sensors-14-05622}
{#f2-sensors-14-05622}
{#f3-sensors-14-05622}
{#f4-sensors-14-05622}
{#f5-sensors-14-05622}
{#f6-sensors-14-05622}
{#f7-sensors-14-05622}
######
Signal reception models in network simulators \[[@b24-sensors-14-05622]\]. SNRT, signal-to-noise ratio threshold; BER, bit error rate.
**Simulator** GloMoSim ns-2 OPNET
---------------------- ----------------------- ------------ -----------
**Signal Reception** SNRT-based, BER-based SNRT-based BER-based
######
Useful and Interferer Chip Phase Combinations.
***b****~C~* ***b****~I~***~1~** ***b****~I~***~2~** **d** **Probability of Error**
-------------- --------------------- --------------------- ------------------ --------------------------
+1 +1 +1 *ϕ~I~* *P~eq~* = 0
+1 +1 −1 *ϕ~I~* + 2*ϕ~t~* *P~mix~*
+1 −1 +1 *ϕ~I~* − 2*ϕ~t~* *P~mix~*
+1 −1 −1 *ϕ~I~* *P~dif~*
−1 +1 +1 *ϕ~I~* *P~dif~*
−1 +1 −1 *ϕ~I~* + 2*ϕ~t~* *P~mix~*
−1 −1 +1 *ϕ~I~* − 2*ϕ~t~* *P~mix~*
−1 −1 −1 *ϕ~I~* *P~eq~* = 0
######
Recurrences of unique Hamming distance values.
----------------------- ---- ---- ---- ---- ----
*H~I~* 12 14 16 18 20
Reoccurrence (*R~I~*) 2 2 3 2 6
----------------------- ---- ---- ---- ---- ----
######
Phase transitions of chip sequences in 2450-MHz PHY of IEEE 802.15.4.
**Data Symbol** **Data Bits** **Phase Transitions**
----------------- --------------- --------------------------------------------
*S*~0~ 0000 ++\-\-\-\-\--+++-++++-+-+++\--++-++\--
*S*~1~ 1000 +\--+++\-\-\-\-\--+++-++++-+-+++\--++-
*S*~2~ 0100 ++-++\--+++\-\-\-\-\--+++-++++-+-+++-
*S*~3~ 1100 ++\--++-++\--+++\-\-\-\-\--+++-++++-+-
*S*~4~ 0010 -+-+++\--++-++\--+++\-\-\-\-\--+++-+++
*S*~5~ 1010 ++++-+-+++\--++-++\--+++\-\-\-\-\--+++
*S*~6~ 0110 +++-++++-+-+++\--++-++\--+++\-\-\-\--
*S*~7~ 1110 \-\-\--+++-++++-+-+++\--++-++\--+++-
*S*~8~ 0001 \--++++++\-\--+\-\-\--+-+\-\--++\--+\--++
*S*~9~ 1001 -++\-\--++++++\-\--+\-\-\--+-+\-\--++\--+
*S*~10~ 0101 \--+\--++\-\--++++++\-\--+\-\-\--+-+\-\--+
*S*~11~ 1101 \--++\--+\--++\-\--++++++\-\--+\-\-\--+-+
*S*~12~ 0011 +-+\-\--++\--+\--++\-\--++++++\-\--+\-\--
*S*~13~ 1011 \-\-\--+-+\-\--++\--+\--++\-\--++++++\-\--
*S*~14~ 0111 \-\--+\-\-\--+-+\-\--++\--+\--++\-\--+++++
*S*~15~ 1111 ++++\-\--+\-\-\--+-+\-\--++\--+\--++\-\--+
######
Recurrences of unique Hamming distance values.
--------------------------- ---- ---- ---- ---- ---- ---- ----
*h~i~* 13 14 15 16 17 18 31
**reoccurrence** (*r~i~*) 2 2 3 3 2 2 1
--------------------------- ---- ---- ---- ---- ---- ---- ----
|
In the interview, Cardinal Burke addresses why a Catholic employer cannot, and should not be expected to, provide health care insurance that includes contraception and sterilization. He explains how this would be not only material cooperation in the sin, but also formal cooperation and for that reason there is no way to justify this implementation.
|
Beneficial effects of fenofibrate in pulmonary hypertension in rats.
Pulmonary hypertension (PH) is a morbid complication of cardiopulmonary as well as several systemic diseases in humans. It is rapidly progressive and fatal if left untreated. In the present study, we investigated the effect of PPARα agonist fenofibrate (FF) on monocrotaline (MCT)-induced PH in rats. FF, because of its pleiotropic property, could be helpful in reducing inflammation, oxidative stress, and reactive oxygen species. On day 1, MCT (50 mg/kg, s.c.) was given to all the rats in MCT, sildenafil, and FF group except normal control rats. After 3 days of giving MCT, sildenafil (175 µg/kg, orally) and FF (120 mg/kg, orally) were given for 25 days. Echocardiography, hemodynamic parameters, fulton's index, histopathology, oxidative stress parameters, inflammatory markers, Bcl2/Bax gene expression ratio in the right ventricle, and protein expression for NOX-1 in lungs were studied in all the groups. FF has shown to prevent decrease in ratio of pulmonary artery acceleration time to ejection time, increase in ratio of right ventricular outflow tract dimension to aortic outflow dimension, rise in right ventricular systolic pressure, right ventricular hypertrophy, increase in the percentage medial wall thickness (%MWT), increase in oxidative stress and inflammation, increase in NADPH oxidase-1 (NOX-1) expression, and decrease in mRNA expression of Bcl2/Bax ratio caused by MCT. To conclude, FF prevented MCT-induced PH in rats by various mechanisms. It might be helpful in preventing PH in patients who are likely to develop PH.
|
Introduction {#s0005}
============
As the COVID-19 pandemic evolves, is evident that around 6% of the patients will require ICU admission [@b0005], [@b0010], [@b0015], [@b0020]. Around 75% of those will need invasive ventilation [@b0020], and approximately 10% will require ventilation beyond 14 days [@b0025], [@b0030], [@b0035], [@b0040], [@b0045]. Undoubtedly, some patients will benefit from a tracheostomy during the weaning recovery phase. A tracheostomy is an aerosol generating procedure with a significant viral spread risk. Identifying who will benefit from it and developing safety procedure protocols requires clear selection criteria [@b0050].
Details around operating protocols have been simultaneously published by our team [@b0055] and an Italian group [@b0060]. The "CORONA-steps" [@b0060] and the "5Ts" [@b0055] cover the entire spectrum of a safe tracheostomy procedure.
Here we aim to share our outcomes in a cohort of COVID-19 patients that had surgical tracheostomies. We focus in selection criteria and outcomes, and share safety lessons-learned.
Methods {#s0010}
=======
Case selection/decision-making {#s0015}
------------------------------
Decisions were made on a case-by-case basis (communication between ICU-OMFS). Decision-making was based on acute and chronic co-morbidities such as acute kidney injury, obesity, anatomy, airway-related difficulties and ICU-related delirium/withdrawal. Prognosis (long-term, short-term) was also a decisive factor.
Most ICU patients were heavily sedated and dependent on benzodiazepines and long-acting opioid infusions; this increased the risk of sedation-related complications (withdrawal/delirium) during sedation holds and extubation attempts.
We developed selection criteria and summarise them based on an **'ABCD'** algorithm:**A *(Airway)***: Intubation for close to 14 days or more**B *(Breathing)*:** FiO2 \< 40%, PEEP below 15**C** ***(Circulation)*:** Apyrexial, cardiovascularly stable, reducing inflammatory markers (WBC:Neutrophil ratio, CRP)**D *(Disability)*:** Tracheostomy requirement for weaning
Two negative tests for COVID-19 were not mandatory. Whilst ideal, the potential for false negatives and false positives ("Positive" PCR from dead virus) makes results unpredictable [@b0065].
Post-tracheostomy decannulation criteria were:(1)48h minimum unsupported spontaneous breathing(2)No signs of infection reactivation for 48h(3)GCS \> 14(4)No signs of ongoing delirium(5)Verified safe upper airway access(6)Hemodynamic stability (no vasopressors/inotropes)
Our cohort consists of ten COVID-19 patients who underwent surgical tracheostomy in the weaning phase. Data were collected from case notes with appropriate institutional ethics.
Results {#s0020}
=======
Patients profile {#s0025}
----------------
All patients were male (average age 57.3) **(** [Table 1](#t0005){ref-type="table"} **)**. Lliterature supports male predominance, but reaching 100% was surprising [@b0070]. Nine patients had co-morbidities. Nine had a BMI greater than 30, (\>100 Kg, \<1.83 m). Eight had pre-existing hypertension and 5 had pre-existing diabetes [@b0075].Table 1COVID-19 patient's medical profile and follow-up.Case NoGenderAgePre- COVID-19 ComorbiditiesPost-ARDS Medical IssuesNo of days Intubated (ETT)Tracheostomy tube sizeNo of days post-tracheostomyWeaned off sedationWeaned off ventilatorDays to decannulationOutcome (Ward Step-down/Discharge)1M40HTN, High BMIRenal Failure199 (Adjustable)22YesYes17Ward2M76NilRenal Failure16922YesMinimal pressure support3M63HTN, High BMI11819YesYes9Discharge4M62HTN, High BMI, Atrial Fibrillation, Type II DM, Hypercholesterolemia129 (Adjustable)18YesYes9Ward5M54HTN, High BMI158.515YesYes7Discharge6M35Schizophrenia, Type II DM, High BMIRenal Failure169 (Adjustable)14YesYes12Ward7M49HTN, Type II DM, High BMI169 (Adjustable)10YesYes88M60HTN, Hypercholesterolemia, High BMIRenal Failure279 (Adjustable)8YesMinimal pressure supportWard9M71HTN, Type II DM, High BMI179 (Adjustable)9YesYes10M63HTN, Type II DM, High BMIRenal Failure239 (Adjustable)2YesNoN/AICU[^1]
Five patients developed renal failure/undergoing haemodialysis. All patients were intubated for a minimum of 11 days. Due to body habitus we used a size-9 adjustable flange tube in 7/10 patients. We aimed to minimise the risk of inadvertent decannulation. We had no incidents of dislodgement.
There were no significant intraoperative/immediate postoperative complications. Two patients experienced tracheostomy obstruction 72h post-procedure. Both were treated with change of inner cannula and bronchoscopy. One tube cuff deflated at day 8 post-op; this tube was changed uneventfully.
Patients were able to wean-off sedation within 24h. All patients required bridging with alternative sedatives (dexmedetomidine, clonidine). Common symptoms observed during the awakening phase were mainly down to sympathetic hyperactivity (hypertension, diaphoresis, tachycardia and tachypnoea). The tracheostomy provided a safe airway during these symptoms. Supplementary medications were effective, without compromising spontaneous breathing.
Overall, we observed the following benefits:(1)Reduction in ICU length of stay, releasing essential capacity(2)Reduction in prolonged use of sedatives/analgesics(3)Earlier spontaneous breathing(4)Better bronchial toilet; less traumatic suctioning(5)Faster delirium resolution(6)Faster rehabilitation/physiotherapy(7)More efficient use of nursing resources
Currently, 6 (60%) patients have been decannulated and stepped down on ward. Patients' profile and outcomes are summarized in [Table 1](#t0005){ref-type="table"}. In the context of a 12-bed ICU, this is a significant number**.**
Procedural pitfalls {#s0030}
-------------------
After each procedure, the team would debrief and reflect. An action plan was introduced to prevent recurring issues **(** [Table 2](#t0010){ref-type="table"} **)**. We aimed to identify human factors contributing towards safety pitfalls. The surgical team remained relatively constant but there was a considerable variation in the anaesthetic/scrub staff. This lack of continuity reinforced the need for a robust SOP and good communication.Table 2Procedural safety pitfalls, solutions sought and lessons learned for future prevention.Case NoSafety pitfallImpact of errorSolution soughtLesson learned1NoneN/AN/AN/A2Early patient transfer to theatreSurgical team not donnedSurgical team scrubbed in the anaesthetic RoomImprove communication with anaesthetic/transfer team3Malfunctioning inner radioImpaired communication with outer teamLoud voice/signsCheck radio prior to procedure4ET Tube advanced too far caudallySingle lung ventilationMeasure ET tube prior to proceedingDo not begin tracheostomy unless confirmation that ET tube is in appropriate position5NoneN/AN/AN/A6NoneN/AN/AN/A72 members of anaesthetic team to be at head end for ET tube manipulationLoss of fluency of ET tube manipulation at a critical pointMandatory 2 members of anaesthetic team to be at head end at time of ET tube manipulationBetter direction to anaesthetic team8ET tube balloon pierced. Pt had a history of previous tracheostomyHad to keep ventilator off and place tracheostomy tube immediatelyNumber 11 blade to be usedBroader blade used to create window. Use an 11 blade9NoneN/AN/AN/A10NoneN/AN/AN/A
We also noticed that doing these cases on a CEPOD list takes longer. A potential solution to streamline the process might be for ICU units to consider a designated area in ICU for performing surgical tracheostomies.
Personnel follow-up {#s0035}
-------------------
All personel used appropriate PPE [@b0055]. None of the staff involved developed COVID-19 symptoms post-operatively (Appendix). One member of the team self-isolated for 2 weeks as his wife tested positive for COVID-19. He subsequently tested negative. This endorses the safety of our protocol.
Discussion {#s0040}
==========
Our early experience with surgical tracheostomies in COVID-19 patients, suggests that this procedure has a positive effect on their outcome. 70% of our patients are no longer ventilator-dependant and 60% have beendecannulated. This releases valuable resources (ICU beds, staff, ventilators) to those that need them. Moreover, is a safe procedure both for patients and staff, if a well-considered SOP is followed.
Our patients are chosen in a multidisciplinary setting, utilising the best available evidence. These patients tend to have high BMI and various comorbidities. We recognise the expanding literature and we react accordingly adjusting our practice. It is important to remain adaptable in challenging times. Recognising the importance of human factors has significant benefits in providing safe/effective service.
Our study has limitations. The sample size is low, but it reflects a significant % of our ICU capacity, translating into effctive use of resources. In addition, our cohort represents all of the patients undergone the procedure (no exclusions). Our study has no control group or experimental setting, as this wouldn't be appropriate, but data were kept in a prospective, protocol-driven fashion. Lastly, we haven't considered a comparison with percutaneous tracheostomies, as this is now considered a procedure with higher AGM transmission risk. Plus, none of our patients would qualify due to anatomy/obesity.
Conclusion {#s0045}
==========
Surgical tracheostomy is an invasive procedure with potentially significant risks. Decision-making should be based on MDT consensus and with a protocol to get the maximum benefit whilst minimizing risk. Doing this in a carefully planned and executed manner with strict inclusion criteria has a positive effect for the patients and the team.
Funding {#s0050}
=======
None.
Declaration of Competing Interest
=================================
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
{#s0055}
See [Table 3](#t0015){ref-type="table"} .Table 3Staff Involved in relation to developing COVID-19 symptoms.TeamTotal NumberDeveloped SymptomsTested positiveNoteScrub Team (Scrub Nurse and Runner)1400Anaesthetics (Consultant, Trainee, Anaesthetic Nurse)2300Surgeons610Surgeon's wife developed symptoms (also a health care professional) prior to surgeon and she subsequently tested positive for COVID-19. (Likely contracted via different route.)
[^1]: HTN: Hypertension, BMI: Body Mass Index, DM: Diabetes Mellitus.
|
Everything That You Need to Know About .NET Core 2
The framework of Microsoft’s .NET platform originates back as far as the late 1990’s. Since then, it has undergone major changes and has become a frontrunner in how we develop software.
Allowing developers to create code using a variety of different languages, such as C# and C++, the language interoperability that .NET allows is due to the Framework Class Library and Common Language Runtime (CLR) components of the framework. The process involves transforming the source code, irrespective of what language it was written in, to a CIL code, known as a bytecode. Before .NET can process the code completely, the CLR transforms it into a native code that can be read universally across the .NET framework.
With a simplified assembly, the .NET framework is a firm favourite amongst those looking for .NET developer jobs in Ireland. Since the addition of .NET Core, code has been able to be shared across platforms with ease due to providing those who use it with a multi-platform library.
From August 14th, 2017, Microsoft’s latest update of their framework, .NET Core 2, has been available for download. With some slight changes since their last release, there are some functions that will change the way you use the platform. Below is a list of what you need to know and how it will impact you as a user:
Expanded API’s
Application Programming Interfaces (API) allows various elements of software to interact with one another. The routines, tools and protocols have the ability to define the way that a program has been developed and ensure that it has been written in the correct way for the specific operating system. Housed within a library class, API influences how the software will interact with external code so is an important element for developers to be aware of.
Previously, one of .NET Core’s limitations was the lack of third party libraries that were available for use. With only 13,000 previously available, the Microsoft-owned platform’s goal with their update was to make more of its API’s consistent across the various versions of .NET. Now, the newest release supports the use of over 32,000 API’s, making it much easier for you to post your code to a .NET standard library without the need to edit your code substantially.
In addition to this, users might appreciate the new compatibility shim that will enable apps using .NET core 2 to reference any framework library they so wish. With the increase of API’s, .NET Core 2 also supports .NET Standard 2. For those who are unfamiliar with .NET Standard, it acts as a central hub for API’s across a mixture of different environments, including desktop applications and games, amongst others.
OS Support
Due to .NET being a Microsoft platform, the initial release of Core saw a disconnect between operating systems. However, now .NET Core has started to see Mac OS, Windows and Linux as single operating systems. Users running an OS system no longer need to use OpenSSL due to changes in the crypto libraries, taking advantage of services that are natively available. This means that developers will find it easier to work on different operating systems such as desktops and mobile devices. For those who prefer Linux, it is also being treated as a single operating system, and the improvements will mean that there are no individual runtimes.
Razor Pages
An additional feature of the .NET framework is ASP.NET Core. Compatible with .NET Core 2, the open-source framework allows an easy build of web and cloud applications. In the latest release, one of the biggest features to be revealed is the introduction of razor pages. With its aim to make coding page-focused scenarios more productive, it simplifies dynamic HTML code and reduces the number of folders and files in the MVC structure.
|
Introduction {#Sec1}
============
Rupture of a vulnerable atherosclerotic plaque is an important underlying cause of clinical ischemic events, such as stroke \[[@CR1]\]. Therefore, visualization of vulnerable plaques may aid in the identification of patients who have an increased risk for a clinical event. Inflammatory cells play an important role during the development and progression of atherosclerosis \[[@CR2]\]. Within atherosclerotic plaques, activated macrophages have a high metabolic rate, inducing hypoxia which stimulates the formation of new microvessels originating from the outer layer of the vessel wall, the adventitia \[[@CR3], [@CR4]\]. These newly formed microvessels generally have impaired endothelial integrity, which can lead to extravasation of inflammatory cells and erythrocytes from the microvessel lumen into plaque tissue \[[@CR4]\]. Extravasation of erythrocytes is generally considered as an important contributing factor to intraplaque hemorrhage (IPH) \[[@CR5]\]. Because lipids constitute 40 % of the erythrocyte membrane \[[@CR5]\], extravasation of erythrocytes leads to increased cholesterol deposition in the plaque tissue, which in turn stimulates further recruitment of inflammatory cells. All the above biological events, especially leaky plaque microvasculature, are considered key features in plaque destabilization \[[@CR6]\].
The microvasculature in plaques are very small (up to ∼100 μm in diameter) but can be studied non-invasively by several imaging modalities, including contrast-enhanced ultrasound (CEUS) \[[@CR7], [@CR8]\], positron emission tomography (PET) \[[@CR9]--[@CR11]\], and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) \[[@CR12]\]. MRI is a well-established imaging modality that can be used to visualize the main plaque components: areas of IPH, the lipid-rich necrotic core, and fibrous cap status \[[@CR13]--[@CR15]\]. Early studies developed MRI for the detection of morphological and chemical components by studying specimen from surgery (carotid endarterectomy (CEA)). These ex vivo specimens were advantageous for testing and developing MRI sequences, but the lack of blood precludes studying of the dynamics from DCE \[[@CR16], [@CR17]\].
In recent years, a number of studies have applied DCE-MRI to study atherosclerotic plaque microvasculature. The present paper reviews the current state and future potential of DCE-MRI in the evaluation of plaque microvasculature with applications in animals and patients. First, because the methods of DCE-MRI are now well-developed and widely applied but are not familiar to a general audience, we begin with principles and acquisition methods of DCE-MRI and methods for (semi)quantitative analysis of DCE-MRI data. Second, an overview is given of publications on DCE-MRI of plaque microvasculature (Table [1](#Tab1){ref-type="table"}) used to study one of the following aspects: (1) associations between plaque microvasculature and other plaque features, (2) longitudinal changes in plaque microvasculature, (3) comparison of different animal groups and human subjects with a different cardiovascular risk profile, and (4) evaluation of therapy response. Finally, future challenges and potential for DCE-MRI to study plaque microvasculature will be discussed.Table 1Overview of DCE-MRI studies of atherosclerotic plaque microvasculature. Overview of studies investigating the atherosclerotic plaque microvasculature using dynamic contrast-enhanced MRI: subjects (human or rabbits), analysis method (quantitative or semi-quantitative), main study purpose, and study outcome are shownReferenceSubjectsMain study purposeMain study outcomeChen et al. \[[@CR18]\]Patients with CVD (AIM-HIGH Trial \[[@CR19]\]Scan-rescan reproducibilityModerate reproducibility for K^trans^ (Patlak) with a 25 % coefficient of variation. To limit dropout, intensive operator training, optimized imaging, and quality control is requiredKerwin et al. \[[@CR20]\]CEA patientsMethod developmentDevelopment of a motion correcting and noise reducing algorithm for the analysis of DCE-MRI of carotid arteriesKerwin et al. \[[@CR21]\]Patients with a carotid lesion ≥ AHA type IVMethod comparisonQuantitative enhancement characteristics, such as K^trans^ (Patlak), depend on the used contrast medium (gadobenate dimeglumine vs gadodiamide)Ramachandran et al. \[[@CR22]\]Humans with CVD riskMethod developmentDevelopment of a registration method for alignment of different time frames of DCE-MRI of carotid arteriesChen et al. \[[@CR23]\]Humans with advanced carotid diseaseMethod developmentExtended graphical model exhibits a reduced bias in K^trans^ estimation compared to the Patlak modelVan Hoof et al. \[[@CR24]\]Symptomatic patients (30--99 % carotid stenosis)Method comparisonComparison between phase- and magnitude-based vascular input functions and resulting effect on pharmacokinetic parameters. No signal saturation due to blood flow for phase-based determined vascular input functionCalcagno et al. \[[@CR25]\]Humans with CVD riskMethod developmentDemonstration of feasibility of simultaneous VIF and vessel wall imaging (extended Tofts)Wan et al \[[@CR26]\]NZW Rabbit^a,b^Method developmentSpatio-temporal texture based features (like AUC) are able to distinguish between vulnerable and stable plaques.Calcagno et al. \[[@CR27]\]NZW Rabbit^c^Method comparisonExcellent reproducibility of DCE-MRI derived AUC (interscan, intraobserver, and interobserver ICCs \> 0.75, *P* \< 0.001)Wu et al. \[[@CR28]••\]NZW Rabbit1Method developmentDemonstration of feasibility of simultaneous VIF and vessel wall imaging with accurate estimation of pharmacokinetic parameters (Patlak)Calcagno et al. \[[@CR29]\]NZW Rabbit3Histological validationPositive correlation (*ρ* = 0.89, *p* = 0.016) between AUC and amount of neovessels in the intimaCalcagno et al. \[[@CR30]•\]NZW Rabbit3Histological validation3D DCE-MRI (AUC (*ρ* = 0.45) and K^trans^ (Patlak) (*ρ* = 0.38)) is able to quantify microvascular permeability in the entire abdominal aorta plaqueChen et al. \[[@CR31]\]NZW Rabbit^d^To study plaque progressionDCE-MRI (AUC) is able to quantitatively assess temporal changes of atherosclerotic plaques over a period of 3 monthsKim et al. \[[@CR32]\]NZW Rabbit3Validation of a chip for the development of nanomedicinesIncreased AUC for atherosclerotic animals compared to control animals. Lipid-polymer hybrid nanoparticle translocation is correlated with AUC (*ρ* = 0.79, *p* \< 0.0001)Lobatto et al. \[[@CR33]\]NZW Rabbit3Evaluation of glucocorticoid treatment for atherosclerosisDCE-MRI (AUC) reveals early changes in plaque microvascular permeability after liposomal glucocorticoid treatmentVucic et al. \[[@CR34]\]NZW Rabbit3Evaluation of pioglitazone treatment for atherosclerosisDCE-MRI (AUC) can demonstrate the anti-inflammatory effect of pioglizatone on atherosclerotic plaquesVucic et al. \[[@CR35]\]NZW Rabbit3Evaluation of LXR agonist R211945 treatment for atherosclerosisDCE-MRI (AUC) showed a trend towards a decreased microvasculature after treatment with atorvastatinChen et al. \[[@CR36]\]Patients with \>50 % carotid stenosisComparison of plaque ComponentsK^trans^ and v~p~ (Patlak) differed significantly between plaque components (lipid core, IPH, calcifications, loose matrix, and fibrous tissue), except between calcifications and IPH.Calcagno et al. \[[@CR37]\]Patients with CHD or CHD risk equivalentCorrelation with ^18^F-FDG PET-CTWeak, inverse relationship between inflammation (^18^F-FDG PET-CT, mean TBR) and plaque perfusion (DCE-MRI, K^trans^ (extended TK))Dong et al. \[[@CR38]\]Humans (carotid plaque thickness ≥2 mm)Evaluation of intensive lipid therapy in the treatment of atherosclerosisIntensive lipid therapy (using atorvastatin, niacin, and colesevelam) results in a reduction in K^trans^ (Patlak) after one yearGaens et al. \[[@CR39]\]Symptomatic patients (30--99 % carotid stenosis)Pharmacokinetic model comparisonThe Patlak model is the most suited quantitative model for description of carotid plaque microvasculatureKerwin et al. \[[@CR40]\]CEA PatientsValidation against microvasculature on histologyStrong correlation (*ρ* = 0.80, *p* \< 0.001) between DCE-MRI and histological measured fractional vascular areasKerwin et al. \[[@CR41]\]CEA PatientsValidation against microvasculature and inflammation on histologyK^trans^ (Patlak) is a quantitative and non-invasive marker of plaque inflammation (*ρ* = 0.75, *p* \< 0.001) and microvasculature (*ρ* = 0.71, *p* \< 0.001)Kerwin et al. \[[@CR42]\]CEA PatientsValidation against microvasculature and inflammation on histologyAdventitial K^trans^ (Patlak) was significantly correlated with the amount of microvasculature (*ρ* = 0.41, *p* = 0.04) and macrophages (*ρ* = 0.49, *p* = 0.01)Mani et al. \[[@CR43]\]Humans with and without exposure to particle matterRisk stratificationHigh exposure to particle matter may be associated with plaque neovascularization, measured with DCE-MRI (AUC)O'Brien et al. \[[@CR44]••\]Patients with CVD (AIM-HIGH Trial) \[[@CR19]\]Association of DCE-MRI with statin therapyShorter duration of statin therapy before occurrence of clinical event is associated with increased v~p~ (Patlak)Sun et al. \[[@CR45]•\]Symptomatic patients (ischemic event \<6 m)Correlation between DCE-MRI (K^trans^) and presence of IPHPresence of IPH was associated with an increase of 28 % of adventitial K^trans^ (Patlak)Truijman et al. \[[@CR46]\]Symptomatic patients (30--69 % carotid stenosis)Correlation with ^18^F-FDG PET-CTWeak, positive relationship between inflammation (^18^F-FDG PET-CT, TBR) and plaque perfusion (DCE-MRI, K^trans^ (Patlak))Wang et al. \[[@CR47]•\]Human (carotid plaque thickness ≥2 mm)Correlation with ^18^F-FDG PET-CTCorrelation between ^18^F-FDG PET (TBR) and DCE-MRI (K^trans^, Patlak) measurements varied with clinical conditions (symptomatic status)*DCE-MRI* dynamic contrast-enhanced MRI, ^*18*^ *F-FDG* ^18^fluorine-fluorodeoxyglucose, *PET*-*CT* positron emission tomography/computed tomography, *AUC* area under the curve, *NIRF* near-infrared fluorescence, *CVD* cardiovascular disease, *CEA* carotid endarterectomy, *CHD* coronary heart disease, *TBR* target-to-background ratio, *NZW* New Zealand White^a^Atherosclerosis was induced by a balloon injury of the aorta in combination with a high cholesterol-enriched diet (1.0 %)^b^Pharmacologic triggering was performed to stimulate plaque disruption^c^Atherosclerosis was induced by a balloon injury of the aorta in combination with a low cholesterol enriched diet (\<1.0 %) combined with palm oil^d^Atherosclerosis was induced by a balloon injury of the aorta in combination with a low cholesterol enriched diet (\<1.0 %)
DCE-MRI Methods to Study Plaque Microvasculature {#Sec2}
================================================
Principles of DCE-MRI {#Sec3}
---------------------
All DCE-MRI experiments require serial acquisition of MR images acquired in a brief time interval to study atherosclerotic plaque microvasculature (Fig. [1](#Fig1){ref-type="fig"}). After acquisition of anatomical references (Fig. [1a](#Fig1){ref-type="fig"}), the first images of the series, acquired before contrast injection, are used to determine baseline signal intensity of the atherosclerotic plaque tissue. Bolus injection of a low molecular weight non-specific Gadolinium-based contrast medium follows, and image acquisition is continued for several minutes. During this period, the bolus of contrast medium will be distributed, resulting in signal enhancement of the blood vessel lumen, vessel wall due to leakage of the contrast medium through damaged endothelial, and other tissues, such as skeletal muscle (Fig. [1b](#Fig1){ref-type="fig"}). In this image, the vessel lumen (circle) appears bright. A ring of enhancement in the outer (adventitial layer) part of the vessel wall (indicated by white arrows) can be clearly observed. The signal enhancement in the vessel wall depends on flow, microvascular density, the ability of the contrast medium to leak from the microvasculature into the extravascular extracellular space, and reflux. After analysis of the DCE-MR images, parametric maps (Fig. [1c](#Fig1){ref-type="fig"}) of the resulting parameter can be generated, indicating local leaky plaque microvasculature. In DCE-MRI studies of atherosclerosis to date, linear or cyclic Gadolinium-based contrast media have been used.Fig. 1MR images (**a**--**c**) of a transverse section of the carotid plaque in the internal carotid artery from a 64-year-old man. In **a**, A *black* blood T1-weighted turbo spin echo MR image as an anatomical reference. In this image, the vessel lumen (*circle*) appears in *black*. The atherosclerotic plaque of this patient appears hyperintense compared to the sternocleidoid muscle (*diamond*). In **b**, a three-dimensional T1-weighted fast field-echo dynamic contrast-enhanced MR image that is acquired 6 min after contrast injection is shown. In this image, the vessel lumen (*circle*) appears bright compared to the atherosclerotic plaque and surrounding tissues. A ring of enhancement can be observed at the outer part of the vessel wall (indicated by *white arrows*), which is attributed to the microvasculature originating from the adventitia. Finally, in **c**, a parametric K^trans^ map is overlaid on DCE-MRI image shown in **b**. In this parametric map, voxel wise determined K^trans^ values are color encoded from 0 to 0.2 min^−1^. Within this overlay, the lipid-rich necrotic core in the center of the plaque, exhibits low K^trans^ values (*dark*), while the highly vascularized adventitia (high K^trans^ values) at the outer rim (indicated by the *arrows*) is clearly visualized (*red regions*). *Circle*, internal carotid artery; *star*, external carotid artery; *diamond*, sternocleidoid muscle. Figure adapted from Truijman et al. \[[@CR46]\]
DCE-MRI studies of brain and tumor perfusion mostly use a contrast medium injection rate of 2 ml/s (typically 0.1 mmol/kg). Such fast injection rates result in quick passage of the bolus through the vessel and a high-contrast medium peak concentration, necessitating a higher temporal resolution for MR acquisition and compromising spatial resolution. For the evaluation of carotid atherosclerosis using DCE-MRI, however, high spatial resolution is required for accurate visualization of the vessel wall. Therefore, some DCE-MRI studies \[[@CR24], [@CR39]\] have used a slower injection rate of 0.5 ml/s in plaque imaging. Previous research has shown that a high injection rate is most beneficial for high K^trans^ values (\>0.2 min^−1^) \[[@CR48]\]. Typically, within the atherosclerotic lesion, mean K^trans^ values below 0.15 are reported \[[@CR24]\], and therefore, a lower injection rate may be applied.
Signal enhancement-time curves of DCE-MR images can be analyzed voxelwise or using a region-of-interest. Especially in the voxelwise analysis, movement of the subject during acquisition of the different DCE-MRI sequence time frames may pose a problem. A solution is to manually shift individual time frames to correctly align the images, or alternatively, to use post-processing methods for automated movement correction and noise reduction \[[@CR20], [@CR22]\].
Pulse Sequences for DCE-MRI of Plaque Microvasculature {#Sec4}
------------------------------------------------------
Currently, two main categories of pulse sequences for DCE-MRI of atherosclerotic plaque microvasculature are employed: "bright blood" or "black blood". Black blood imaging facilitates improved delineation of the inner vessel wall, whereas bright blood imaging enables to determine the CM concentration in the vessel lumen for each patient individually. Because the luminal CM concentration cannot be quantified accurately, quantitative analysis of black blood DCE-MRI with pharmacokinetic models can only be performed using a reference region model \[[@CR49]\] or previously determined generalized input functions \[[@CR50]\]. Recently, dedicated imaging methods have been proposed combining bright and black blood images in an interleaved fashion, allowing improved delineation of the vessel wall from black blood images as well as extraction of vascular input function based on lumen signal intensity from bright blood images \[[@CR25], [@CR28]••\].
A compromise between the desired spatial and the required temporal resolution must be made regardless of the imaging method used. Current studies (both in rabbits and patients) employed an in-plane spatial acquisition resolution of approximately 0.5 × 0.5 mm^2^. The preclinical rabbit studies have employed a temporal resolution of 5 s for 2D acquisition techniques and lower temporal resolution (30 s) for 3D techniques. In patient studies, the temporal resolution ranges from 15 to 30 s per time frame.
Semi-Quantitative Assessment of the Microvasculature {#Sec5}
----------------------------------------------------
The microvasculature can be assessed semi-quantitatively using the area-under-the-curve (AUC) of the (relative) signal enhancement curve. This requires that start- and end-time points are selected over which the AUC will be calculated. Generally, the moment of contrast arrival in the tissue of interest is chosen as the starting point, and the end time point is chosen empirically. It must be noted that when the end point is chosen relatively close to the contrast injection, the AUC reflects early contrast arrival, whereas a later end point will cause the AUC to reflect total leakage (and entrapment) of contrast medium in the plaque tissue.
The main advantage of semi-quantitative analyses is the relatively easy implementation. However, the information is limited because there is no direct relationship between the AUC and (patho)physiological parameters. Although research in the field of oncology \[[@CR51]\] has shown that the AUC reflects pathophysiology, it does so non-specifically, meaning that one particular AUC value can indicate a number of biological properties. Thus, an increased AUC can indicate increased leakage of the contrast medium from the microvasculature, increased density of microvessels, increased flow through the microvasculature, a decrease in reflux from the extracellular extravascular space to the microvasculature, or a combination of these. Therefore, changes or differences in the AUC may result from a variety of phenomena so that it may be difficult to attribute these changes to a single, underlying physiological cause. Similarly, effects of therapeutic interventions may potentially be obscured using the AUC. Another drawback of semi-quantitative analysis is the difficulty of direct comparison of results between studies because the AUC also depend on settings of the MR system, such as receiver gain.
### Validation of Semi-Quantitative DCE-MRI Parameters {#Sec6}
Validation of semi-quantitative DCE-MRI was performed in several balloon injured cholesterol-fed New Zealand White rabbit studies. It was found that the AUC positively correlated with microvessel count in the intima of histological specimens (Pearson's *ρ* of 0.89 (*p* = 0.016) and 0.91 (*p* = 0.011) for the AUC 2 and 7 min after contrast injection, respectively) \[[@CR29]\]. Furthermore, later research \[[@CR27]\] showed a good interscan and excellent intra- and inter-observer reproducibility (all ICCs \> 0.75, *p* \< 0.01).
Another atherosclerotic rabbit study compared two three-dimensional (3D) high spatial resolution DCE-MRI sequences (3D turbo field echo (TFE) with motion-sensitized-driven equilibrium (MSDE) preparation and a 3D turbo spin echo (TSE) sequence) \[[@CR30]•\]. A moderate Pearson correlation was found between AUC and ex vivo permeability measurements using Evans Blue (an albumin-binding dye used for quantification of ex vivo vascular permeability) near-infrared fluorescence (NIRF) (*ρ* = 0.45 for 3D TFE MRI and *ρ* = 0.39 for 3D TSE MRI). In addition, a fourfold improvement of temporal resolution was achieved when using compressed sensing by retrospective undersampling and reconstruction. In another study, comparison between in vivo (3D DCE-MRI) and ex vivo (Cy7-labeled near-infrared fluorescence \[NIRF\]) measures of microvascular permeability in the aortic wall of atherosclerotic rabbits showed a high degree of correlation between both imaging modalities (*r*^2^ = 0.65, *p* \< 0.0001) \[[@CR32]\].
These studies \[[@CR27], [@CR29], [@CR30]•, [@CR32]\] have demonstrated reproducible representation of plaque microvasculature through semi-quantitative DCE-MRI parameters.
Quantitative Assessment of the Microvasculature {#Sec7}
-----------------------------------------------
### Pharmacokinetic Modeling {#Sec8}
Pharmacokinetic modeling allows the quantification of contrast medium distribution over a tissue of interest with the main advantage of deriving parameters of the in vivo physical quantities of the amount, flow, and leakiness of the microvasculature.
A number of quantitative DCE-MRI data analysis models have been applied in the evaluation of atherosclerotic plaque microvasculature (Table [2](#Tab2){ref-type="table"}). These models describe the relationship between the concentration of the (extracellular) contrast medium in the blood plasma (C~p~) and the extracellular extravascular space (C~e~) according to the two-compartment model and using the parameters K^trans^, v~e~, and v~p~. K^trans^, the transfer constant of contrast medium from plasma to the tissue compartment, serves as an indicator of blood supply and vessel permeability within the atherosclerotic tissue. The parameters v~e~ and v~p~ represent the extravascular extracellular space and the plasma fractional volume, respectively. A schematic representation of the physiological meaning of the parameters is shown in Fig. [2](#Fig2){ref-type="fig"}.Table 2Overview of quantitative DCE-MRI models used in the analysis of atherosclerosis. Quantitative pharmacokinetic models used for the analysis of atherosclerosis based on the two-compartment model. The modified/extended Tofts and Kermode model is the analytical solution for the two-compartment model. The extended graphical model is based on a second order Taylor expansion of the modified/extended Tofts and Kermode modelMathematical descriptionParameters Two-compartment model$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ \begin{array}{l}\frac{{\mathrm{dC}}_{\mathrm{e}}\left(\mathrm{t}\right)}{\mathrm{dt}}=\frac{{\mathrm{K}}^{\mathrm{t}\mathrm{rans}}}{{\mathrm{V}}_{\mathrm{e}}}\left({\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right)-{\mathrm{C}}_{\mathrm{e}}\left(\mathrm{t}\right)\right)\hfill \\ {}{\mathrm{C}}_{\mathrm{t}}\left(\mathrm{t}\right)={\mathrm{v}}_{\mathrm{p}}{\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right)+{\mathrm{v}}_{\mathrm{e}}{\mathrm{C}}_{\mathrm{e}}\left(\mathrm{t}\right)\hfill \end{array} $$\end{document}$Mathematical descriptionK^trans^v~e~v~p~ Modified/extended Tofts and Kermode (TK)$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ {\mathrm{C}}_{\mathrm{t}}\left(\mathrm{t}\right)={\mathrm{v}}_{\mathrm{p}}{\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right)+{\mathrm{K}}^{\mathrm{t}\mathrm{rans}}{\displaystyle \underset{0}{\overset{\mathrm{t}}{\int }}{\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right){\mathrm{e}}^{-\frac{{\mathrm{K}}^{\mathrm{t}\mathrm{rans}}}{{\mathrm{v}}_{\mathrm{e}}}\left(\mathrm{t}-\uptau \right)}\mathrm{d}\uptau} $$\end{document}$XXX Tofts and Kermode$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ {\mathrm{C}}_{\mathrm{t}}\left(\mathrm{t}\right)={\mathrm{K}}^{\mathrm{t}\mathrm{rans}}{\displaystyle \underset{0}{\overset{\mathrm{t}}{\int }}{\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right){\mathrm{e}}^{-\frac{{\mathrm{K}}^{\mathrm{t}\mathrm{rans}}}{{\mathrm{v}}_{\mathrm{e}}}\left(\mathrm{t}-\uptau \right)}\mathrm{d}\uptau} $$\end{document}$XX Patlak$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ {\mathrm{C}}_{\mathrm{t}}\left(\mathrm{t}\right)={\mathrm{v}}_{\mathrm{p}}{\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right)+{\mathrm{K}}^{\mathrm{t}\mathrm{rans}}{\displaystyle \underset{0}{\overset{\mathrm{t}}{\int }}{\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right)\mathrm{d}\uptau} $$\end{document}$XX Extended Graphical Model$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$ {\mathrm{C}}_{\mathrm{t}}\left(\mathrm{t}\right)={\mathrm{v}}_{\mathrm{p}}{\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right)+{\mathrm{K}}^{\mathrm{t}\mathrm{rans}}{\displaystyle \underset{0}{\overset{\mathrm{t}}{\int }}{\mathrm{C}}_{\mathrm{p}}\left(\mathrm{t}\right)\mathrm{d}\uptau -\frac{{\mathrm{K}}^{{\mathrm{t}\mathrm{rans}}^2}}{{\mathrm{v}}_{\mathrm{e}}}}{\displaystyle \underset{0}{\overset{\mathrm{t}}{\int }}{\displaystyle \underset{0}{\overset{\uptau_1}{\int }}{\mathrm{C}}_{\mathrm{p}}\left({\uptau}_2\right){\mathrm{d}\uptau}_2{\mathrm{d}\uptau}_1}} $$\end{document}$XXXFig. 2Schematic representation of parameters used in pharmacokinetic models for analysis of atherosclerotic plaque microvasculature. Within a single region of interest or voxel, the fractional blood volume (microvasculature) is represented by v~p~, while the fraction of the extracellular extravascular space is represented by v~e~. Contrast medium transfer rate from the microvasculature to the extracellular extravascular space is given by K^trans;^ the reflux is described by K^trans^/v~e~. In most DCE-MRI studies, an extracellular contrast medium with a low molecular weight is used. For quantitative data analysis, therefore, a two-compartment model can be used (i.e., vascular and extracellular extravascular compartments). Based on this general concept and setting various assumptions, several different quantitative models can be derived. An overview of these models is presented in Table [2](#Tab2){ref-type="table"}
The modified/extended Tofts and Kermode (TK) model \[[@CR52], [@CR53]\] is a commonly employed analytical solution for the two-compartment model \[[@CR54], [@CR55]\], estimating all three pharmacokinetic parameters (K^trans^, v~e~, and v~p~). The original TK model, which was proposed for the study of multiple sclerosis \[[@CR56]\], does not take vascular contribution into account (i.e., v~p~ is assumed to be negligible). The Patlak model \[[@CR57]\] assumes that reflux, i.e., transfer of contrast medium from the tissue compartment back to the blood plasma (K^trans^/v~e~), is negligible. Recently, an approximation of the modified TK model has been introduced as an intermediate solution between the modified TK and the Patlak model: the extended graphical model \[[@CR23]\]. This model uses the first-order term of a Taylor series from the modified TK model to estimate v~e~.
### Vascular Input Function {#Sec9}
One essential requirement for quantitative analysis of DCE-MRI data is knowledge of the contrast medium concentration in the blood vessel over time, commonly, referred to as the arterial or vascular input function (AIF/VIF). Two main features of the VIF are a high relative peak concentration and a short bolus passage compared to other tissues. Accurate determination of the VIF requires a relatively high temporal resolution, which usually results in compromise with regard to the spatial resolution that can be achieved.
Two strategies can be employed for the determination of VIF. The first strategy is based on the assumption that VIF is similar in all subjects and a generalized population-averaged VIF, obtained from literature or determined in a cohort is used \[[@CR24], [@CR39], [@CR46], [@CR50]\]. An advantage of this method is that data acquisition and analysis requirements are simplified \[[@CR49]\]. The second strategy involves measurement of patient-specific function, giving the potential advantage of accounting for variations between subjects \[[@CR58]\]. Previous research in oncology found comparable results using either method, and the use of population-averaged vascular input functions resulted in increased \[[@CR50]\] or comparable \[[@CR59]\] reproducibility. In clinical studies of atherosclerotic plaque microvasculature, a generalized VIF is the most commonly chosen method, probably because of the required spatial resolution for accurate imaging of atherosclerotic plaque in the carotid artery. The generalized VIF can be obtained from a separate study cohort where acquisition is performed with a higher temporal resolution and a lower spatial resolution.
The VIF with MRI can be calculated by two different methods. The first method uses the magnitude of the acquired MR signal and is based on conversion of the relative signal enhancement to contrast medium concentration using the Ernst equation \[[@CR60]\]. For this conversion, blood relaxation and contrast medium relaxivity rates are taken into account. A second method based on MR signal phase has been developed more recently \[[@CR61], [@CR62]\]. First used in brain perfusion studies with dynamic susceptibility MRI \[[@CR63]\], the technique is increasingly used in DCE-MRI \[[@CR24], [@CR64]\]. Efforts have been made to compare the magnitude- and phase-based techniques \[[@CR24], [@CR59], [@CR65]--[@CR67]\], showing a strong potential for the phase-based technique, allowing accurate VIF quantification. In a DCE-MRI study of 17 symptomatic patients with a mild to severe carotid stenosis, it was found that the magnitude-based VIF resulted in a strong underestimation of lumen contrast medium concentration as compared to the phase-based VIF \[[@CR24]\]. Simulations and phantom experiments showed that this underestimation is caused by local blood flow velocity, which leads to saturation of the magnitude MR signal caused by the shortened T~1~ relaxation time in the presence of contrast medium. Analysis of K^trans^ values using population-averaged input functions showed a strong positive correlation between the two methods, although absolute values significantly differed.
### Validation of Quantitative DCE-MRI Parameters {#Sec10}
Histological validation of carotid plaque DCE-MRI has been carried out using reference specimens from patients after carotid endarterectomy (CEA). However, the drawback of all such validation studies is that these are performed in patients scheduled for CEA. Large randomized trials have shown that symptomatic patients with severe ipsilateral stenosis benefit the most from CEA \[[@CR68]\]. This population is more likely to have developed advanced atherosclerotic plaques. In addition, the surgeon removes the intima and part of the media of the vessel wall and the adventitia, from which microvasculature originates \[[@CR6]\], is missing in the CEA specimen. An additional limitation of the comparison of in vivo MRI with histological measurements as a reference standard is the comparison of a thin histological slice to thicker MR imaging slice (typically 2 mm). Due to the heterogeneous nature of atherosclerotic lesions, this may result in partial volume effects.
Despite these drawbacks, a strong and positive correlation between fractional blood volume derived from in vivo MRI and post-surgical histology (0.80, *p* \< 0.001) was found in 16 CEA patients \[[@CR40]\]. In addition, a significant Pearson correlation was reported between the transfer constants K^trans^ calculated from in vivo DCE-MRI with postsurgical histologic measurements of the microvessel area (*ρ* = 0.71, *p* \< 0.001 for the entire vessel wall K^trans^ and *ρ* = 0.41, *p* \< 0.04 for adventitial K^trans^). Additionally, an association between K^trans^ and other postsurgical histological parameters was reported, i.e., macrophage density (*ρ* = 0.75, *p* \< 0.001 for the vessel wall K^trans^ and *ρ* = 0.49, *p* \< 0.01 for adventitial K^trans^), loose matrix area (*ρ* = 0.50, *p* = 0.01, for vessel wall K^trans^) \[[@CR41], [@CR42]\]. It was also shown that K^trans^ and v~p~ differed significantly between different plaque components (lipid core, IPH, calcifications, loose matrix, and fibrous tissue), except between calcifications and IPH \[[@CR36]\].
Reproducibility, fit error, parameter uncertainty, and correlation with histology of carotid plaque DCE-MRI were compared for four pharmacokinetic models in patients with mild to severe carotid stenosis \[[@CR39]\]. Analysis of 43 patients showed the highest relative fit error for the Tofts model, while the other three models did not differ in this regard. The Patlak model had a significant lower parameter uncertainty for K^trans^ as compared to the other models. Reproducibility was studied in 16 asymptomatic patients with 30--69 % carotid stenosis who underwent imaging twice with several (4.3 ± 2.8) days between the two examinations. Results showed a good reproducibility for all considered pharmacokinetic models (ICC \> 0.6, *p* \< 0.05) for K^trans^ and significant scan-rescan ICCs for v~e~ (Tofts) and v~p~ (Patlak). Correlation with histologic findings in 13 CEA patients showed significant positive Pearson's correlation (*ρ* = 0.7; *p* \< 0.01) with the entire vessel wall microvasculature for all models, with the exception of the Tofts model. It was concluded that the Patlak model was the most suited of these four models for pharmacokinetic modeling of the microvasculature in atherosclerotic plaques \[[@CR39]\]. Another study \[[@CR23]\], however, found favorable results for the extended graphical model for simulated and selected in vivo data of carotid plaques with good to excellent image quality. Their results showed that a compromise between noise and bias sensitivity has to be made when choosing between the Patlak and extended graphical models.
The scan-rescan reproducibility of DCE-MRI was also investigated in a multi-center study \[[@CR18]\] of 35 subjects with established cardiovascular disease recruited from 15 hospitals. Results showed a moderate reproducibility for K^trans^ with a coefficient of variation of 25 %. The relatively high dropout rate within the study (31.4 %) suggested a need for intensive operator training, an optimized imaging protocol, and quality control.
The dependence of model parameters on contrast medium was investigated in a study comparing two extracellular contrast media \[[@CR21]\]. Quantitative analysis of DCE-MR images demonstrated a lower K^trans^ when using gadobenate dimeglumine (0.0846 min^−1^) as compared to gadodiamide (0.101 min^−1^, *p* \< 0.01), while no difference in v~p~ was found. In order to facilitate direct comparison of quantitative DCE-MRI parameters between- or in longitudinal studies, the use of the same contrast medium is recommended.
Taken together, despite the recognized limitations, the above studies demonstrate the suitability of quantitative DCE-MRI parameters for reproducibly determining plaque microvasculature characteristics.
Overview of DCE MRI Studies to Study Plaque Microvasculature {#Sec11}
============================================================
Association Between DCE-MRI Parameters and Other Plaque Features {#Sec12}
----------------------------------------------------------------
Many plaque characteristics and pathological features contribute to the risk for disruption and thrombosis, and studies have been designed to investigate possible associations between plaque microvasculature and other plaque features. In recent years, several studies \[[@CR31], [@CR37], [@CR45]•, [@CR46], [@CR47]•\] were carried out to investigate associations between DCE-MRI parameters, plaque inflammation, and the presence of IPH. In a preclinical study of cholesterol-fed balloon-injured atherosclerotic rabbits \[[@CR31]\], a positive Pearson correlation (*ρ* = 0.70, *p* = 0.01) was found between DCE-MRI derived parameters and histologically determined plaque macrophage content.
The relationship between DCE-MRI parameters and plaque inflammation using ^18^fluorine-fluorodeoxyglucose (^18^F-FDG) PET-computed tomography (CT) has been investigated in several clinical studies \[[@CR37], [@CR46], [@CR47]•\]. One study of 49 symptomatic patients with mild to moderate carotid stenosis \[[@CR46]\] reported a weak positive correlation (Spearman *ρ* = 0.30, *p* = 0.035) between plaque inflammation (mean Target-to-Background Ratio (TBR) on ^18^F-FDG PET-CT) and plaque perfusion (mean K^trans^). Another study of 33 patients \[[@CR37]\] with coronary heart disease (CHD) or CHD risk equivalent and a carotid plaque with TBR ≥ 1.6 on ^18^F-FDG PET-CT \[[@CR69]\] found a significant inverse relationship between plaque perfusion (K^trans^) and plaque inflammation on ^18^F-FDG PET-CT of *ρ* = -0.24 (*p* \< 0.05). A subsequent study of 41 patients with carotid plaque \[[@CR47]•\] found that correlations depend on the clinical condition of patients. Overall, a weak, marginal non-significant correlation (Spearman *ρ* = 0.22, *p* = 0.068) was found for all, both symptomatic and asymptomatic, carotid plaques. A significant difference in Spearman correlation coefficients between TBR and K^trans^ was found when grouped according to the symptomatic and asymptomatic carotid plaques (*p* = 0.033): a significant correlation (Spearman *ρ* = 0.59, *p* = 0.006) was found for symptomatic carotid plaques, not seen for asymptomatic plaques (Spearman *ρ* = 0.07, *p* = 0.625). Also, an inverse relationship was found between the time since the last neurological event and both parameters (Spearman *ρ* = −0.94 for TBR and Spearman *ρ* = −0.69 for K^trans^). These results point towards a complex interplay between inflammation and microvasculature in atherosclerotic plaques that is difficult to capture in clinical imaging.
The link between plaque microvasculature and the specific feature of intraplaque hemorrhage (IPH) has been investigated in symptomatic patients with moderate to severe carotid stenosis \[[@CR45]•\]. The presence of IPH on MP-RAGE MR images was associated with a significant increase in K^trans^ of 28 % (*p* \< 0.001) in the adventitial layer of the vessel wall as compared to arteries where IPH was absent (*p* \< 0.001). A multivariate analysis adjusting for symptomatic status, degree of stenosis, and male sex showed that the increased K^trans^ in arteries with IPH remained significant (*p* = 0.018).
These studies show the potential of DCE-MRI as a tool to gain more insight in relation between plaque microvasculature and other features of vulnerable atherosclerotic lesions.
Monitoring Longitudinal Changes in Plaque Microvasculature {#Sec13}
----------------------------------------------------------
DCE-MRI can be used to follow progression of atherosclerotic plaques, as illustrated by a preclinical study \[[@CR31]\] of cholesterol-fed atherosclerotic rabbits. One group of rabbits was imaged 3 months after balloon denudation, immediately followed by euthanasia, and a second group at 3 and 6 months after balloon denudation. From 3 to 6 months after balloon denudation, an increase of 40 % in K^trans^ was found measured by DCE-MRI, suggesting that DCE-MRI can be used to investigate plaque microvasculature development.
Differences Between Different Animal Groups and Human Subjects with a Different Cardiovascular Risk Profile {#Sec14}
-----------------------------------------------------------------------------------------------------------
In a recent rabbit study \[[@CR32]\], investigating the development of a microfluidic chip for potential future nanomedicines an increased AUC within the abdominal aorta for atherosclerotic animals as compared to control animals was reported. In another study of cholesterol-fed rabbits with induced plaque disruption \[[@CR26]\], it was shown that ruptured plaques can be distinguished from stable plaques by spatial-temporal texture-based features of DCE-MRI. The effect of exposure to high particulate airborne matter on atherosclerosis was investigated in "Ground Zero" workers in New York City with high and low exposure to particulate matter using DCE-MRI \[[@CR43]\]. Subjects with high exposure had a significantly higher AUC in the carotid artery (+41 %) as compared to subjects with low exposure (*p* = 0.016), indicating increased changes of the plaque microvasculature. These changes may range from increased leakage of contrast medium from the microvasculature, increased microvessel density, increased flow through the microvasculature, decreased reflux from the extracellular extravascular space to the microvasculature, or a combination. The authors of the study concluded that a high exposure to particulate matter may lead to increased plaque microvasculature, potentially indicating an increased risk for further development of atherosclerosis.
Evaluation of Therapies {#Sec15}
-----------------------
DCE-MRI enables the study of plaque microvasculature changes over time, making it useful in animal and patient drug effect studies. Changes in microvasculature may reflect changes in phenotype and/or vulnerability of the atherosclerotic plaque.
DCE-MRI has been employed in several preclinical cholesterol-fed balloon-injured atherosclerotic rabbit studies investigating potential anti-inflammatory treatments of atherosclerosis \[[@CR33]--[@CR35]\]. The effect of liposome-encapsulated prednisolone phosphate (L-PLP) on atherosclerosis was investigated using MR imaging before treatment, immediately after injection with L-PLP, and over time \[[@CR33]\]. A reduction of the plaque AUC was found from pre-treatment to 2 days post-treatment, revealing early changes in microvascular permeability after treatment. In a further study, the anti-inflammatory effects of pioglitazone on atherosclerotic plaques were investigated \[[@CR34]\]. DCE-MRI analysis showed a 22 % decrease in AUC for the treatment group (*p* \< 0.01) over the study time period of 3 months, while no decrease in plaque enhancement was found for the control group. No changes in vessel wall area measurements were found during the study period for either animal group. A third study \[[@CR35]\] evaluated the anti-inflammatory effects of a liver X receptor (LXR) agonist which induces reversal cholesterol transport, as compared to atorvastatin. The 3-month treatment with LXR did not lead to changes of the microvasculature, whereas treatment with atorvastatin caused a trend towards a decrease in microvasculature (*p* = 0.06). No differences in vessel wall area measurements were found. Combined, these studies have shown the potential of DCE-MRI to study changes of the plaque microvasculature in the evaluation of potential new therapies. A limitation of these studies, however, is that the rabbits did not exhibit plaque disruption with luminal thrombosis, the clinical endpoint of high risk plaques.
DCE-MRI has been used to study the effect of intensive lipid therapy over a period of 12 months \[[@CR38]\] in patients with coronary artery disease or carotid disease and increased levels (≥120 mg/dl) of apolipoprotein B from the Carotid Plaque Composition study \[[@CR70]\]. Results of the study show that 12-month therapy leads to a significant reduction of 21 % in K^trans^. This is consistent with the hypothesis that intensive lipid therapy results in a reduction of the extent and permeability of atherosclerotic plaque microvasculature. A study with 98 subjects with established cardiovascular disease \[[@CR44]••\] selected from the AIM-HIGH trial \[[@CR19]\] found an inverse association between v~p~ (plaque microvasculature fraction) and the duration of statin therapy. Statins are commonly used to lower lipid levels and also possess anti-inflammatory properties \[[@CR71]\]. These results suggest that a relationship exists between duration of statin therapy and plaque microvasculature, which could reflect a decreased level of vascular inflammation.
The above studies on DCE-MRI of plaque microvasculature have measured differences between treatment groups or subjects with increased cardiovascular risk and shown that DCE-MRI can be employed effectively as an evaluation tool.
Challenges and Future Perspectives in DCE-MRI of Atherosclerosis {#Sec16}
================================================================
To further advance DCE-MRI for wider use in clinical practice, uniform acquisition and analysis methods need to be agreed upon. Previous studies have shown that DCE-MRI-derived parameters are influenced by the contrast medium, vascular input function, and which pharmacokinetic model is used, making direct cross-study comparisons difficult. Use of a standard imaging and data analysis protocol is essential, therefore, for longitudinal studies of plaque microvasculature. A very important clinical precaution is use of a stable Gadolinium-based contrast medium \[[@CR72]\] using low dosages to minimize the risks for nephrogenic systemic fibrosis and deposition of the contrast medium in the brain.
The recent introduction of interleaved acquisition methods \[[@CR28]••\], providing both bright and black blood images, may be an important step toward the determination of an individualized vascular input function. In addition, 3D acquisition techniques \[[@CR30]•\] may provide increased spatial accuracy as compared to currently employed 2D techniques, although at the expense of temporal resolution. Currently, these 3D acquisition techniques have only been explored in preclinical rabbit studies; their potential in clinical studies remains to be investigated.
All clinical DCE-MRI studies, to date, have been performed at 1.5 and 3.0 T. The potential of carotid MRI at 7.0 T has already been explored \[[@CR73], [@CR74]\], and results show a potential increase in signal-to-noise ratio (SNR) due to the increased field strength. However, the increase in SNR may be diminished by increased relaxivity of the contrast medium, which may also require longer scan times. The potentially increased SNR at 7.0 T would allow increased spatial and/or temporal resolution, but these studies also demonstrate that further technical developments are required to enable complete plaque characterization.
Associations between plaque microvasculature (measured using DCE-MRI) and plaque inflammation (measured by ^18^F-FDG uptake or macrophage content) remain an important area to be studied further since varying results have been reported to date. The reported association between and plaque microvasculature and intraplaque hemorrhage could be studied longitudinally. The recent introduction of hybrid PET-MRI systems provides excellent opportunities for further investigation of the relationships between these processes using a single imaging system. Recent research \[[@CR75]--[@CR79]\] has already shown the potential of hybrid PET-MR systems for the imaging of atherosclerosis. However, the additional value of DCE-MRI in PET/MR imaging is yet to be explored.
The predictive value of DCE-MRI for plaque progression or development of vulnerable plaque features is of great interest and remains to be determined; in addition, its predictive value for cerebrovascular ischemic events needs to be investigated in a prospective clinical trial.
Applications of DCE-MRI can be extended beyond the carotid artery to other (human) vascular territories, such as the microvasculature in the aortic wall of abdominal aortic aneurysms \[[@CR80], [@CR81]\]. These measurements were reproducible with a high technical success rate, and the Patlak model was the most suited pharmacokinetic model. Future studies are warranted to investigate the predictive potential of DCE-MRI derived parameters for abdominal aortic aneurysm rupture risk.
Conclusion {#Sec17}
==========
Over the past decade, DCE-MRI has developed from a novel imaging tool to a useful non-invasive research tool used in animal and patient studies of plaque microvasculature. DCE-MRI has been used to investigate the relationship between plaque microvasculature and other plaque features such as inflammation and intraplaque hemorrhage, for assessing effectiveness of therapeutic interventions, and in the evaluation of plaque microvasculature changes over time and between groups with increased cardiovascular risks. Future studies could apply DCE-MRI to elucidate plaque development mechanisms, specifically the interplay between inflammation, increased microvasculature, and intraplaque hemorrhage. Also of great interest is the potential predictive value of plaque microvasculature DCE-MRI for plaque progression and future cerebrovascular ischemic events (such as stroke).
Search Strategy {#Sec18}
===============
The studies discussed in the present review have been identified through a database search in MEDLINE in December 2015 using the following search terms: "carotid atherosclerosis"/"atherosclerosis"/"atherosclerotic plaque"/"atherosclerotic plaques"/"plaque" AND "human"/"rabbit" AND "DCE-MRI"/"dynamic contrast enhanced MRI"/"MRI"/"dynamic contrast enhanced magnetic resonance imaging"/"magnetic resonance imaging" AND "neovessels"/"neovascularization"/"neovasculature"/"vasa vasorum"/"microvasculature"/"inflammation". Resulting abstracts and articles were screened and references checked for possible additional studies.
This article is part of the Topical Collection on *Vascular Biology*
This research was performed within the framework of CTMM, the Center for Translational Molecular Medicine ([www.ctmm.nl](http://www.ctmm.nl/)), project PARISk (grant 01C-202), and supported by the Dutch Heart Foundation. M.E. Kooi is supported by Aspasia Grant 015.008.047 from the Netherlands Organization for Scientific Research. J.E. Wildberger and M.E. Kooi are supported by Stichting de Weijerhorst. The authors would like to thank Estelle C. Nijssen for critically reviewing the manuscript.
Conflict of Interest {#FPar1}
====================
Raf H.M. van Hoof declares grant support from CTMM. Sylvia Heeneman declares no conflict of interest. Joachim E. Wildberger declares grant support from Siemens, Philips, Bayer, and AGFA, and declares personal fees from Siemens and Bayer. M. Eline Kooi declares grant support from CTMM, Stichting de Weijerhorst, NWO (Aspasia), and Servier.
Human and Animal Rights and Informed Consent {#FPar2}
============================================
All studies by R.H.M. van Hoof, S. Heeneman, J.E. Wildberger, and M.E. Kooi involving animal and/or human subjects were performed after approval by the appropriate institutional review boards. Where required, written informed consent was obtained from all participants.
|
Tip: You can add any amount to your meal plan, not just common serving units. Want a Arnold Dutch Country smooth texture 100% Whole Wheat serving size to fit your carbs limit? No problem, just use the Arnold Dutch Country smooth texture 100% Whole Wheat nutrition facts serving size tool to the right.
Find more foods like Arnold Dutch Country smooth texture 100% Whole Wheat. FitClick has over 60,000 foods and recipes with detailed nutrition information to view.
Search Foods
Browse this section for quick links to our diet plans and other popular diet and nutrition features. From diet plans to food calories data, FitClick has the content you need to lead a healthy life. Find out how many calories in your food and more at your source for diet and nutrition information.
|
Q:
How to check device is Online or Offline Phonegap
What is the easiest way to check whether the device(smartphone) is online or offline.
I'm working with phonegap, jquery mobile.
I found this one.
document.addEventListener("online", yourCallbackFunction, false);
What I want to do is to check the Internet connection and decide about to get the data from the internet or from the device local database.
if(deviceoffline)
getDataFromLokalDatabase();
else
getDataFromInternet();
A:
This is the code using native phonegap code:
function checkInternet() {
var networkState = navigator.connection.type;
if(networkState == Connection.NONE) {
onConnexionError();
return false;
} else {
return true;
}
}
A:
You can use navigator.onLine:
var isOffline = 'onLine' in navigator && !navigator.onLine;
if ( isOffline ) {
//local db
}
else {
// internet data
}
But be aware of that if navigator.onLine isn't supported isOffline will always be false
|
Q:
mapview random annotation image
I'm having great trouble figuring out what I'm doing wrong here...
I am adding annotations to my mapview through loops, but the annotation image is completely random every time... when I NSLog the order, it's random - i'm not sure if that is the problem.
- (MKAnnotationView *)mapView:(MKMapView *)mapView1 viewForAnnotation:(id<MKAnnotation>)annotation {
if ([annotation isKindOfClass:[MKUserLocation class]])
return nil;
if ([[annotation subtitle] isEqual:@"Bar"]) {
MKAnnotationView *view = nil;
view = [self.mapView dequeueReusableAnnotationViewWithIdentifier:@"myAnnotationIdentifier"];
if (!view) {
// Could not reuse a view ...
// Creating a new annotation view
view = [[MKAnnotationView alloc] initWithAnnotation:annotation reuseIdentifier:@"myAnnotationIdentifier"];
view.enabled = YES;
view.canShowCallout = YES;
view.rightCalloutAccessoryView = [UIButton buttonWithType:UIButtonTypeDetailDisclosure];
view.image = [UIImage imageNamed:@"beer.png"];
}
return view;
}
else if ([[annotation subtitle] isEqual:@"Club"]) {
MKAnnotationView *view = nil;
view = [self.mapView dequeueReusableAnnotationViewWithIdentifier:@"myAnnotationIdentifier"];
if (!view) {
// Could not reuse a view ...
// Creating a new annotation view
view = [[MKAnnotationView alloc] initWithAnnotation:annotation reuseIdentifier:@"myAnnotationIdentifier"];
view.enabled = YES;
view.canShowCallout = YES;
view.rightCalloutAccessoryView = [UIButton buttonWithType:UIButtonTypeDetailDisclosure];
view.image = [UIImage imageNamed:@"clubs.png"];
}
return view;
}
}
the view.image is totally random... either clubs.png or beer.png..
How do I make it correctly?
this is how I add the annotations:
- (void)barLoop {
for (int i = 0; i<barArray.count; i++) {
int index = [[barArray objectAtIndex:i]intValue];
NSString *lati = [[[self.usersLocationArray objectAtIndex:index]
valueForKeyPath:@"items.Latitude"]componentsJoinedByString:@""];
NSString *longi = [[[self.usersLocationArray objectAtIndex:index]
valueForKeyPath:@"items.Longitude"]componentsJoinedByString:@""];
NSString *barNavn = [[[self.usersLocationArray objectAtIndex:index] valueForKeyPath:@"items.Navn"]componentsJoinedByString:@""];
float latitude = [lati floatValue];
float longitude = [longi floatValue];
MKCoordinateRegion region = { {0.0, 0.0} , {0.0, 0.0} };
region.center.latitude = latitude;
region.center.longitude = longitude;
region.span.longitudeDelta = 0.20f;
region.span.latitudeDelta = 0.20f;
[mapView setRegion:region animated:NO];
CLLocationCoordinate2D location;
location.latitude = latitude;
location.longitude = longitude;
Annotation *ann = [[Annotation alloc]initWithPosition:location];
ann.title = barNavn;
ann.subtitle = @"Bar";
[self.mapView addAnnotation:ann];
}
}
thanks in advance :)
A:
You have two types of annotations, but you are only setting the image when the annotation is originally created. Thus, if an annotation for a bar scrolls off the map view and another annotation for a club scrolls on, it might reuse the bar's annotation view for the club.
There are two ways of fixing this:
Use different reuseIdentifier parameters for each of the two types of annotation views; or
Set/reset the annotation view's image regardless of whether your call to dequeueReusableAnnotationViewWithIdentifier succeeded in returning a value or not.
Unrelated, but your viewForAnnotation method should:
You might want to use isEqualToString instead of isEqual when checking for @"Bar" vs @"Club".
Make sure to return nil if neither of those if clauses return true (this should never happen, but nonetheless you have a potential path in this routine in which you neglect to return any value). I would guess that this would have been brought to your attention if you ran the code through the static analyzer ("Analyze" on Xcode's "Product" menu).
A more subtle observation, but I'd probably not rely upon the annotation's subtitle for determining which annotation view to employ, but rather either
have two annotation subclasses, one for bars and one for clubs; or
have a custom property in your existing Annotation class that indicates whether it's a bar or club (and I'd probably use an enum for that).
At some future date, you might want to use the callout's subtitle for something other than "Bar" vs "Club" (e.g., maybe the address of the bar/club?), and the current model seems to conflate a UI attribute (i.e. what shows up in the callout) with a state property (i.e. a variable that controls which annotation view type to use). Not a big deal, but just an suggestion.
|
Foxconn CEO Terry Gou just can't help himself: When he sees an opening to stir up controversy, he goes all in.
Last month, Gou added fuel to the Apple TV rumor fire when he said that his company was making preparations for the iTV. Now, Gou is urging consumers to hold off on buying Samsung's new flagship smartphone. Instead, he says, we should all get an iPhone 5 when it eventually is released.
At a shareholders' meeting Monday, Gou reportedly said that the new model "will put Samsung's Galaxy S III to shame."
Gou also reportedly called Samsung "a company with a track record of snitching on its competitors" – a reference to a European price-fixing investigation of the flat panel industry in 2010. And how's this for an inflammatory statement: "I respect the Japanese and especially like their execution and communication styles. Unlike the Koreans, they will not hit you from behind."
According to gaming blog Kotaku, Japanese blogs are actually reporting that Gou used more colorful language, calling the Koreans a slang word for "country hicks."
Foxconn confirmed an alliance with Japan's Sharp Electronics Co. on Monday. Gou said the partnership should help his company defeat Samsung Electronics in 3 to 5 years.
Via Patently Apple
|
More colors filled in. Just ignore the dots in the top-left corner. I made those just to remember which colors I had used.
Fill the white in the eye with a light gray with some blue in it. I choose a green iris color. You can of course select another color. You can actually get away with pretty 'wild' eye colors. I've read somewhere that like with fingerprints no eyes look
the exact same. Some have very bright eyes, looking almost spookey if you ask me. I find dark eyes more friendly and warm, but thats just me.
Some people have eyes with an almost awkward color, a mix between blue and green. But then you see them in a different light and now they look green and yellow. My advice is to go for an "established" eye color combination in a painting, unless you want that extra attention to the eyes though.
I paint my eyes the same way each time; You have a round and black pupil (the pupil can vary a lot in size, depending on the lightning situation or drug abuse :-)) The iris is the round, colored area around the pupil. On most people the iris has a dark edge both towards the pupil and the eyeball. More ramblings on the iris later in this lesson!
The Iris and the pupil suggested
Start suggesting the depths in the image. The top/center of the eyelid is rendered light because it 'stands out' in the image. The same goes for the area above the eyelid and below the eye brow. As the opposite, the left part of the skin area is painted darker to illustrate that the skin curves back. The area around the eyes is particularly skinful.
You have probably seen drawings/photographs of the human skull and noticed that the eyesockets are much bigger than our eyes. With age, human skin looses some of its firmness and develop wrinkles and creases in the area below the eyes (maybe gravity plays an effect too, I don't know :))
The skin above the eyelid is affected too if not exactly creasing, it starts sagging and can eventually cover the eyelid completely. This probably isn't new to anyone, but I want to illustrate that such should be remembered when drawing people. Age is a part of our identity...
|
Information in our posts and case histories is of tremendous value to all our members and is also used for formal studies. It is extremely important that it be factual. To help protect the data and your privacy only members of the main ECIR Group are granted access to the Case History Group. If you are concerned about privacy you need only give your first name and general location. The use of pseudonyms to keep your equine professionals anonymous is encouraged.
Don't delay in beginning your Case History with the information you have. Missing information it can be added as you obtain it. The faster your case history is uploaded, the faster you will receive support from the volunteers.
A complete and up-to-date case history is invaluable when an equine emergency crops up. Accurate documentation of management changes, the equine's responses, blood work results and links to the equine's photo album not only helps the ECIR Group gain the trust and respect of science-based equine professionals, but can be used by vets who are on the ground trying to help your horse. Your Case History is the best way to effectively monitor and illustrate the results in PPID and EMS management changes.
When the time comes that you or your equine no longer need the ECIR Forum's advice, please leave your photos and case history folder intact. Your information will be used in the future to confirm or re-shape ECIR protocols and help other members learn. This is the best way to give back to the ECIR for the help you and your horse have received.
|
<?php defined('BASEPATH') || exit('No direct script access allowed');
$lang['tr_translate_title'] = 'Translate to %s';
$lang['tr_no_modules'] = 'No Modules found with language files.';
$lang['tr_no_core'] = 'No Core language files found.';
$lang['tr_core'] = 'Core';
$lang['tr_modules'] = 'Modules';
$lang['tr_current_lang'] = 'Translate to';
$lang['tr_language'] = 'Language';
$lang['tr_other'] = 'Other...';
$lang['tr_select_lang'] = 'Select';
$lang['tr_edit_title'] = 'Translate';
$lang['tr_save_success'] = 'Language file saved successfully.';
$lang['tr_save_fail'] = 'There was an error trying to save language file. Check file permissions.';
$lang['tr_translate_file'] = 'File';
$lang['tr_translate'] = 'Translate';
$lang['tr_export'] = 'Export Language Files';
$lang['tr_export_short'] = 'Export';
$lang['tr_export_note'] = 'Exporting language files provides you with a .zip file with all of the language files for that particular language so that you can easily share.';
$lang['tr_include'] = 'Select Parts';
$lang['tr_include_core'] = 'Include Core';
$lang['tr_include_mods'] = 'Include Custom Modules';
$lang['tr_translate_success'] = 'selected language entries were successfully translated.';
$lang['tr_translate_part_success'] = 'selected language entries were successfully translated but %s failed.';
$lang['tr_translate_failed'] = 'The selected language entries were not translated. See the error log for details.';
$lang['tr_new_lang'] = 'New Language';
|
Effect of routine bedside procedures on intracranial pressure.
The effect on intracranial pressure (ICP) of routine bedside procedures, such as changing the patient's position, suction, rotation, flexion or extension of the head, was investigated in 21 comatose patients with brain edema. Simple maneuvers, which under physiological conditions have no effect on ICP, often led to significant changes in the ICP in these patients. It is suggested that monitoring of ICP and adjustment of body position in accordance with the level of ICP become a standard procedure in neurosurgical wards.
|
<?xml version="1.0" encoding="UTF-8"?>
<level>
<information>
<Property name="game_version" value="1.8" />
<Property name="engine_version" value="34" />
<Property name="save_time" value="1236475902" />
</information>
<settings>
<Property name="lvl_author" value="" />
<Property name="lvl_version" value="" />
<Property name="lvl_music" value="land/land_5.ogg" />
<Property name="cam_limit_x" value="0" />
<Property name="cam_limit_y" value="0" />
<Property name="cam_limit_w" value="20000" />
<Property name="cam_limit_h" value="-4000" />
<Property name="cam_fixed_hor_vel" value="0" />
</settings>
<background>
<Property name="type" value="103" />
<Property name="bg_color_1_red" value="0" />
<Property name="bg_color_1_green" value="0" />
<Property name="bg_color_1_blue" value="0" />
<Property name="bg_color_2_red" value="0" />
<Property name="bg_color_2_green" value="0" />
<Property name="bg_color_2_blue" value="0" />
</background>
<player>
<Property name="posx" value="482" />
<Property name="posy" value="7" />
<Property name="direction" value="right" />
</player>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="167" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="559" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="602" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="645" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="688" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="731" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="774" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="609" />
<Property name="posy" value="18" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="right" />
<Property name="max_distance" value="360" />
<Property name="speed" value="40" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="0" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<sprite>
<Property name="posx" value="344" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="258" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="-91" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="-134" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="furball" />
<Property name="posx" value="901" />
<Property name="posy" value="-222" />
<Property name="color" value="brown" />
<Property name="direction" value="left" />
</enemy>
<sprite>
<Property name="posx" value="688" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="645" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="602" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="559" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="516" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="559" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="602" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="645" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="688" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="731" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="774" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="903" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="946" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1075" />
<Property name="posy" value="38" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1208" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/left.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1258" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="946" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/wood/1.png" />
<Property name="type" value="halfmassive" />
</sprite>
<sprite>
<Property name="posx" value="903" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/wood/1.png" />
<Property name="type" value="halfmassive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<levelexit>
<Property name="posx" value="1206" />
<Property name="posy" value="29" />
<Property name="type" value="1" />
<Property name="motion" value="0" />
<Property name="direction" value="right" />
</levelexit>
<sprite>
<Property name="posx" value="1308" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1358" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1408" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1458" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="516" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/wood/1.png" />
<Property name="type" value="halfmassive" />
</sprite>
<sprite>
<Property name="posx" value="473" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/wood/1.png" />
<Property name="type" value="halfmassive" />
</sprite>
<sprite>
<Property name="posx" value="871" />
<Property name="posy" value="19" />
<Property name="image" value="signs/default_1/1_ending_big.png" />
<Property name="type" value="passive" />
</sprite>
<sprite>
<Property name="posx" value="688" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="215" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="258" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="301" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="215" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="301" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-5" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="38" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="215" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="258" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="301" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="344" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-134" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-91" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="68" />
<Property name="posy" value="-126" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="down" />
<Property name="max_distance" value="150" />
<Property name="speed" value="2" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<moving_platform>
<Property name="posx" value="194" />
<Property name="posy" value="-127" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="halfmassive" />
<Property name="direction" value="down" />
<Property name="max_distance" value="150" />
<Property name="speed" value="2" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<sprite>
<Property name="posx" value="731" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="774" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="903" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-91" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="946" />
<Property name="posy" value="-91" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="561" />
<Property name="posy" value="-111" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="right" />
<Property name="max_distance" value="280" />
<Property name="speed" value="3" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<enemy>
<Property name="type" value="turtle" />
<Property name="posx" value="843" />
<Property name="posy" value="-128" />
<Property name="color" value="red" />
<Property name="direction" value="right" />
</enemy>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="210" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="furball" />
<Property name="posx" value="151" />
<Property name="posy" value="161" />
<Property name="color" value="brown" />
<Property name="direction" value="right" />
</enemy>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="167" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="210" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="turtle" />
<Property name="posx" value="190" />
<Property name="posy" value="131" />
<Property name="color" value="red" />
<Property name="direction" value="left" />
</enemy>
<enemy>
<Property name="type" value="spika" />
<Property name="posx" value="260" />
<Property name="posy" value="142" />
<Property name="color" value="green" />
</enemy>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="210" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="167" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="516" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="516" />
<Property name="posy" value="-91" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="473" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="43" />
<Property name="posy" value="210" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="down" />
<Property name="max_distance" value="70" />
<Property name="speed" value="2" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="7" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<sprite>
<Property name="posx" value="473" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="602" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="645" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="688" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="731" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="774" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="903" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="946" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="344" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="167" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="167" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="eato" />
<Property name="posx" value="396" />
<Property name="posy" value="193" />
<Property name="image_dir" value="enemy/eato/brown/" />
<Property name="direction" value="right_bottom" />
</enemy>
<enemy>
<Property name="type" value="furball" />
<Property name="posx" value="690" />
<Property name="posy" value="-222" />
<Property name="color" value="blue" />
<Property name="direction" value="right" />
</enemy>
<sprite>
<Property name="posx" value="1508" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1558" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1608" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1658" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1708" />
<Property name="posy" value="-12" />
<Property name="image" value="pipes/orange/hor.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1032" />
<Property name="posy" value="-134" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1075" />
<Property name="posy" value="-134" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="thromp" />
<Property name="posx" value="474" />
<Property name="posy" value="221" />
<Property name="image_dir" value="enemy/thromp/" />
<Property name="direction" value="right" />
<Property name="max_distance" value="500" />
<Property name="speed" value="7" />
</enemy>
<enemy>
<Property name="type" value="thromp" />
<Property name="posx" value="959" />
<Property name="posy" value="188" />
<Property name="image_dir" value="enemy/thromp/" />
<Property name="direction" value="left" />
<Property name="max_distance" value="500" />
<Property name="speed" value="7" />
</enemy>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="296" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="339" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="382" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="473" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="516" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="559" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="602" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="645" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="688" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="731" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="774" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="903" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="946" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1032" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1075" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1118" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1161" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="382" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="339" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="296" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="253" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="210" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="167" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="124" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1161" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1118" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1075" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1032" />
<Property name="posy" value="81" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="flyon" />
<Property name="posx" value="342" />
<Property name="posy" value="359" />
<Property name="direction" value="left" />
<Property name="image_dir" value="enemy/flyon/orange/" />
<Property name="max_distance" value="100" />
<Property name="speed" value="5.8" />
</enemy>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="thromp" />
<Property name="posx" value="392" />
<Property name="posy" value="367" />
<Property name="image_dir" value="enemy/thromp/desert_1/" />
<Property name="direction" value="down" />
<Property name="max_distance" value="200" />
<Property name="speed" value="7" />
</enemy>
<sprite>
<Property name="posx" value="344" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="static" />
<Property name="posx" value="249" />
<Property name="posy" value="362" />
<Property name="rotation_speed" value="2.5" />
<Property name="image" value="enemy/static/blocks/spike_1/1_grey.png" />
<Property name="path" value="" />
<Property name="speed" value="0" />
</enemy>
<sprite>
<Property name="posx" value="301" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="static" />
<Property name="posx" value="293" />
<Property name="posy" value="361" />
<Property name="rotation_speed" value="0" />
<Property name="image" value="enemy/static/blocks/spike_1/2_grey.png" />
<Property name="path" value="" />
<Property name="speed" value="0" />
</enemy>
<sprite>
<Property name="posx" value="258" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="57" />
<Property name="posy" value="364" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="right" />
<Property name="max_distance" value="280" />
<Property name="speed" value="3" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<sprite>
<Property name="posx" value="215" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="382" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="425" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="339" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="296" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="thromp" />
<Property name="posx" value="1033" />
<Property name="posy" value="123" />
<Property name="image_dir" value="enemy/thromp/" />
<Property name="direction" value="down" />
<Property name="max_distance" value="300" />
<Property name="speed" value="7" />
</enemy>
<enemy>
<Property name="type" value="thromp" />
<Property name="posx" value="1121" />
<Property name="posy" value="397" />
<Property name="image_dir" value="enemy/thromp/" />
<Property name="direction" value="up" />
<Property name="max_distance" value="300" />
<Property name="speed" value="7" />
</enemy>
<sprite>
<Property name="posx" value="1075" />
<Property name="posy" value="-5" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<box>
<Property name="posx" value="516" />
<Property name="posy" value="-220" />
<Property name="type" value="text" />
<Property name="invisible" value="0" />
<Property name="useable_count" value="-1" />
<Property name="text" value="Bugs:<br/>-Climbable platform doesn't function correctly<br/><br/>" />
</box>
<moving_platform>
<Property name="posx" value="473" />
<Property name="posy" value="296" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="halfmassive" />
<Property name="direction" value="down" />
<Property name="max_distance" value="85" />
<Property name="speed" value="2" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="5" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<moving_platform>
<Property name="posx" value="774" />
<Property name="posy" value="296" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="climbable" />
<Property name="direction" value="down" />
<Property name="max_distance" value="85" />
<Property name="speed" value="2" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="4" />
<Property name="image_top_left" value="blocks/stone/2.png" />
<Property name="image_top_middle" value="blocks/stone/2.png" />
<Property name="image_top_right" value="blocks/stone/2.png" />
</moving_platform>
<sprite>
<Property name="posx" value="215" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-306" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-306" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-349" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-349" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-392" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-392" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-435" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-435" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-392" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-349" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-306" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-306" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-349" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-306" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="339" />
<Property name="posy" value="-319" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="right" />
<Property name="max_distance" value="280" />
<Property name="speed" value="3" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="1" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<enemystopper>
<Property name="posx" value="700" />
<Property name="posy" value="-277" />
</enemystopper>
<enemy>
<Property name="type" value="furball" />
<Property name="posx" value="511" />
<Property name="posy" value="-307" />
<Property name="color" value="brown" />
<Property name="direction" value="right" />
</enemy>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-134" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-177" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-220" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="946" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="903" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="774" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="731" />
<Property name="posy" value="-263" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="288" />
<Property name="posy" value="-502" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="right" />
<Property name="max_distance" value="400" />
<Property name="speed" value="3" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<enemy>
<Property name="type" value="gee" />
<Property name="posx" value="375" />
<Property name="posy" value="-505" />
<Property name="direction" value="horizontal" />
<Property name="max_distance" value="0" />
<Property name="always_fly" value="0" />
<Property name="wait_time" value="-1" />
<Property name="fly_distance" value="400" />
<Property name="color" value="green" />
</enemy>
<moving_platform>
<Property name="posx" value="424" />
<Property name="posy" value="-502" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="right" />
<Property name="max_distance" value="400" />
<Property name="speed" value="3" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-306" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-349" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-392" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-435" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-478" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-521" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-564" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-607" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-650" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-693" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-478" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-521" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-564" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-607" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-650" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-693" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="946" />
<Property name="posy" value="-650" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="903" />
<Property name="posy" value="-650" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-650" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="-650" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-736" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="903" />
<Property name="posy" value="-1123" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="halfmassive" />
<Property name="direction" value="down" />
<Property name="max_distance" value="400" />
<Property name="speed" value="2" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="0" />
<Property name="middle_img_count" value="0" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-822" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="774" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="731" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="688" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="645" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="602" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="559" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="516" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="473" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="344" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="301" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="258" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="215" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-779" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-736" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-865" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-908" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-951" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-994" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-1037" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-1080" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-1123" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-822" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-865" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-908" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-951" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-994" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-1037" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-1080" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-1123" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-1166" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-1209" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-822" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-822" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-822" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-822" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-822" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-865" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-865" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-865" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-865" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-865" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-908" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-908" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-908" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-908" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-908" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-951" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-951" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-951" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-951" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-951" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-994" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-994" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-994" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-994" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-994" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-1037" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-1037" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-1037" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-1037" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-1037" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-1080" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-1080" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-1080" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-1080" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-1080" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-1123" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-1123" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-1123" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-1123" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-1123" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<moving_platform>
<Property name="posx" value="215" />
<Property name="posy" value="-1123" />
<Property name="move_type" value="0" />
<Property name="massive_type" value="massive" />
<Property name="direction" value="down" />
<Property name="max_distance" value="280" />
<Property name="speed" value="1" />
<Property name="touch_time" value="0" />
<Property name="shake_time" value="16" />
<Property name="touch_move_time" value="10" />
<Property name="middle_img_count" value="13" />
<Property name="image_top_left" value="blocks/stone/1.png" />
<Property name="image_top_middle" value="blocks/stone/1.png" />
<Property name="image_top_right" value="blocks/stone/1.png" />
</moving_platform>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-1252" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-1295" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="989" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="946" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="903" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="860" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="817" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="774" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="731" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="688" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="645" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="602" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="559" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="516" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="473" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="430" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="387" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="344" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="301" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="258" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="215" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-1338" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<box>
<Property name="posx" value="946" />
<Property name="posy" value="-1295" />
<Property name="type" value="text" />
<Property name="invisible" value="0" />
<Property name="useable_count" value="-1" />
<Property name="text" value="Crush the boss and spika, NOW!<br/>" />
</box>
<enemy>
<Property name="type" value="turtleboss" />
<Property name="posx" value="222" />
<Property name="posy" value="-982" />
<Property name="color" value="red" />
<Property name="direction" value="right" />
<Property name="max_hit_count" value="0" />
<Property name="max_downgrade_count" value="0" />
<Property name="shell_time" value="2.5" />
</enemy>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-1295" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-1295" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-1295" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-1295" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-1295" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-1252" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-1252" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-1252" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-1252" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-1252" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-1209" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-1209" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-1209" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-1209" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-1209" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="172" />
<Property name="posy" value="-1166" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="129" />
<Property name="posy" value="-1166" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="86" />
<Property name="posy" value="-1166" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="43" />
<Property name="posy" value="-1166" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="0" />
<Property name="posy" value="-1166" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<enemy>
<Property name="type" value="static" />
<Property name="posx" value="814" />
<Property name="posy" value="-1062" />
<Property name="rotation_speed" value="0" />
<Property name="image" value="enemy/static/blocks/spike_1/2_grey.png" />
<Property name="path" value="" />
<Property name="speed" value="0" />
</enemy>
<enemy>
<Property name="type" value="static" />
<Property name="posx" value="757" />
<Property name="posy" value="-1068" />
<Property name="rotation_speed" value="-5.5" />
<Property name="image" value="enemy/static/saw/default.png" />
<Property name="path" value="" />
<Property name="speed" value="0" />
</enemy>
<enemy>
<Property name="type" value="static" />
<Property name="posx" value="712" />
<Property name="posy" value="-1062" />
<Property name="rotation_speed" value="2.5" />
<Property name="image" value="enemy/static/blocks/spike_1/1_grey.png" />
<Property name="path" value="" />
<Property name="speed" value="0" />
</enemy>
<enemy>
<Property name="type" value="flyon" />
<Property name="posx" value="539" />
<Property name="posy" value="-1072" />
<Property name="direction" value="left" />
<Property name="image_dir" value="enemy/flyon/orange/" />
<Property name="max_distance" value="200" />
<Property name="speed" value="5.8" />
</enemy>
<enemy>
<Property name="type" value="thromp" />
<Property name="posx" value="227" />
<Property name="posy" value="-1028" />
<Property name="image_dir" value="enemy/thromp/desert_1/" />
<Property name="direction" value="right" />
<Property name="max_distance" value="200" />
<Property name="speed" value="7" />
</enemy>
<enemy>
<Property name="type" value="spika" />
<Property name="posx" value="719" />
<Property name="posy" value="-913" />
<Property name="color" value="grey" />
</enemy>
<enemystopper>
<Property name="posx" value="701" />
<Property name="posy" value="-853" />
</enemystopper>
<enemy>
<Property name="type" value="static" />
<Property name="posx" value="200" />
<Property name="posy" value="360" />
<Property name="rotation_speed" value="0" />
<Property name="image" value="enemy/static/blocks/spike_1/2_grey.png" />
<Property name="path" value="" />
<Property name="speed" value="0" />
</enemy>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="-48" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1118" />
<Property name="posy" value="-134" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1161" />
<Property name="posy" value="-134" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="-134" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
<sprite>
<Property name="posx" value="1204" />
<Property name="posy" value="-91" />
<Property name="image" value="blocks/stone/1.png" />
<Property name="type" value="massive" />
</sprite>
</level>
|
1. Field of the Invention
The invention is directed to a coupling part for a fluid line coupling.
2. Description of the Related Art
One known coupling part is disclosed in DE 43 00 037 C1. The prior coupling part is provided with a locking element held in a receiving portion and designed to secure a plug-in part that is configured with an annular shoulder and is to be introduced into a receiving space of the receiving portion. The locking element is configured as an annularly closed spring element, in which two pressure plates emerging from the outer side of the receiving portion have two annular portions formed on them mutually oppositely, each annular portion being configured with a respective locking catch that engages behind the annular shoulder of the plug-in part. If the plug-in part is not inserted to the proper depth into the receiving space, however, there is a risk that the plug-in part will not be properly secured in the coupling part by the locking sections engaging behind the annular shoulder.
Known from DE 10 2004 062 887 B3 is a coupling part for a fluid line coupling in which a locking element is configured both with two retaining arms, which are devised to engage behind an annular shoulder configured on a plug-in part that is to be introduced into the coupling part, and with two locking arms. The locking arms are flexible in the axial direction of the coupling part, and each has at its free end an engaging arrangement that cooperates in such a way with a locking projection configured on the coupling element that when the plug-in part is introduced to the proper depth, the engagement between the engagement structure and the locking projection is released by the action of the annular shoulder butting against the locking arms and the locking arms flexing in the insertion direction, and the locking element can be shifted from an open position into a closed position that secures the plug-in part by virtue of the re-engagement of the locking structure and the locking projection. To release the connection between the plug-in part and the locking part, however, the locking arms must again be flexed in the insertion direction in order to release the engagement between the engagement structure and the locking projection in the closed position, a manipulation which, however, is very problematic or downright impossible, especially under very tight conditions.
|
###
# #%L
# thinkbig-service-app
# %%
# Copyright (C) 2017 ThinkBig Analytics
# %%
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# #L%
###
#
# Spring Datasource properties for spring batch and the default data source
#
spring.datasource.url=jdbc:mysql://localhost:3306/kylo
spring.datasource.username=
spring.datasource.password=
spring.datasource.maxActive=10
spring.datasource.validationQuery=SELECT 1
spring.datasource.testOnBorrow=true
spring.datasource.driverClassName=org.mariadb.jdbc.Driver
spring.jpa.database-platform=org.hibernate.dialect.MySQL5InnoDBDialect
spring.jpa.open-in-view=true
#
#Postgres datasource configuration
#
#spring.datasource.url=jdbc:postgresql://localhost:5432/pipeline_db
#spring.datasource.driverClassName=org.postgresql.Driver
#spring.datasource.username=
#spring.datasource.password=
#spring.jpa.database-platform=org.hibernate.dialect.PostgreSQLDialect
#
# Used just for the SimpleAuthenticationService
#
authenticationService.username=
authenticationService.password=
###Ambari Services Check
#ambariRestClientConfig.username=
#ambariRestClientConfig.password=
#ambariRestClientConfig.serverUrl=http://127.0.0.1:8080/api/v1
##ambari.services.status=HIVE/HIVE_CLIENT,HDFS
#ambari.services.status=HDFS,HIVE,MAPREDUCE2,SQOOP
#hive.datasource.url=jdbc:hive2://127.0.0.1:10000/default
#hive.datasource.username=
#hive.datasource.password=
###Cloudera Services Check
clouderaRestClientConfig.username=
clouderaRestClientConfig.password=
clouderaRestClientConfig.serverUrl=127.0.0.1
cloudera.services.status=HDFS/[DATANODE,NAMENODE,SECONDARYNAMENODE],HIVE/[HIVEMETASTORE,HIVESERVER2],YARN,SQOOP
#
# Server port
#
server.port=8425
#
# General configuration - Note: Supported configurations include STANDALONE, BUFFER_NODE_ONLY, BUFFER_NODE, EDGE_NODE
#
application.mode=STANDALONE
#
# Prevents execution of jobs at startup. Change to true, and the name of the job that should
# be run at startup if we want that behavior
#
spring.batch.job.enabled=false
spring.batch.job.names=
#spring.jpa.show-sql=true
#spring.jpa.hibernate.ddl-auto=validate
metadata.datasource.driverClassName=org.mariadb.jdbc.Driver
metadata.datasource.url=jdbc:mysql://localhost:3306/kylo
metadata.datasource.username=
metadata.datasource.password=
metadata.datasource.validationQuery=SELECT 1
metadata.datasource.testOnBorrow=true
hive.datasource.driverClassName=org.apache.hive.jdbc.HiveDriver
hive.datasource.url=jdbc:hive2://localhost:10000/default
hive.datasource.username=
hive.datasource.password=
hive.datasource.validationQuery=SELECT 1
hive.datasource.testOnBorrow=true
hive.metastore.datasource.driverClassName=org.mariadb.jdbc.Driver
hive.metastore.datasource.url=jdbc:mysql://localhost:3306/metastore
hive.metastore.datasource.username=
hive.metastore.datasource.password=
hive.metastore.validationQuery=SELECT 1
hive.metastore.testOnBorrow=true
nifi.rest.host=localhost
nifi.rest.port=8079
elasticsearch.host=localhost
elasticsearch.port=9300
elasticsearch.clustername=demo-cluster
## used to map Nifi Controller Service connections to the User Interface
## naming convention for the property is nifi.service.NIFI_CONTROLLER_SERVICE_NAME.NIFI_PROPERTY_NAME
##anything prefixed with nifi.service will be used by the UI. Replace Spaces with underscores and make it lowercase.
nifi.service.mysql.password=
nifi.service.example_mysql_connection_pool.password=
jms.activemq.broker.url=tcp://localhost:61616
jms.client.id=thinkbig.feedmgr
## nifi Property override with static defaults
##Below are Cloudera configuration options for Hive Metastore and Spark location
##Static property override supports 2 usecases
# 1) store properties in the file starting with the prefix defined in the "PropertyExpressionResolver class" default = config.
# 2) store properties in the file starting with "nifi.<PROCESSORTYPE>.<PROPERTY_KEY> where PROCESSORTYPE and PROPERTY_KEY are all lowercase and the spaces are substituted with underscore
config.hive.schema=metastore
nifi.executesparkjob.sparkhome=/usr/lib/spark
|
Q:
Get Woocommerce product tag post count via a shortcode
I need to count the number of times a "product tag" is used across the website. Each product has a number of tags associated with it.
I thought about creating a shortcode which I can then reference when needed. The shortcode I have created below crashes the website.
// function
function tag_count_shortcode() {
// Identify the tag ID and then run a count.
$term = get_tag( $tag_ID );
$count = $term->count;
// Output the total number of times a tag is used
return $count;
}
// register shortcode
add_shortcode('tagcount', 'tag_count_shortcode');
Im not sure where I am going wrong with this. Would really appreciate any assistance.
Platform: WordPress |
File with code: "Functions.php"
Cheers
A:
This is about product tags which is a WooCommerce custom taxonomy and not WordPress Tags.
Also a product can have many product tags, so the following code will handle the count on the first product tag term. This shortcode also handle some arguments:
the taxonomy (can handle any custom taxonomy, WordPress tag and category too) - By default: product tag
the term_id (can handle any defined term ID) - by default it gets the term on single product pages
the post_id - By default the current product ID
The code:
function term_count_shortcode( $atts ) {
extract( shortcode_atts( array(
'taxonomy' => 'product_tag', // Product tag taxonomy (by default)
'term_id' => 0,
'post_id' => get_queried_object_id(), // The current post ID
), $atts ) );
// For a defined term ID
if( $term_id > 0 ) {
// Get the WP_term object
$term = get_term_by( 'id', $term_id, $taxonomy );
if( is_a( $term, 'WP_Term' ) )
$count = $term->count;
else
$count = 0;
}
// On product single pages
elseif ( is_product() && $term_id == 0 && $post_id > 0 ) {
// Get the product tag post terms
$terms = get_the_terms( $post_id, $taxonomy );
// Get the first term in the array
$term = is_array($terms) ? reset( $terms ) : '';
if( is_a( $term, 'WP_Term' ) )
$count = $term->count;
else
$count = 0;
} else {
$count = false;
}
return $count;
}
add_shortcode('term_count', 'term_count_shortcode');
Code goes in function.php file of your active child theme (or active theme). Tested and works.
USAGE:
1). Basic usage: Display the first product tag count from product single page: [term_count]
2). Usage with arguments:
With a defined term ID: [term_count term_id="58"]
With a defined term ID and Taxonomy: [term_count term_id="15" taxonomy="product_cat"]
With a defined term ID and post ID: [term_count term_id="15" post_id="37"]
|
Q:
¿Cómo insertar varios registros a una misma laravel y laravelcollective?
Quiero ingresar en una misma tabla varios registros, para ello utilizo este código en la vista:
<h3>Resultados</h3>
<div class="form-group">
{{Form::label('descripcion_resultado', 'Resultado 1')}}
{{Form::text('descripcion_resultado', null, ['class' => 'form-control'])}}
</div>
<div class="form-group">
{{Form::label('descripcion_resultado', 'Resultado 2')}}
{{Form::text('descripcion_resultado', null, ['class' => 'form-control'])}}
</div>
este es mi controlador:
public function store(Request $request){
$resultado = Resultado::create($request->all());
return 'resultados añadidos con éxito';
}
Pero el problema está en que no me guarda los dos registro si no que guarda solo uno (el ultimo)
A:
Utiliza un array en el nombre del campo, y luego iteras sobre esta, algo así:
<div class="form-group">
{{Form::label('descripcion_resultado[]', 'Resultado 1')}}
{{Form::text('descripcion_resultado[]', null, ['class' => 'form-control'])}}
</div>
<div class="form-group">
{{Form::label('descripcion_resultado[]', 'Resultado 2')}}
{{Form::text('descripcion_resultado[]', null, ['class' => 'form-control'])}}
</div>
En el controlador utilizas el método create():
public function store(Request $request){
foreach ($request->description_resultado as $resultado) {
Resultado::create(['description_resultado' => $resultado]);
}
return 'resultados añadidos con éxito';
}
|
Cara Greenberg, a notorious writer and art historian, was the one who created the expression “ Mid Century Modern ”, in 1983. It was in the title of one of her most relevant books (“Mid-Century Modern: Furniture of the 1950s”), which reflected what is now recognized as a global and iconic design movement. The movement spread to fields such as interior, product and graphic design, architecture and urban development, from about 1933 to 1965 and it is a style recognized today by scholars and museums.
The label is straightforward and easy to understand, as candid as the interior style it describes, which defended concepts of functionality, comfort and modern simplicity. When verbally spoken, the sounds that compose the words “Mid-century Modern” also have a satisfying melodic value to them, thanks to the mirrored, doubling ‘Ms’ which roll smoothly off the tongue, evoking the clean sculptural lines of a perfectly balanced aesthetic.
Bubble shapes, immaculate proportions and appealing colors made mid-century be appropriately defined as “furniture candy”. Nowadays, chic vintage stores and the mid-century modern furniture fairs bring back this historic movement which prevails in our sense of what contemporary means. This is a movement from the past that continues to adorn our homes in the present, specially our kitchens and living rooms.
Origin
Mid-century Modern is a furniture and interior design style which was a trend from the 1940s-1960s, being the natural descendant of Modernism. It was rooted in philosophies of functionality, elegance and simplicity, following the words of the Bauhaus and Le Corbusier which declared: “a house is a machine for living”.
George Nelson’s “Platform Bench”(1946) is often mentioned as one of the first notable designs of the genre. It was created to be mass-produced in order to be affordable to the average homeowner, an ideal Nelson inherited from the Bauhaus mindset: a good design must be for everyone.
After Nelson’s casual, simple yet stylish bench, the movement continued to promote the romantic idea that good design could change everyone’s lives, not just the wealthy people’s ones. Design could change the entire world and improve it.
Ray and Charles Eames, a designer couple, created lovely Californian chairs which are now symbols of the movement. Their ideology was simple but powerful: “Getting the most of the best to the greatest number of people for the least amount of money.”
Materials
Mid-century designers utilized materials according to their own specific, sometimes even artificial, qualities and never to imitate natural materials like wood or marble. They embraced new materials such as metal, glass, vinyl, and plywood, combining them with wood to create innovative and stimulating compositions. Often a piece of furniture would include only two materials or two colors, creating tension and harmony without any superfluous adornments.
The artist and architect Isamu Noguchi’s table made of walnut hardwood and glass demonstrates this: it was described as a “functional sculpture”, and its duality of two elements created something graceful and useful at the same time.
The Eames couple also mastered new technologies, creating plastic resin and wire mesh chairs that were produced by the manufacturer Herman Miller. Their “Molded Plastic and Fiberglass Armchair” was low-cost and had good quality, mixing several elements. The customer could choose from three plastic colors (elegant greige, elephant-hide, or parchment) and could select a metal, wood or rocker base. Materials that were once associated with cheap and industrial objects were becoming the standart in contemporary interior design.
Aesthetic
George Nelson was the one who determined what were three main Mid Century Modern categories: the biomorphic; the machine; and the handcrafted.
Biomorphic Mid Century Modern is a term used to define furniture with organic, curved, smooth surfaces, often with designs molded into the shapes of kidneys and boomerangs. This furniture was more enjoyable and energetic than the harsh machine aesthetic of the Bauhaus.
Eva Zeisel, a talented artist born in Hungary and based in the United States, explored the natural world in her designs. Her famous salt and pepper shakers radiate personality, and her belly-button shaped room-divider combined human body forms with efficient division of space. Other biomorphic classics include the dreamy, cloud-like Eames’s “Plastic Chaise Longue” and Verner Panton’s elegant “Panton Chair”.
The machine Mid Century Modern style was created in Bauhaus. It is based on blunt, high-tech geometric forms that play with form and function. George Nelson’s “Associate’s Ball Clock” for the Howard Miller clock company is a great example. At the time of its birth it looked as if it had fallen out of the 21st century, although now it feels vintage.
Despite its designation, the third category of “handmade” Mid Century Modern Furniture was also designed for industrial production, and it expressed ideals of utility and minimalism in the sculptural lines of shaped wood. This style is mostly associated with the Danish vision of the movement. Denmark’s Finn Juhl was a pioneer, his “45 Chair” is a true classic, a perfect combination of graceful leather and finely sculptured wood that looks light and delicate.
Mid-century Modern Furniture in Our Days
Despite the passing of time, this furniture designs from the past that tried to predict the future are still being made by manufacturers such as Herman Miller and Knoll. Eames’ “Lounge Chair” has never even gone out of production since it was first released. However, ironically for an artistic movement that prided itself on its accessibility, many of its iconic pieces, such as Eames’ “Molded Plywood Folding Screen” or the “Marshmallow Sofa” by George Nelson, are now being sold for extremely high prices.
Mid-century Modern in Scandinavian Design
Scandinavian design was very influential in Mid-century Modern Furniture, with a unique style characterized by simplicity, design “for everyone” and natural silhouettes. Glassware (Iittala – Finland), ceramics (Arabia – Finland), tableware (Georg Jensen – Denmark), lighting (Poul Henningsen – Denmark), and furniture (Danish modern) were some of the product designs affected by this influence.
Danish Modern
Danish modern is a style of minimalist furniture and housewares from Denmark associated with the Danish design movement. In the 1920s, Kaare Klint embraced the principles of Bauhaus modernism in furniture design, creating clean lines based on an understanding of classical furniture and craftsmanship united with a careful research of materials, proportions and the necessities of the human body.
With designers such as Arne Jacobsen and Hans Wegner and associated cabinet makers, Danish furniture flourished from the 1940s through the 1960s. Adopting mass-production techniques and concentrating on form rather than just function, Finn Juhl also contributed to the style’s success. Danish housewares also adopted a similar simple design such as those produced in Denmark for Dansk in its early years, expanding the Danish modern aesthetic beyond furniture.
Incredible Designers of Danish Modern Style
Kaare Klint (1888-1954)
Thanks to the furniture design school he founded at the Royal Academy in 1924, Klint had a strong influence on Danish furniture, shaping designers such as Kjærholm and Mogensen. His wisely studied designs are based on functionality, proportions in line with the human body, craftsmanship and the use of high-quality materials. Noteworthy examples of his work include the “Propeller Stool” (1927), the “Safari Chair” and the “Deck Chair” (both from 1933), and the “Church Chair” (1936).
Poul Henningsen (1894-1967)
Poul Henningsen, an architect with a strong belief in the functionalist ideology, was a vital member of the Danish Modern school, not for furniture but for lighting design. His attempt to prevent the blinding glare of the electric lamp bulb succeeded in 1926 with the creation of a three-shade lamp, known as the “PH lamp”. The curvature of the shades allowed his hanging lamp to illuminate both the table and the rest of the room. He went on to design many similar lamps, some with frosted glass, including desk lamps, chandeliers and wall-mounted fixtures. Though he died in 1967, many of his designs have remained popular to this day.
Mogens Lassen (1901-1987)
In addition to his architectural work, Lassen was also a talented furniture designer. Influenced both by Le Corbusier and Ludwig Mies van der Rohe, he developed a unique approach to Functionalism. Thanks to his fine craftsmanship and his search for simplicity, his steel-based furniture from the 1930s added a new dimension to the modernist movement. His later designs in wood are still part of classical Danish Mid Century Modern, especially his three-legged stool and folding “Egyptian coffee table” (1940) originally produced by A. J. Iversen.
Arne Jacobsen (1902-1971)
Jacobsen graduated from the Royal Academy in 1924 and quickly proved his mastery in both architecture and furniture design. With the accomplishment of his Royal Hotel in Copenhagen and all its internal furniture in 1960, his talents became widely recognized, especially due to his chairs “Egg” and “Swan”, now considered universal icons. His stackable, three-legged “Ant Chair” (1952) with a one-piece plywood seat and back and its four-legged counterpart, the “7 Chair” (1955), were also particularly popular all around the world.
Ole Wanscher (1903-1985)
Inspired by Kaare Klint, Wanscher later followed his footsteps, becoming a professor of the Royal Academy’s furniture school. Particularly interested in 18th-century English furniture and in early Egyptian furniture, one of his most successful pieces was his carefully designed “Egyptian Stool” (1960) crafted from exquisite materials. Another successful work was his “Colonial Chair” in Brazilian rosewood. He was awarded the Grand Prix for furniture at Milan’s triennale in 1960.
Finn Juhl (1912-1989)
Although he studied architecture at the Royal Academy, Juhl was a self-taught designer as far as furniture is concerned. In the late 1930s, he created furniture for himself but from 1945 on he started to become notorious for his sculptural designs, placing an emphasis on form rather than function, breaking the tradition of Klint’s school. His successful interior design project at the UN Headquarters in New York spread Danish Modern, paving the way for the international recognition of his colleagues. Two key pieces of his furniture, in which the seat and backrest are separated from the wooden frame, are “45-Chair”, with its elegant armrests, and “Chieftain Chair” (1949).
Børge Mogensen (1914-1972)
After studying at the Copenhagen School of Arts and Crafts and at the Royal Danish Academy of Fine Arts, Mogensen adopted Klint’s approach to simple, functional furniture design. Taking an almost scientific approach to an object’s utility, most of his furniture is characterized by strong, simple lines and was designed for industrial production. Famous pieces by him include oak-framed “Hunting Chair” (1950) with a strong leather back and seat, his light “Spokeback Sofa” (1945), and the low robust “Spanish Chair” (1959).
Hans Wegner (1914-2007)
Known as the “Master of Chairs”, Wegner created charming furniture with clean, organic and aesthetic lines, balanced by a minimalist and serene aspect. He was a modernist with emphasis on the practicality and elegance of each piece he crafted. He believed the versatility and usability of his designs were as vital for him as their aesthetics.
After graduating in architecture in 1938, he worked in Arne Jacobsen and Eric Møller’s office before establishing his own office in 1943. Striving for functionality as well as beauty, he became the most prolific Danish designer, producing over 500 different chairs. His “Round Chair” (technically Model 500) in 1949 was called “the world’s most beautiful chair” before being labelled simply “The Chair”. His “Wishbone Chair”, also from 1949, with a Y-shaped back split and a curved back, was inspired by a Chinese child’s chair. Wegner’s designs can now be found in several of the world’s best design museums including New York’s Museum of Modern Art.
Grete Jalk (1920-2006)
After training as a cabinetmaker, she studied at the Danish Design School in 1946, while receiving additional instruction from Kaare Klint at the Royal Academy’s Furniture School. Inspired by Alvar Aalto’s laminated bent-plywood furniture and Charles Eames’ molded plywood designs, she began to develop her own boldly curved models in the 1950s. In 1963, she won a Daily Mirror competition with her “He Chair” and “She Chair”. With the help of the furniture manufacturer Poul Jeppesen, she designed simpler models with comfortable lines, which became popular both in Denmark and the United States thanks to their competitive prices. Jalk also edited the Danish design magazine Mobilia and compiled a four-volume work on Danish furniture.
Verner Panton (1926-1998)
After graduating from the Royal Academy in 1951, Panton worked briefly with Arne Jacobsen. During the 1960s, he designed furniture, lamps and textiles with an imaginative combination of innovative materials, playful shapes and bold colors. Among his earliest designs were the “Bachelor Chair” and “Tivoli Chair” (1955), both produced by Fritz Hansen. His “Panton Chair” (1960) was the world’s first one-piece molded plastic chair. Sometimes seen as a pop artist, unlike most of his generation, he continued to be successful in the 1970s, not only with furniture but also with interior designs, including lighting.
Poul Kjærholm (1929-1980)
In addition to an academic career at the School of Arts and Crafts and at the Institute of Design at the Royal Academy, Kjærholm always understood the importance a piece of furniture has in its surrounding architectural space. Functionality was the second most important thing in his artistic approach, which was centered on sophisticatedly clean lines and attention to detail. Unlike many of his contemporaries, he worked essentially with steel, combining it with wood, leather, cane or marble.
Kjærhom developed a close relationship with the cabinetmaker E. Kold Christensen who produced most of his designs. Today a wide selection of his furniture is produced by Fritz Hansen. Kjærholm’s work can be seen in New York’s Museum of Modern Art and the Victoria and Albert Museum in London.
Jens Risom (1916-2016)
Often credited with having introduced Danish Mid Century Modern design to America, Risom was a graduate of Copenhagen School of Industrial Arts and Design. He emigrated to the United States in 1939 to study American design, working first as a textile designer and later as a freelance furniture designer. In 1941, he joined Hans Knoll at the Hans Knoll Furniture Company, and together they explored the country promoting Risom’s designs. A true minimalist, Risom worked mainly in wood because it was cheap, and one of his most successful pieces, “Knoll Chair #654” was made with a seat of nylon webbing that was discarded by the army.
Stay with us to find out more about Mid Century Modern, the best of contemporary art, design and craftsmanship.
See More Related Stories
Patron of the Arts: A Figure Essential to Every Page of Art’s History
Contemporary Design: Best Designers and Architects in the World
|
Women Clarks Beige Leather Wedges - Clarks Women Heels F95x5616
Size
Description
Steal the show wherever you go by wearing this pair of wedges from the house of Clarks. We suggest that you team it up with a pair of jeans or treggings and a tunic to get ready for a weekend party.Product Details:A pair of open-toed beige wedges with yellow and orange accents, has slip-on stylingLeather upper with stitched and cut-out detail, has crisscross straps on the forefoot, elasticated gusset on the midfoot strapSoft footbed with stitched and perforated detail on the forefoot region and branding on the heelTextured outsole with branding, has a wedge heel with textured detail on the sides
|
A density-functional-theory study of atomic nitrogen abstraction from Si(100)-(2 x 1) by a gaseous O(3P) atom.
We have employed density-functional theory (DFT) to investigate the abstraction of a nitrogen atom from the Si(100)-(2 x 1) surface by a gas-phase O(3P) atom for different initial bonding configurations of nitrogen at the surface. For the N-Si(100) structures investigated, nitrogen abstraction by an O(3P) atom is predicted to be exothermic by at least 1.9 eV. Abstraction in a single elementary step is found only for the interaction of an O(3P) atom with nitrogen bound in a coordinatively saturated configuration, and an energy barrier of 0.20 eV is computed for this reaction. For nitrogen bound in coordinatively unsaturated configurations, abstraction is predicted to occur by precursor-mediated pathways in which the initial O-surface collision results in the formation of a N-O bond and the concomitant release of between 2.7 and 4.8 eV of energy into the surface, depending on the initial N-Si(100) structure. This initial step produces different surface structures containing an adsorbed NO species, which can then undergo a series of elementary steps leading to NO desorption. Since the barriers for these steps are found to be less than 1 eV in all cases, a significant excess of energy is available from initial N-O bond formation that could activate NO desorption within no more than a few vibrational periods after the initial gas-surface collision. Nitrogen abstraction by such a pathway is essentially an Eley-Rideal process since NO desorption occurs rapidly after the initial gas-surface collision, without the reactants thermally accommodating with the surface. These computational results indicate that nitrogen abstraction by gaseous O(3P) atoms should be facile, even at low surface temperatures, if nitrogen is bound to the Si(100) surface in coordinatively unsaturated configurations.
|
Additional Photos
Raising the bar: Rachel Hanson, 10, top bar, practices the uneven bars at Decal Gymnastics in Farmington on Saturday. Decal re-opened today after closing because of flooding. Staff photo by Michael G. Seamans
Glad to be back: Abby Drummond, 10, performs a back flip as she practices her vault routine at Decal Gymnastics in Farmington on Saturday. Decal re-opened today after closing because of flooding. Staff photo by Michael G. Seamans
Hanson and Parker, who said she’s had her own bouts with bullies, are two of the 150 gymnastics students who were devastated to learn that Decal was being closed for repairs a month ago, when a thaw followed by subzero temperatures caused the roof to leak and flooded the gymnasium with more than an inch of water.
The loss temporarily broke up a community of young people who look to each other for the friendship and camaraderie they can’t always find elsewhere.
On Saturday, Decal’s co-owners, 28-year-old Delani Evans and her mother, Carol Hamilton, held an open house to celebrate the grand reopening.
Hamilton estimates that the ordeal cost the business about $7,000, partly in lost revenue and partly in ruined equipment, with some of the sodden mats having to be replaced at a cost of as much as $1,000 apiece.
And for the girls and boys who look forward to going to gymnastics classes together, the cost was emotional.
“I was sad,” Hanson said. “This is like my home. I made a lot of friends here.”
Her father, Dennis Hanson, who watched without a hint of apprehension as his daughter hurled herself into the air from a crossbar, said gymnastics has provided her with an opportunity to grow strong.
“It’s self confidence,” he said. “She is really committed to this. She broke her arm two years ago and still made the team.”
Parker, who holds her blue-framed glasses on with a strap during gymnastics, said her skills have given her a reason to be confident even when others try to tear her down.
“I just shrug it off,” she said. “I can do things they can’t. It makes me feel unique.”
She’s learning how to do a sequence involving a roundoff, a backhand spring and a backtuck.
Evans said many of the students are like Hanson and Parker.
“We’re taking ordinary kids and teaching them something that not everyone can do,” she said. We’re taking average joes, kids getting picked on in school who come here to find a home.”
Aside from the physical benefits — gymnastics teaches kids power, core strength, coordination and balance — the very nature of the sport teaches children something even more important, Evans said.
“We do a lot of failing here,” she said. “I tell the kids, you’re not going to land on your feet the first time you try a move. It will take you a thousand tries before you land on your feet.”
Hamilton said the unexpected closure might also have long-term costs for the business. She said it remains to be seen whether all of the students who were attending classes in December will return after the monthlong hiatus.
After weeks of worry and hard work to restore the gymnasium, Hamilton said Saturday’s open house, which drew about 50 students in the first two hours, was a good sign that the small business would be able, like a gymnast performing a difficult maneuver, to land on its feet.
That level of resolve to bounce back after a setback is a lesson she hopes the students will pick up.
“We’ve seen a lot of new faces,” she said. “It’s like we try to teach them. We took something really bad and are making it work for us.”
Here at MaineToday Media we value our readers and are committed to growing our community by encouraging you to add to the discussion.
To ensure conscientious dialogue we have implemented a strict no-bullying policy. To participate, you must follow our Terms of Use. Click here to flag and report a comment that violates our terms of use.
|
Couple, Daughter Found Dead In South Delhi Home; Injured Son Questioned
NDTVPublished on Oct 10, 2018
Three members of a family were found dead at their home in South Delhi's Vasant Kunj this morning. Mithilesh and Seeya, a couple in their 40s, and their 16-year-old daughter Neha were reportedly found with stab wounds. Their 19-year-old son was found with minor injuries. The incident occurred in a building in Kishangarh area around 5 am. The family's domestic help alerted neighbours after she found the bodies.
|
Q:
regarding java version 1.3 to 1.4
I am using this Java version:
java version "1.3.1_01" Java(TM) 2 Runtime Environment, Standard Edition (build 1.3.1_01) Java HotSpot(TM) Client VM (build 1.3.1_01, mixed mode)`
But I have written an application and I am getting run time error
Exception in thread "main" java.lang.NoClassDefFoundError: java/lang/CharSequence
CharSequence only exists since 1.4.
How can I overcome this problem?
A:
Sounds like you have some 1.4 code that you're trying to run on 1.3.
You're making the classic mistake of assuming that your assumptions are correct. Check your assumptions.
Both JVMs are well past the end of their support lives. Perhaps you should consider upgrading. JDK 6 is the current standard.
|
Pages
FLickr Friends: Ashley and Emily from Frances Suzanne
Today we have two ladies we've been smitten with since their first entry into the flickr group. And I'm sure after you see their clean, polished style you'll know why. Here they are----Emily and Ashley from Frances Suzanne !
Hello Project Run and Play readers! We are absolutely thrilled that liZ and Elizabeth invited us to be here today! Two seasons ago, we participated in our first pr&p sew-along with the flickr group. The camaraderie and sewing support 'hooked us' from the onset.
Unlike previous guest bloggers, we are "sewing aunts," rather than "sewing mommas." We are two sisters, Ashley {frances} and Emily {suzanne} who love creating and sewing children's clothing for our two nieces {our 'baby' sister's children}.
When trying to determine our tutorial focus, we decided why not go with what we love? Accessorizing with PIPING!
And, since there was plenty of white sheet to spare from season 5, white sheet challenge (here and here), we put it to use! Piping is probably 'old news' to a lot of you, but to add intrigue, we opted to use multiple piping layers to add effect on an outfit.
We used this technique in one of our flickr entries, and a friend asked, "How did you make double piping?" Hence, the tutorial idea was conceived. Then, for a bit of a twist, and since we were doing multiple layers, we thought: why not make ombre piping?Since Christmas is quickly approaching, we decided use the opportunity to make one of our niece's Christmas outfits. *Although in order to do this, we broke our "standing rule" of not wearing Christmas clothes before Thanksgiving albeit only to make pictures.
Truth be told, pattern drafting is practically foreign to us at this point, so we chose to start with a pattern and modify it, although the ombre piping technique could be used on practically anything. We began with the Oliver + SAfter School Shirt, which we lengthened to dress length (see figure below for measurements). We also added 1" to the sleeve length (right above the notch) to accommodate for piping seams.
This tutorial eliminates the need to piece individual bias strips together, which to us was just lovely. We chose to cut our strips 1" wide, and then ironed them in half.
Next, we inserted "baby piping cord" into the fold. Of course, you could use any size cording that you prefer.
Then we stitched as close to the cording as we could, so we would end up with "fat" piping, rather than "flat" piping.
Fortunately, we have a handy sewing foot that makes this all possible. It has a small groove where the cording feeds almost effortlessly.
Our goal was to make a piped ombre effect going from lighter to darker down the dress. So, we incorporated the lightest colored piping around the entire yoke. Here are some pictures to give you a general idea, but you can find a very detailed tutorial here.
After we had made the single piping cords of all three colors and completed the above-mentioned steps based on the pattern instructions and tutorials, we were ready to get down to the real fun: ombre piping.
Starting with the sleeves, we lengthened them 1" to accommodate 1/2" seam allowances on either side of the cut for the piping. The inch was added above the notch. We had decided we wanted the darker color to be at the edge of the sleeve and the hem of the dress. In order to accomplish this, we stacked and sewed all three piping ombre pieces to make one piece of piping. We first sewed the medium green piping on top of the darkest green piping.
Once those two were attached, we added the lightest green piping on the top of them.
*Ignore the clipping of the seam....this was left over from the yoke piping....
Next, we cut the sleeve 1/2" above the notch, and pinned the piping to the right side of the main part of the sleeve, and stitched on top of the piping stitch closest to the sleeve's edge.
Finally, we placed the end of the sleeve on top of the piping - right sides together (piping sandwiched in between) - and sewed on top of the stitching line that we just sewed. You will possibly have to flip the sleeve to the opposite side to see this.
Then, we referenced the pattern instructions to finish the sleeves.
After attaching the sleeves, we seamed up ONE of the side seams. This allowed piping to be attached at the bottom of the dress in one continuous line.We determined that we wanted to have three tiers of piping at the base of the dress. The first tier would consist of a single piping cord made of the lightest green piping. Then, we would move down the dress 1.5" and have double piping made of the lightest green piping on the top and the medium shade of piping on the bottom. Finally, on the third tier, we would move down the dress an additional 2.25" and have triple piping made up of two medium green piping cords and one of the darkest green piping. We would finish it off with a bias strip of the darkest shade of green.
To prepare for this process, we sewed each tier together as one piece.Then, we cut 5.75" from the bottom of the dress, and began with the top tier.
We laid the single piping cord on top of the main dress, right sides together, and stitched.
Next, we added the bottom part of the dress right sides together (sandwiching piping in between), and stitched following the previous stitched line. Although some brave souls might be able to do this step in one fatal swoop...not us. Then, we ironed the seam down toward the bottom of the dress. Do NOT leave this step out! The more you iron the happier your dress will be. We then edgestitched on the right side of the dress (under the piping) to allow the seam to lay flat against the dress.
For the second tier, we measured 3.25" from the bottom of the main dress, and cut.
We went through the same process as above, only this time we were adding double piping and sewed on top of the piping stitch line closest to the bottom of the dress.
After everything was attached back together (by sandwiching the attached piping and main dress with the piece you cut off), we pressed our seam up this time, and edgestitched.For the third and final tier. . .we measured up 2.75" from the bottom of the dress and cut one final time.
This time, however the triple piping layer was sandwiched between the main dress and dark green bias band (ours was cut 7" to accommodate for 1/2" piping seam, 5/8" band, and hem facing to cover ALL piping seams).
We created a bias band by ironing 5/8" from final piping. The folded hem facing easily fell above the first tier of piping. Then, we edgestitched the hem facing to the main dress directly above the first tier to complete the ombre effect.
Is this a difficult way to add a little "pop" to an outfit? Absolutely not. Trust us, if we can do it, anyone can!
Is it time consuming? Yes, definitely.
Was it worth it? We'd like to think yes, and one precious little girl seems to think so too!
We couldn't stop with ombre piping though....we had to add just a tiny bit of handstitching to finish the dress off! The hardest part: keeping it simple, and not 'overdoing' the simplicity of the outfit. So many ideas....so hard to simplify!
We added a Christmas tree above the 1st tier (lightest color) of piping.
The tree was stitched with a straight stitch of medium green thread, and the trunk was stitched with a light taupe color. Then, red French knots were added on the tree to accentuate the red buttons.
Finally, a few red French knots were added above the ombre sleeve openings.
And the finished product was...........
Now, we'll be the first to admit, our wheels are turning with what we might do next because the possibilities are endless (thicker cording, various color combinations, vertical piping, etc)! But first up will be a Christmas / winter outfit for her little sis, so we'd love for you to tune into our blog to see what we have up our sleeves!
Thank you again liZ and Elizabeth for having us today, it truly has been fun!
I love this simply little dress with the beautiful details. I enjoyed reading your tutorial too. I've never thought about making triple piping, and the effect of light to dark was perfect. Thanks for sharing your techniques with us, and the pictures made it so understandable.Deborah
|
Many machines are mobile machines configured to perform one or more tasks while traveling over a work surface, such as a road surface. A cold planer is an example of such a mobile machine. The cold planer may include a grinding mechanism such as a rotor that grinds a top layer of the work surface as the cold planer moves in a machine travel direction. The cold planer may include a conveyor connected to a frame of the cold planer that receives the material that is ground from the road surface by the rotor. The conveyor may convey the material to another vehicle, such as a dump truck, traveling next to the cold planer.
The rotor may be surrounded by a rotor housing to contain ground material from the work surface so that the material may be removed by the conveyor. A portion of the rotor extends below the rotor housing so that it can dig into the work surface. The rotor housing may include side plates that extend below the rotor housing to further retain the material ground out of the work surface by the rotor. In known cold planers, the side plate may lie on the work surface under the force of gravity and be dragged by the cold planer in a floating condition. An example of an alternative side plate arrangement is provided in U.S. Patent Appl. Publ. No. 2013/0082508 published on Apr. 4, 2013 to Orefice. In the published application, a side plate arrangement for a milling device includes a milling roller box arranged on a frame of the milling device, a side plate whose height can be adjusted and a side plate support having a swivel bearing with a swivel axis around which the side plate can swivel against the frame in a swivel area. One element of the milling device is a guide curve running concentrically to the swivel axis, with the help of which the swivel movement of the side plate around the swivel axis is carried out. Height adjustment of the side plate occurs actively through a lifting device linked to the side plate, specifically through a one cylinder piston unit. After being positioned at the desired height, the side plate skids along a work surface and rotates about a swivel axis as the milling device passes over an uneven work surface.
|
Q:
basic linear algebra question, proving it is a diagonal matrix and scalar matrix.
Let A = $(a_{ij})_{n\times n}$ such that $AB = BA$ for every square matrix $B$ of order $n$.
$(I)$ Prove that $A$ is a diagonal matrix.
[Hint: Let $B_i$ denote the matrix whose $(i,i)$ entry is $1$, and $0$ elsewhere. Show
that $a_{ij} = 0$ whenever $i ≠ j$.]
$(II)$ Prove that $A$ is a scalar matrix.
[Hint: For $i ≠ j$, let $B_{ij}$ denote the matrix whose $(i,j)$-entry is $1$ and $0$ elsewhere.Show that, $a_{ii} = a_{jj}$ for all $i ≠ j$.]
Can someone help me with this question using basic linear algebra concepts? Thanks.
A:
i) You have the matrix $(B_i)_{kl}=\delta_{ik}\delta_{il}$, where $\delta_{rs}$ is the matrix with only $1$'s in the diagonal and the rest zeros.
The multiplication $AB_i$ is then
$$(AB_i)_{rs}\,=\,\sum_{r}A_{rk}(B_i)_{ks}\,=\,\sum_{k}A_{rk}\,\delta_{ik}\delta_{is}\,=\,A_{ri}\delta_{is}$$
Thus $AB_i$ is the matrix with all zeros except for its $s=i$-th column which contains the $i$-th column of $A$ -given by $A_{ri}$ where $r=1,\cdots,n$.
Analogously the multiplication $B_iA$ gives
$$(B_iA)_{rs}\,=\,\sum_{k}(B_i)_{rk}A_{ks}\,=\,\sum_{k}\,\delta_{ir}\delta_{ik}\,A_{ks}\,=\,A_{is}\delta_{ir}$$
This, however, is is the matrix with all zeros except for its $r=i$-th row which contains the $i$-th row of $A$ -given by $A_{is}$ where $s=1,\cdots,n$.
From the arrangement of both results, you can readily infer the conclusion sought. Indeed, a "row matrix" (by that I mean it's all zeros except within a given row) can never be equal to a "column matrix" (analogous but within a column) unless all values of the row in one matrix and the column in the other are zero, except maybe the diagonal elements in each case. This means
$$A_{ri}=0 \; (r\neq i)$$
ii) Let's consider now the matrix $B$ whose entries $k,l$ are $(B_{ij})_{kl}=\delta_{ki}\delta_{jl}$. Proceeding as before we see that $(AB_{ij})_{rs}=A_{ri}\delta_{js}$ and $(B_{ij}A)_{rs}=\delta_{ri}A_{js}$. Equating this two results, and considering that the deltas bound the values of $r$ and $s$ to $i$ and $j$, respectively, leads to $A_{ii}=A_{jj}$.
We can call that number $a$. Thus we have proven that $A=a\mathbb{1}$ a multiple of the identity matrix.
|
In a gas-phase polymerization process, traditional Ziegler-Natta catalysts have been shown to readily polymerize ethylene with one or more higher alpha-olefin comonomers such as propylene, 1-butene, 1-pentene, 1-hexene, 4-methyl-1-pentene, 1-octene and 3,5,5-trimethyl-hexene, to produce a linear low-density polyethylene (LLDPE) with minimal if any long-chain branching (LCB). The increasing development of metallocene catalyzed olefin polymers has resulted in the ability to produce similar polymers with a more well-defined molecular structure than can be achieved with conventional Ziegler-Natta catalysts . Metallocene linear low density polyethylenes made according to U.S. Pat. Nos. 5,420,220 and 5,324,800, for example, possess narrow comonomer and molecular weight distributions. U.S. Pat. Nos. 5,527,752 describes further a family of metallocene catalyst precursors which are useful, when combined with a cocatalyst or catalyst activator, in the manufacture of polyolefins.
The various linear low density polyethylenes referred to above are known to be useful, among a wide variety of applications, in those applications in which a crosslinkable polymer is requited.
In this vein, industry has been seeking polymers, which have improved crosslinking properties without sacrificing processability.
|
On the Abolition of All Political Parties (NYRB Classics)
Simone Weil—philosopher, activist, mystic—is some of the most uncompromising of contemporary non secular masters. In “On the Abolition of All Political events” she demanding situations the root of the fashionable liberal political order, making a controversy that has specific resonance this present day, while the apathy and anger of the folk and the self-serving partisanship of the political type current a danger to democracies world wide. Dissecting the dynamic of strength and propaganda because of get together spirit, the expanding overlook for fact in desire of opinion, and the ensuing corruption of schooling, journalism, and paintings, Weil forcefully makes the case precise politics can basically commence the place social gathering spirit ends.
This quantity additionally comprises an admiring portrait of Weil by means of the good poet Czeslaw Milosz and an essay approximately Weil’s friendship with Albert Camus by way of the translator Simon Leys.
Quick preview of On the Abolition of All Political Parties (NYRB Classics) PDF
Best Philosophy books
French celebrity thinker Luc Ferry encapsulates an enlightening treatise of pop-philosophy in a full of life narrative of Western thought—explaining how the background of philosophy can educate us how one can stay greater lives at the present time. This blockbuster larger dwelling advisor has already offered approximately a part one million copies in Europe; now it's on hand around the globe in an exhilarating, hugely readable translation: a short heritage of proposal.
What's it so you might be wide awake? there isn't any contract no matter what in philosophy or technological know-how: it has remained a troublesome challenge, a secret. is that this in part or almost always owed to the present theories now not even having an identical topic, no longer answering an identical query? In genuine recognition, Ted Honderich units out to supersede dualisms, goal physicalisms, summary functionalism, basic externalisms, and different positions within the debate.
This quantity provides an interconnected set of 16 essays, 4 of that are formerly unpublished, by way of Allan Gotthelf--one of the major specialists within the research of Aristotle's organic writings. Gotthelf addresses 3 major themes throughout Aristotle's 3 major organic treatises. beginning together with his personal ground-breaking research of Aristotle's typical teleology and its illuminating courting with the iteration of Animals, Gotthelf proceeds to the axiomatic constitution of organic clarification (and the 1st rules such rationalization proceeds from) within the elements of Animals.
There were notable advancements within the box of human rights long ago few many years. nonetheless, hundreds of thousands of asylum-seekers, refugees, and undocumented immigrants proceed to discover it demanding to entry human rights. during this e-book, Ayten Gündogdu builds on Hannah Arendt's research of statelessness and argues that those demanding situations display the perplexities of human rights.
Additional info for On the Abolition of All Political Parties (NYRB Classics)
It really is an increasing number of seen (in the nations of jap Europe in addition) that fridges and tv units, or perhaps rockets despatched to the moon, don't switch guy into God. previous conflicts among human teams were abolished yet are changed by way of new ones, might be extra acute. I translated the chosen works of Simone Weil into Polish in i958 now not simply because I pretended to be a 'Weil ian . ' I wrote frankly within the preface that I examine THE I M PO RT AN CE O F S I M O N E WE I L myself a Caliban, too fleshy, too heavy, to tackle the feathers of an Ariel.
Or 'As a Catholic, i feel that . . . a few little ladies, who declared they have been dedicate ted to Gaullism because the French similar of Hitlerism, further: 'Truth is relative, even in geometry. ' certainly, this is often the guts of the problem. If there have been no fact, it might be correct to imagine in such or any such method, whilst one occurs to be in such or this kind of place. simply as one's hair is black, brown, purple or blond simply because one occurred to be born that means, one can also show such or the sort of inspiration. proposal, like hair, is then the fabricated from a actual technique of removal.
39 -40. Milosz, The Captive brain. Secker & Warburg, London, i nine five three , p. 28. S I M ON L E Y S in a short time that serving a Stalinist regime could entail not just morally and intellectually unacceptable compromise, yet extra easily may impress down correct revulsion: 'A guy could convince himself by way of the main logical reasoning that he'll enormously gain his overall healthiness by means of swallowing dwell frogs; and therefore rationally confident, he might swallow a primary frog, then a moment, yet on the 3rd his belly will insurrection.
Her various political articles at the probabilities of the employees' strug gle in France, on monetary coverage, at the reasons of Nazism in Germany, in addition to her stories at the mechanism of society and at the background of Europe, were lately accumulated in a couple of volumes. just some of them were released in her lifetime, in little identified magazines. the need to proportion the destiny of the oppressed led her to a momentous determination. despite undesirable overall healthiness, she labored for a yr (1934-35) as an easy employee in Paris metallurgical factories; she therefore got a primary hand wisdom of guide labour.
S I M O N E WE I L This then quantities to idolatry, for God by myself is legitimately his personal finish. The transition is definitely completed. First, an axiom is determined: for the occasion to serve successfully the concept that of the general public curiosity that justifies its lifestyles, there's one priceless and adequate situation: it may safe an enormous quantity of strength. but, as soon as acquired, no finite volume of strength will ever be deemed enough. The absence of inspiration cre ates for the social gathering an enduring kingdom of impotence, which, in flip, is attributed to the inadequate quantity of strength already bought.
|
Suggestions on making a progress bar/waiting message
I've got this program which saves data into TXT files and it takes very long to do so. The problem is not how long it take though, the more annoying problem is that everything in my program is blocked until it's done saving and doesn't notif the user of what's going on.... What I mean is I want a message (i was thinking a little window) to appear where saying "Saving..." and disappearing when the program is done saving.
Which SWING class would the best to do this? I was thinking of using JOptionPane but those windows come with buttons... which is unnecessary
Must I create an extension of the JFrame class or is there an alternative?
Any suggestion will be much appreciated
Thank-you
P.S. I was also thinking of adding a progress bar and I could
Olly
Rudolf Meerkotter
Greenhorn
Posts: 14
posted 8 years ago
A progress bar would make sense
SCJP 5, SCWCD 5, SCBCD 5
Michael Dunn
Ranch Hand
Posts: 4632
posted 8 years ago
> the more annoying problem is that everything in my program is blocked until it's done saving and doesn't notif the user of what's going on....
if you're doing this via actionPerformed() or similar listener, you're probably blocking the swing thread (EDT),
in which case it would be better to save the data via a separate thread, thus leaving your GUI 'responsive',
then perhaps you wouldn't need a progress bar, as the data would be being saved in the background
I've checked out the SwingWorker class but we're not allowed to use this since it's part of Java 6 (and this project is meant to be coded in Java 5). What is the best alternative?
So far my efforts reveal a problem which seems to be what Michael Dunn pointed out:
I have an algorithm that looks somewhat like this:The problem with this is that the bar doesn't move as the data is being saved. Instead it just magically goes from 0 to 100 when the process is done (which through my understanding seems to be a Thread problem is everything else works when I test it on its own). So to try get around this problem I changed my algorithm a bit to something like:But I still get a similar problem....
I pressume I need to take my first algorithm and run it on a separate Thread but I don't understand why my second algorithm doesn't work :S could anyone please explain this.
I'm pretty sure that SwingWorker was a class that appeared in tutorials and web articles going back several years, and it was only inducted into Java recently. So if you could track down one of those articles you could get the code for it and use it in your project (legalisms permitting, of course).
|
851 F.2d 356Unpublished Disposition
NOTICE: Fourth Circuit I.O.P. 36.6 states that citation of unpublished dispositions is disfavored except for establishing res judicata, estoppel, or the law of the case and requires service of copies of cited unpublished dispositions of the Fourth Circuit.DAIRY KING, INC., Plaintiff-Appelleev.KRAFT, INC., Defendant-AppellantDAIRY KING, INC., Plaintiff-Appellantv.KRAFT, INC., Defendant-Appellee
Nos. 87-2631, 87-2635.
United States Court of Appeals, Fourth Circuit.
Argued: March 10, 1988.Decided: July 11, 1988.
John A. MacColl and David Clarke, Jr. (Piper & Marbury), for appellant and cross-appellee.
Judith D. O'Neill and William K. Meyer (Weinberg and Green), for appellee and cross-appellant.
Before MURNAGHAN and CHAPMAN, Circuit Judges, and JAMES B. McMILLAN, District Judge for the Western District of North Carolina, sitting by designation.
PER CURIAM:
1
Plaintiff Dairy King, Inc., and defendant Kraft, Inc., are disputing about the alleged terms and conditions of a 32-year-old distribution contract which no one has seen for over thirty years and which ostensibly made Dairy King the exclusive Maryland distributor of Breakstone dairy products. The background of this case is summarized in Dairy King, Inc. v. Kraft, Inc., 645 F.Supp. 126 (D.Md.1986). Plaintiff's complaint for breach of the contract was tried to a jury from December 17 to December 22, 1986, before District Judge Joseph Young. On December 22, 1986, the jury returned a verdict in favor of Dairy King. The jury found that in 1953 Dairy King and Breakstone Brothers, defendant Kraft's predecessors, had executed a contract giving Dairy King an exclusive right to distribute Breakstone products in Maryland and requiring Breakstone to buy out Dairy King before re-entering the Maryland market and selling its products to other distributors. The jury also awarded Dairy King damages of $20,000 and a judgment was entered to that effect.
2
After consideration of memoranda tendered by counsel and a hearing on the scope of the contract's buy-out provision, Judge Young entered an order denying defendant's motions for a new trial and for judgment notwithstanding the verdict, and enjoining Kraft either to resume its exclusive arrangement with Dairy King or offer to buy out the Breakstone product-line share of Dairy King's business. Kraft appeals the portion of Judge Young's order denying its post-trial motions and Dairy King appeals Judge Young's order with regard to the scope of the buy-out provision. Since the jury's verdict is amply supported by the evidence and Judge Young's order is consistent with the evidence and that verdict, we affirm.
3
The substantive legal issues as to the contract's validity and enforceability are governed by the law of New York, the place where the alleged contracting took place. Kraft correctly points out that New York law does not enforce vague and indefinite contract provisions; but contract provisions are not vague and indefinite simply because the contracting parties disagree about their meaning. Contract provisions are vague and indefinite if they do not provide an adequate basis or standard for deciding whether the agreement has been kept or broken and there are no extrinsic means by which the terms may be made certain, e.g., trade usage, past practice of the parties. See, Candid Productions v. International Skating Union, 530 F.Supp. 1330 (S.D.N.Y.1982); Bankers Trust Co. of Western New York, 409 N.Y.S.2d 51, 62 (1978).
4
Kraft and Dairy King adamantly disagree whether a contract ever existed and if it did, what its terms were. Dairy King presented evidence of its existence and Kraft presented evidence of its nonexistence.
5
After listening to three days of testimony the jury found that in 1953 an exclusive distributorship contract was entered into between Dairy King and Breakstone Brothers. The jury was apparently persuaded by the testimony of three witnesses who said they had read and still remembered the terms of the contract. Murray Goldstein, former manager of Breakstone and founder of Dairy King, Inc., testified that he negotiated and signed the contract with Larry Becker, the sales manager, vice president and eventual president of Breakstone Brothers, at Becker's New York office. Edith Goldstein testified that she read the contract when her husband returned from New York. Walter Kushner, Dairy King's accountant, testified that he also read the contract when Goldstein returned from Baltimore. Becker's administrative assistant, Eugene Pitrick, did not read the contract and did not witness the actual signing, but testified that he was there when the contract was negotiated and heard it being dictated. The vice president of one of Dairy King's competitors testified that Becker had told him that Breakstone Brothers had an agreement with Dairy King which it had to abide by. And finally, there was the circumstantial evidence that the contract had been performed according to its alleged terms for over thirty years.
6
The jury also found that Kraft could terminate the contract by buying out Dairy King at its fair market value. That issue was presented to the jury and answered as follows:
7
3. Did the alleged contract give Kraft's predecessor Breakstone, the option to buy out Dairy King at its fair market value? (Jury answered "yes" )
8
If yes, does the contract require Breakstone to exercise the option
9
a. before entering market? (Jury answered "Yes" )
10
b. after entering the market? (Jury answered "No" )
11
Kraft contends that even if there was enough evidence to show that some sort of contract was entered into between the parties, there was not enough evidence from which the jury could conclude that Breakstone was required to exercise the buy-out option before entering the market, and that the buy-out provision was the means for ending the contract rather than the means for compensating Dairy King in the event of a breach. Kraft's assertion is incorrect. There were some inconsistencies in the testimony but for the most part Murray Goldstein, Edith Goldstein, and Walter Kushner testified that the buy-out provision was the mechanism by which Dairy King's interest in the distributorship was protected and Breakstone Brothers' right to re-enter the Maryland market was preserved.
12
The evidence was sketchy on some points and hotly contested on others. Judge Young did not err in submitting this case to the jury, and having them wade through, weigh and decide the conflicting evidence. He appropriately accepted the jury's findings, and his order denying Kraft's motions for judgment notwithstanding the verdict and for a new trial is AFFIRMED.
13
The one contested issue which Judge Young did not submit to the jury was the scope of the buy-out provision. Judge Young concluded that the buy-out provision was ambiguous and correctly resolved the ambiguity by evaluating the hearing testimony and concluding that had the parties contemplated the scope of the buy-out provision they would have limited it to the buy out of Dairy King's interest in Breakstone products only. See, National Distillers v. First National Bank, 804 F.2d 978, 982 (7th Cir.1986) ("Ambiguities and gaps in contracts should be resolved by finding what the parties would have bargained for had they addressed the matter explicitly at the time.").
14
The evidence shows that the parties intended to provide Dairy King with a measure of security in the form of an exclusive distributorship and the buy-out provision, and to provide Breakstone with a means of re-entering the Maryland market. When the contract was entered into, Breakstone products accounted for almost all of Dairy King's business. A promise to buy Dairy King's other product lines would have afforded Dairy King only a small increase in financial security and would have been of no benefit to Breakstone Brothers since they were only interested in the distribution of Breakstone products. Today, Breakstone products account for only 10% of Dairy King's product lines. To require Kraft to buy out Dairy King in its entirety would give Dairy King an un-bargained for benefit and Kraft an unassumed burden.
15
Accordingly, the judgment of the district court is AFFIRMED.
|
Box office results show Universal Studios’ “Safe House” ranking number one this week, but a retired Chicago police officer is hoping to close the door on the new film.
According to The Hollywood Reporter, Marty Christopher Lawson claims that the ideas for the film were taken directly from his novel “Truth or Treason,” as well as from a slew of videos he posted on YouTube.
Lawson is attempting to obtain an injunction to stop distribution of the film.
The complaint he filed does not give specific details regarding the similarities or explain how the company may have come across his work, but he claims certain scenes in the movie are similar to his own attempts to uncover alleged financial fraud within the Chicago Police Pension Fund.
He also argues, in his videos, that the movie’s plot about CIA documents and pressures to uncover facts could be related to his own trial, with small alterations to court transcripts.
The retired officer said the movie caused him “pain and suffering” and the lawsuit is currently pending in an Illinois federal court.
Watch the video below to determine whether “Safe House” was, indeed, a rip off of Lawson’s life.
|
The Weekly Standard reserves the right to use your email for internal use only. Occasionally,
we may send you special offers or communications from carefully selected advertisers we believe may be of benefit to our subscribers.
Click the box to be included in these third party offers. We respect your privacy and will never rent or sell your email.
Please include me in third party offers.
Add Brazil to the unhappy trading nations. It attributes the woes of its manufacturing sector to cheap Chinese imports and dumping by developed countries. “We are not going to just sit by while other countries devalue their currencies to give them a competitive advantage…. We don’t want to lose our manufacturing sector,” announced Brazil’s finance minister Guido Mantega. So taxes on foreign cars have been raised, and state-owned Petrobas will direct about 75 percent of its $225 billion capital programme, the world’s largest for any corporation, to local suppliers, a buy-local move also being considered by the EU. More important, Brazil is re-introducing currency controls to prevent the value of its real from rising. These are not “protectionist measures,” claims Mr. Mantega, they are “defensive measures” in response to “non-competitive mechanisms.”
Another developing nation has joined the flight from globalization. “India is not a no-tax country, … not a tax haven, zero-tax or low-tax country,” announced finance minister Pranab Mukherjee. His new budget proposes a tax on some international mergers, retroactive to 1962. Of the foreign companies that have bought assets in India, Vodaphone is the most at risk, liable for $2.2 billion in taxes on its purchase of Indian wireless operations. Mr. Mukherjee denies this will have an adverse effect on much-needed direct foreign investment.
Then there is the problem of China’s restrictions on the export of minerals, including rare earths essential to the manufacture of high-tech goods such as hybrid cars, iPads, and missiles. The EU and Japan have joined our complaint to the World Trade Organisation; China claims its export restrictions are aimed at protecting the environment rather than distorting trade.
More by Irwin M. Stelzer
Whether all of this represents a fraying around the edges of the trading regime that has accompanied globalization, or the beginning of a return to autarky is difficult to say. But we can say that the constituency for free trade, always dispersed and less noisy than advocates of protectionist measures, is in retreat.
|
/*
Copyright 2015 The Kubernetes Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package metrics
import (
"fmt"
"sync"
"time"
"k8s.io/component-base/metrics"
"k8s.io/component-base/metrics/legacyregistry"
corev1 "k8s.io/api/core/v1"
"k8s.io/apimachinery/pkg/types"
utilfeature "k8s.io/apiserver/pkg/util/feature"
"k8s.io/klog/v2"
"k8s.io/kubernetes/pkg/features"
)
// This const block defines the metric names for the kubelet metrics.
const (
KubeletSubsystem = "kubelet"
NodeNameKey = "node_name"
NodeLabelKey = "node"
PodWorkerDurationKey = "pod_worker_duration_seconds"
PodStartDurationKey = "pod_start_duration_seconds"
CgroupManagerOperationsKey = "cgroup_manager_duration_seconds"
PodWorkerStartDurationKey = "pod_worker_start_duration_seconds"
PLEGRelistDurationKey = "pleg_relist_duration_seconds"
PLEGDiscardEventsKey = "pleg_discard_events"
PLEGRelistIntervalKey = "pleg_relist_interval_seconds"
PLEGLastSeenKey = "pleg_last_seen_seconds"
EvictionsKey = "evictions"
EvictionStatsAgeKey = "eviction_stats_age_seconds"
PreemptionsKey = "preemptions"
VolumeStatsCapacityBytesKey = "volume_stats_capacity_bytes"
VolumeStatsAvailableBytesKey = "volume_stats_available_bytes"
VolumeStatsUsedBytesKey = "volume_stats_used_bytes"
VolumeStatsInodesKey = "volume_stats_inodes"
VolumeStatsInodesFreeKey = "volume_stats_inodes_free"
VolumeStatsInodesUsedKey = "volume_stats_inodes_used"
// Metrics keys of remote runtime operations
RuntimeOperationsKey = "runtime_operations_total"
RuntimeOperationsDurationKey = "runtime_operations_duration_seconds"
RuntimeOperationsErrorsKey = "runtime_operations_errors_total"
// Metrics keys of device plugin operations
DevicePluginRegistrationCountKey = "device_plugin_registration_total"
DevicePluginAllocationDurationKey = "device_plugin_alloc_duration_seconds"
// Metric keys for node config
AssignedConfigKey = "node_config_assigned"
ActiveConfigKey = "node_config_active"
LastKnownGoodConfigKey = "node_config_last_known_good"
ConfigErrorKey = "node_config_error"
ConfigSourceLabelKey = "node_config_source"
ConfigSourceLabelValueLocal = "local"
ConfigUIDLabelKey = "node_config_uid"
ConfigResourceVersionLabelKey = "node_config_resource_version"
KubeletConfigKeyLabelKey = "node_config_kubelet_key"
// Metrics keys for RuntimeClass
RunPodSandboxDurationKey = "run_podsandbox_duration_seconds"
RunPodSandboxErrorsKey = "run_podsandbox_errors_total"
)
var (
// NodeName is a Gauge that tracks the ode's name. The count is always 1.
NodeName = metrics.NewGaugeVec(
&metrics.GaugeOpts{
Subsystem: KubeletSubsystem,
Name: NodeNameKey,
Help: "The node's name. The count is always 1.",
StabilityLevel: metrics.ALPHA,
},
[]string{NodeLabelKey},
)
// ContainersPerPodCount is a Counter that tracks the number of containers per pod.
ContainersPerPodCount = metrics.NewHistogram(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: "containers_per_pod_count",
Help: "The number of containers per pod.",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
)
// PodWorkerDuration is a Histogram that tracks the duration (in seconds) in takes to sync a single pod.
// Broken down by the operation type.
PodWorkerDuration = metrics.NewHistogramVec(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: PodWorkerDurationKey,
Help: "Duration in seconds to sync a single pod. Broken down by operation type: create, update, or sync",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
[]string{"operation_type"},
)
// PodStartDuration is a Histogram that tracks the duration (in seconds) it takes for a single pod to go from pending to running.
PodStartDuration = metrics.NewHistogram(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: PodStartDurationKey,
Help: "Duration in seconds for a single pod to go from pending to running.",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
)
// CgroupManagerDuration is a Histogram that tracks the duration (in seconds) it takes for cgroup manager operations to complete.
// Broken down by method.
CgroupManagerDuration = metrics.NewHistogramVec(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: CgroupManagerOperationsKey,
Help: "Duration in seconds for cgroup manager operations. Broken down by method.",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
[]string{"operation_type"},
)
// PodWorkerStartDuration is a Histogram that tracks the duration (in seconds) it takes from seeing a pod to starting a worker.
PodWorkerStartDuration = metrics.NewHistogram(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: PodWorkerStartDurationKey,
Help: "Duration in seconds from seeing a pod to starting a worker.",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
)
// PLEGRelistDuration is a Histogram that tracks the duration (in seconds) it takes for relisting pods in the Kubelet's
// Pod Lifecycle Event Generator (PLEG).
PLEGRelistDuration = metrics.NewHistogram(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: PLEGRelistDurationKey,
Help: "Duration in seconds for relisting pods in PLEG.",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
)
// PLEGDiscardEvents is a Counter that tracks the number of discarding events in the Kubelet's Pod Lifecycle Event Generator (PLEG).
PLEGDiscardEvents = metrics.NewCounter(
&metrics.CounterOpts{
Subsystem: KubeletSubsystem,
Name: PLEGDiscardEventsKey,
Help: "The number of discard events in PLEG.",
StabilityLevel: metrics.ALPHA,
},
)
// PLEGRelistInterval is a Histogram that tracks the intervals (in seconds) between relisting in the Kubelet's
// Pod Lifecycle Event Generator (PLEG).
PLEGRelistInterval = metrics.NewHistogram(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: PLEGRelistIntervalKey,
Help: "Interval in seconds between relisting in PLEG.",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
)
// PLEGLastSeen is a Gauge giving the Unix timestamp when the Kubelet's
// Pod Lifecycle Event Generator (PLEG) was last seen active.
PLEGLastSeen = metrics.NewGauge(
&metrics.GaugeOpts{
Subsystem: KubeletSubsystem,
Name: PLEGLastSeenKey,
Help: "Timestamp in seconds when PLEG was last seen active.",
StabilityLevel: metrics.ALPHA,
},
)
// RuntimeOperations is a Counter that tracks the cumulative number of remote runtime operations.
// Broken down by operation type.
RuntimeOperations = metrics.NewCounterVec(
&metrics.CounterOpts{
Subsystem: KubeletSubsystem,
Name: RuntimeOperationsKey,
Help: "Cumulative number of runtime operations by operation type.",
StabilityLevel: metrics.ALPHA,
},
[]string{"operation_type"},
)
// RuntimeOperationsDuration is a Histogram that tracks the duration (in seconds) for remote runtime operations to complete.
// Broken down by operation type.
RuntimeOperationsDuration = metrics.NewHistogramVec(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: RuntimeOperationsDurationKey,
Help: "Duration in seconds of runtime operations. Broken down by operation type.",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
[]string{"operation_type"},
)
// RuntimeOperationsErrors is a Counter that tracks the cumulative number of remote runtime operations errors.
// Broken down by operation type.
RuntimeOperationsErrors = metrics.NewCounterVec(
&metrics.CounterOpts{
Subsystem: KubeletSubsystem,
Name: RuntimeOperationsErrorsKey,
Help: "Cumulative number of runtime operation errors by operation type.",
StabilityLevel: metrics.ALPHA,
},
[]string{"operation_type"},
)
// Evictions is a Counter that tracks the cumulative number of pod evictions initiated by the kubelet.
// Broken down by eviction signal.
Evictions = metrics.NewCounterVec(
&metrics.CounterOpts{
Subsystem: KubeletSubsystem,
Name: EvictionsKey,
Help: "Cumulative number of pod evictions by eviction signal",
StabilityLevel: metrics.ALPHA,
},
[]string{"eviction_signal"},
)
// EvictionStatsAge is a Histogram that tracks the time (in seconds) between when stats are collected and when a pod is evicted
// based on those stats. Broken down by eviction signal.
EvictionStatsAge = metrics.NewHistogramVec(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: EvictionStatsAgeKey,
Help: "Time between when stats are collected, and when pod is evicted based on those stats by eviction signal",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
[]string{"eviction_signal"},
)
// Preemptions is a Counter that tracks the cumulative number of pod preemptions initiated by the kubelet.
// Broken down by preemption signal. A preemption is only recorded for one resource, the sum of all signals
// is the number of preemptions on the given node.
Preemptions = metrics.NewCounterVec(
&metrics.CounterOpts{
Subsystem: KubeletSubsystem,
Name: PreemptionsKey,
Help: "Cumulative number of pod preemptions by preemption resource",
StabilityLevel: metrics.ALPHA,
},
[]string{"preemption_signal"},
)
// DevicePluginRegistrationCount is a Counter that tracks the cumulative number of device plugin registrations.
// Broken down by resource name.
DevicePluginRegistrationCount = metrics.NewCounterVec(
&metrics.CounterOpts{
Subsystem: KubeletSubsystem,
Name: DevicePluginRegistrationCountKey,
Help: "Cumulative number of device plugin registrations. Broken down by resource name.",
StabilityLevel: metrics.ALPHA,
},
[]string{"resource_name"},
)
// DevicePluginAllocationDuration is a Histogram that tracks the duration (in seconds) to serve a device plugin allocation request.
// Broken down by resource name.
DevicePluginAllocationDuration = metrics.NewHistogramVec(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: DevicePluginAllocationDurationKey,
Help: "Duration in seconds to serve a device plugin Allocation request. Broken down by resource name.",
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
[]string{"resource_name"},
)
// Metrics for node config
// AssignedConfig is a Gauge that is set 1 if the Kubelet has a NodeConfig assigned.
AssignedConfig = metrics.NewGaugeVec(
&metrics.GaugeOpts{
Subsystem: KubeletSubsystem,
Name: AssignedConfigKey,
Help: "The node's understanding of intended config. The count is always 1.",
StabilityLevel: metrics.ALPHA,
},
[]string{ConfigSourceLabelKey, ConfigUIDLabelKey, ConfigResourceVersionLabelKey, KubeletConfigKeyLabelKey},
)
// ActiveConfig is a Gauge that is set to 1 if the Kubelet has an active NodeConfig.
ActiveConfig = metrics.NewGaugeVec(
&metrics.GaugeOpts{
Subsystem: KubeletSubsystem,
Name: ActiveConfigKey,
Help: "The config source the node is actively using. The count is always 1.",
StabilityLevel: metrics.ALPHA,
},
[]string{ConfigSourceLabelKey, ConfigUIDLabelKey, ConfigResourceVersionLabelKey, KubeletConfigKeyLabelKey},
)
// LastKnownGoodConfig is a Gauge that is set to 1 if the Kubelet has a NodeConfig it can fall back to if there
// are certain errors.
LastKnownGoodConfig = metrics.NewGaugeVec(
&metrics.GaugeOpts{
Subsystem: KubeletSubsystem,
Name: LastKnownGoodConfigKey,
Help: "The config source the node will fall back to when it encounters certain errors. The count is always 1.",
StabilityLevel: metrics.ALPHA,
},
[]string{ConfigSourceLabelKey, ConfigUIDLabelKey, ConfigResourceVersionLabelKey, KubeletConfigKeyLabelKey},
)
// ConfigError is a Gauge that is set to 1 if the node is experiencing a configuration-related error.
ConfigError = metrics.NewGauge(
&metrics.GaugeOpts{
Subsystem: KubeletSubsystem,
Name: ConfigErrorKey,
Help: "This metric is true (1) if the node is experiencing a configuration-related error, false (0) otherwise.",
StabilityLevel: metrics.ALPHA,
},
)
// RunPodSandboxDuration is a Histogram that tracks the duration (in seconds) it takes to run Pod Sandbox operations.
// Broken down by RuntimeClass.Handler.
RunPodSandboxDuration = metrics.NewHistogramVec(
&metrics.HistogramOpts{
Subsystem: KubeletSubsystem,
Name: RunPodSandboxDurationKey,
Help: "Duration in seconds of the run_podsandbox operations. Broken down by RuntimeClass.Handler.",
// Use DefBuckets for now, will customize the buckets if necessary.
Buckets: metrics.DefBuckets,
StabilityLevel: metrics.ALPHA,
},
[]string{"runtime_handler"},
)
// RunPodSandboxErrors is a Counter that tracks the cumulative number of Pod Sandbox operations errors.
// Broken down by RuntimeClass.Handler.
RunPodSandboxErrors = metrics.NewCounterVec(
&metrics.CounterOpts{
Subsystem: KubeletSubsystem,
Name: RunPodSandboxErrorsKey,
Help: "Cumulative number of the run_podsandbox operation errors by RuntimeClass.Handler.",
StabilityLevel: metrics.ALPHA,
},
[]string{"runtime_handler"},
)
// RunningPodCount is a gauge that tracks the number of Pods currently running
RunningPodCount = metrics.NewGauge(
&metrics.GaugeOpts{
Subsystem: KubeletSubsystem,
Name: "running_pods",
Help: "Number of pods currently running",
StabilityLevel: metrics.ALPHA,
},
)
// RunningContainerCount is a gauge that tracks the number of containers currently running
RunningContainerCount = metrics.NewGaugeVec(
&metrics.GaugeOpts{
Subsystem: KubeletSubsystem,
Name: "running_containers",
Help: "Number of containers currently running",
StabilityLevel: metrics.ALPHA,
},
[]string{"container_state"},
)
)
var registerMetrics sync.Once
// Register registers all metrics.
func Register(collectors ...metrics.StableCollector) {
// Register the metrics.
registerMetrics.Do(func() {
legacyregistry.MustRegister(NodeName)
legacyregistry.MustRegister(PodWorkerDuration)
legacyregistry.MustRegister(PodStartDuration)
legacyregistry.MustRegister(CgroupManagerDuration)
legacyregistry.MustRegister(PodWorkerStartDuration)
legacyregistry.MustRegister(ContainersPerPodCount)
legacyregistry.MustRegister(PLEGRelistDuration)
legacyregistry.MustRegister(PLEGDiscardEvents)
legacyregistry.MustRegister(PLEGRelistInterval)
legacyregistry.MustRegister(PLEGLastSeen)
legacyregistry.MustRegister(RuntimeOperations)
legacyregistry.MustRegister(RuntimeOperationsDuration)
legacyregistry.MustRegister(RuntimeOperationsErrors)
legacyregistry.MustRegister(Evictions)
legacyregistry.MustRegister(EvictionStatsAge)
legacyregistry.MustRegister(Preemptions)
legacyregistry.MustRegister(DevicePluginRegistrationCount)
legacyregistry.MustRegister(DevicePluginAllocationDuration)
legacyregistry.MustRegister(RunningContainerCount)
legacyregistry.MustRegister(RunningPodCount)
legacyregistry.MustRegister(RunPodSandboxDuration)
legacyregistry.MustRegister(RunPodSandboxErrors)
if utilfeature.DefaultFeatureGate.Enabled(features.DynamicKubeletConfig) {
legacyregistry.MustRegister(AssignedConfig)
legacyregistry.MustRegister(ActiveConfig)
legacyregistry.MustRegister(LastKnownGoodConfig)
legacyregistry.MustRegister(ConfigError)
}
for _, collector := range collectors {
legacyregistry.CustomMustRegister(collector)
}
})
}
// GetGather returns the gatherer. It used by test case outside current package.
func GetGather() metrics.Gatherer {
return legacyregistry.DefaultGatherer
}
// SinceInSeconds gets the time since the specified start in seconds.
func SinceInSeconds(start time.Time) float64 {
return time.Since(start).Seconds()
}
const configMapAPIPathFmt = "/api/v1/namespaces/%s/configmaps/%s"
func configLabels(source *corev1.NodeConfigSource) (map[string]string, error) {
if source == nil {
return map[string]string{
// prometheus requires all of the labels that can be set on the metric
ConfigSourceLabelKey: "local",
ConfigUIDLabelKey: "",
ConfigResourceVersionLabelKey: "",
KubeletConfigKeyLabelKey: "",
}, nil
}
if source.ConfigMap != nil {
return map[string]string{
ConfigSourceLabelKey: fmt.Sprintf(configMapAPIPathFmt, source.ConfigMap.Namespace, source.ConfigMap.Name),
ConfigUIDLabelKey: string(source.ConfigMap.UID),
ConfigResourceVersionLabelKey: source.ConfigMap.ResourceVersion,
KubeletConfigKeyLabelKey: source.ConfigMap.KubeletConfigKey,
}, nil
}
return nil, fmt.Errorf("unrecognized config source type, all source subfields were nil")
}
// track labels across metric updates, so we can delete old label sets and prevent leaks
var assignedConfigLabels map[string]string
// SetAssignedConfig tracks labels according to the assigned NodeConfig. It also tracks labels
// across metric updates so old labels can be safely deleted.
func SetAssignedConfig(source *corev1.NodeConfigSource) error {
// compute the timeseries labels from the source
labels, err := configLabels(source)
if err != nil {
return err
}
// clean up the old timeseries (WithLabelValues creates a new one for each distinct label set)
if !AssignedConfig.Delete(assignedConfigLabels) {
klog.Warningf("Failed to delete metric for labels %v. This may result in ambiguity from multiple metrics concurrently indicating different assigned configs.", assignedConfigLabels)
}
// record the new timeseries
assignedConfigLabels = labels
// expose the new timeseries with a constant count of 1
AssignedConfig.With(assignedConfigLabels).Set(1)
return nil
}
// track labels across metric updates, so we can delete old label sets and prevent leaks
var activeConfigLabels map[string]string
// SetActiveConfig tracks labels according to the NodeConfig that is currently used by the Kubelet.
// It also tracks labels across metric updates so old labels can be safely deleted.
func SetActiveConfig(source *corev1.NodeConfigSource) error {
// compute the timeseries labels from the source
labels, err := configLabels(source)
if err != nil {
return err
}
// clean up the old timeseries (WithLabelValues creates a new one for each distinct label set)
if !ActiveConfig.Delete(activeConfigLabels) {
klog.Warningf("Failed to delete metric for labels %v. This may result in ambiguity from multiple metrics concurrently indicating different active configs.", activeConfigLabels)
}
// record the new timeseries
activeConfigLabels = labels
// expose the new timeseries with a constant count of 1
ActiveConfig.With(activeConfigLabels).Set(1)
return nil
}
// track labels across metric updates, so we can delete old label sets and prevent leaks
var lastKnownGoodConfigLabels map[string]string
// SetLastKnownGoodConfig tracks labels according to the NodeConfig that was successfully applied last.
// It also tracks labels across metric updates so old labels can be safely deleted.
func SetLastKnownGoodConfig(source *corev1.NodeConfigSource) error {
// compute the timeseries labels from the source
labels, err := configLabels(source)
if err != nil {
return err
}
// clean up the old timeseries (WithLabelValues creates a new one for each distinct label set)
if !LastKnownGoodConfig.Delete(lastKnownGoodConfigLabels) {
klog.Warningf("Failed to delete metric for labels %v. This may result in ambiguity from multiple metrics concurrently indicating different last known good configs.", lastKnownGoodConfigLabels)
}
// record the new timeseries
lastKnownGoodConfigLabels = labels
// expose the new timeseries with a constant count of 1
LastKnownGoodConfig.With(lastKnownGoodConfigLabels).Set(1)
return nil
}
// SetConfigError sets a the ConfigError metric to 1 in case any errors were encountered.
func SetConfigError(err bool) {
if err {
ConfigError.Set(1)
} else {
ConfigError.Set(0)
}
}
// SetNodeName sets the NodeName Gauge to 1.
func SetNodeName(name types.NodeName) {
NodeName.WithLabelValues(string(name)).Set(1)
}
|
As Republicans decide how to respond to the new agreement between the West and Iran, it is right to remain skeptical of both Iran’s intentions and Presidents Obama’s instincts and abilities.
The usual suspects are lining up in opposition to the deal, but when an agreement draws criticism from those as informed and diverse as Israeli Prime Minister Benjamin Netanyahu and Saudi Arabia’s Prince Alwaleed bin Talal, you have to take notice. These two even appear to be working as a tag team. Alwaleed can say things that Netanyahu cannot — specifically, that no one trusts President Obama to do the right thing. Netanyahu is pretty plainspoken himself. He said the deal reached in Geneva on Saturday night is “not a historic agreement, but a historic mistake.”
Whether or not they say so publicly, everyone agrees that as the plan takes effect, part of what will need to be verified is President Obama’s almost-certain claim that the deal is a great success, regardless of what transpires in the months ahead. No one has any reason to believe this president will admit mistakes once he starts taking credit. President Obama doesn’t appear to have been closely involved in the construction of the deal and at least so far, the agreement isn’t all about Obama. Some very good, able people shaped this deal and they deserve some deference, even if the president does not. Republicans should think of effective ways to police the terms of this agreement rather than rush to cable news studios with shrill hyperbole that only contributes noise.
Puneet Talwar is currently serving as special assistant to the president and senior director for Iran, Iraq and the Gulf States on the White House National Security Staff. According to the Wall Street Journal, Mr. Talwar, among others, had been involved with secret negotiations with the Iranians that produced the agreement. By all accounts, Mr. Talwar is a real pro with no agenda other than a safer America. He isn’t trying to create a smokescreen to hide Obamacare, and he isn’t driven by egotistical anxieties about his personal legacy or what thanks and praise he will receive.
Republicans can and should be cautious. They have every right to criticize the agreement, but it would be more useful if they state their concerns with specifics and define with clarity what results they will be looking for over the next six months. It is also wise for Congress to continue to build on the sanctions they already put in place. Republicans should ignore the White House opposition to more sanctions. It’s no surprise the White House is against further sanctions — remember that this administration was against the previous sanctions, which have proven to be effective.
Also, the GOP doesn’t need to maneuver for partisan advantage. Foreign policy won’t drive any votes in 2014 unless war breaks out. For Republicans, the future is tied to health care and the economy, not how many centrifuges Iran has. When Americans gather around their Thanksgiving dinner tables, they won’t be talking about Iranian uranium enrichment. They will be sharing their Obamacare horror stories and bemoaning the bleak economy.
The bottom line is, the GOP should give this deal a chance to work. Iran was never going to suddenly and completely capitulate, so any agreement had to have a first step. And this first step seems reasonable, verifiable and worth trying.
Side note: The economy may be in for a lift. The booming stock market could soon be accompanied by lower fuel prices. The trickle-down of stock market wealth and cheaper gasoline prices might combine to give some real-time relief to the working class. This could have some interesting political consequences. More about this later.
Ed Rogers is a contributor to the PostPartisan blog, a political consultant and a veteran of the White House and several national campaigns. He is the chairman of the lobbying and communications firm BGR Group, which he founded with former Mississippi Gov. Haley Barbour in 1991.
Comments our editors find particularly useful or relevant are displayed in Top Comments, as are comments by users with these badges: . Replies to those posts appear here, as well as posts by staff writers.
To pause and restart automatic updates, click "Live" or "Paused". If paused, you'll be notified of the number of additional comments that have come in.
Comments our editors find particularly useful or relevant are displayed in Top Comments, as are comments by users with these badges: . Replies to those posts appear here, as well as posts by staff writers.
|
Q:
Power 14 red LED's and arduino from same source
I am new to the science behind electronics (software guy at heart) but I am working on a project where I need to power an Arduino and 14 standard 20 mA red LED's preferably from the same source.
I want to be able to control the LED's from the Arduino, nothing fancy, just fade them all on and off (together). I have a Darlington TIP120 on hand, will this be useful?
A:
Since you want them all controlled from a single pin, you will need to use the transistor (it is unlikely that a single Arduino pin can safely source or sink enough current for that many LEDs).
simulate this circuit – Schematic created using CircuitLab
The output will be PWM, which uses a variable duty cycle to control the brightness.
|
Q:
Control + F4 not working in ubuntu 16.04
Control + F4 option which is used to close current tab in google chrome is not working in ubuntu 16.04.
A:
Try this for closing the tab in chrome:
ctrl + w
|
MANILA, Philippines - Veteran director and comedian Manny Castañeda is now at the center of controversy after his Facebook post alleging the Aquino administration of protecting and benefitting from narco-politicians spread like wildfire on social media.
PHOTO CREDIT: Facebook/Manny Castañeda/shminhe.com
Castañeda speculated that the members of the Aquino administration, having ignored the illegal drug menace and consequently allowing it to grow rapidly in the country, may be recipients of money from the drug trade. He explained that the suspicious actions of Sen. Leila de Lima and Sen. Franklin Drilon, who are members of Aquino’s Liberal Party, appear to be efforts to make President Rodrigo Duterte’s list of narco-politicians seem “useless.”
“Instead of being jubilant, it was noticeable that the yellow personalities became suspiciously defensive. Everything was done to discredit the list and to make it look irrelevant and useless. All concerted efforts seem very suspicious indeed. There must be a more substantial reason for this tirade other than just sheer dislike for Pres. Duterte as a political adversary. The noose is getting tight maybe?” Castañeda explained.
In his post, Castañeda also alleged the Aquino administration of doing nothing despite knowledge of the existence of narco-politics in the Philippines. Castañeda even cited former Department of Interior Local Government (DILG) chief and losing presidential candidate Mar Roxas as an example saying that the former possessed explicit data about the drug problem in Davao City. It can be recalled that Roxas once said that he knew where to buy illegal drugs in Davao City to debunk Duterte’s claim that Davao City is drug-free.
Castañeda, with all the details he mentioned, concluded that the Aquino administration was indeed receiving drug-money.
“With all the circumstances mentioned, I cannot help but conclude that the Aquino administration was a recipient of drug-money,” Castañeda explained. “It is only this motive that can explain their behavior.”
|
VOICE OF THE PEOPLE (letter)
Pilot seniority
March 03, 2007|By Karin Matusiak
La Grange Highlands — While two letters to the editor make outstanding points on why commercial pilots' retirement age should not be raised from the current 60 years to the age of 65, citing health concerns, there is one point that neither letter writer nor your "Gray hair in the cockpit" Editorial on Feb. 10 addresses. Many pilots want to put off retirement solely for financial reasons.
Commercial airline pilots' pay is based on seniority, hence their pay scale. From the size of the aircraft they fly to the hours in a month they work, it all is determined by where they rank in the seniority list at the domicile that they are based. Their seniority at their base is the ultimate factor in determining which aircraft they fly, which will determine their hourly flying rate. The only way that a pilot can advance in seniority is if a pilot senior to him or her quits, dies or retires from the company.
Pilots have sacrificed much over the past five years. Most have had to endure substantial pay cuts and reduction of benefits, and some have had a substantial reduction in retirement benefits. And now a small group of pilots wants to delay the pilots junior to them the opportunity to reap the benefits that they have had during their careers, for another five years.
I have read articles in numerous publications in which pilots state that that they "live to fly." While I do not question their sincerity, I do question why none of them mentions the additional five years of income at generally the highest rates in the business, and for some carriers, the benefits that would be eliminated for them if they would retire at age 60.
Would these men and women still be interested in flying for their carriers at age 60 plus one day as the most junior co-pilot in the system taking the substantial decrease in pay that comes with their new position? What if they agreed to that without any additional benefits and no additional funds placed in their retirement funds?
In my opinion, the number of pilots who "live to fly" would be reduced substantially.
|
Minutes of the 6th meeting of EFSA’s CEP Panel working group on food contact materials now available online; discussion of criteria and tools for prioritizing food contact plastics substances on the Union list that do not have specific migration limits
EFSA’s CEP Panel working group on phthalates reviews summary statistics on the outcome of a public consultation on draft opinion of DBP, BBP, DEHP, DINP, DIDP; focus placed on reviewing comments related to the exposure assessment
Minutes of the 5th meeting of EFSA’s CEP Panel working group on food contact materials now available online; continues risk-assessing several new FCMs, presents extensive summary of views on the current framework regulation
About Us
The Food Packaging Forum Foundation is a science communication organisation. The Food Packaging Forum provides independent and balanced information on issues related to food packaging and health. In doing so the Food Packaging Forum addresses all its stakeholders, including business decision makers, regulators, media and communication experts, scientists and consumers.
|
Evolution of caprine and ovine beta-defensin genes.
Defensins comprise an important family of anti-microbial peptides. Among vertebrates, numerous defensin genes have been detected, but their evolutionary background is still discussed. We investigated the molecular evolution and variability of beta-defensins of Caprini via sequence analyses of defensin introns. Screening of several domestic and wild species of Caprini revealed a total of 13 discrete beta-defensin coding sequences, with three of them described before this study. Phylogenetic analyses revealed that the array of newly described defensin genes is of monophyletic origin and has arisen in numerous independent duplication events after separation of the ancestral defensins. As a result of that scenario, recent defensin genes are distributed in a species-specific manner. Values of synonymous and non-synonymous substitutions demonstrated that both modes of evolutionary pressure, positive as well as negative selection, have acted. In addition, conservation of some beta-defensin exons is demonstrated. Discrimination of certain beta-defensin genes was possible only due to intron-specific differences. Therefore, sequence analyses restricted to the exons would result in underestimation of the number of beta-defensin genes. Our study shows that for reconstruction of the phylogenetic history data of defensin introns are more appropriated. Comparisons among the amino acid sequences show moderate substitutions without changing the net charge of the mature peptides.
|
Introduction
============
Complications of pregnancy and childbirth are the leading causes of disability and death among women of reproductive age in developing countries, accounting for at least 18% of the global burden of disease in this age-group.[@b1-ijwh-5-523] Similarly, the pattern of leading causes of maternal death and disability are closely linked to high prevalence of vitamin D deficiency, poor maternal health during pregnancy, inadequate care during delivery, and lack of newborn care.[@b1-ijwh-5-523] Every year, almost 8 million stillbirths and early neonatal deaths occur. In addition to maternal deaths, more than 50 million women experience maternal health problems annually.[@b2-ijwh-5-523] In underdeveloped countries, those high vitamin D deficiency and maternal complications often pose an immediate financial burden on women and their households.
Vitamin D deficiency during pregnancy is common in many parts of the world,[@b3-ijwh-5-523] and there is a very strong relationship between vitamin deficiency D and multiple potential adverse pregnancy outcomes.[@b4-ijwh-5-523]--[@b7-ijwh-5-523] However, the role and metabolism of vitamin D in the pregnant state is not well known or understood.[@b5-ijwh-5-523],[@b8-ijwh-5-523]--[@b10-ijwh-5-523] Despite the reported high prevalence of deficiency and the possible consequences, the desired optimal level needed for pregnant women in their body and the amount of vitamin D intake required to maintain adequate levels is not very well documented.[@b3-ijwh-5-523]--[@b12-ijwh-5-523]
The vitamin D stores in the infant start with transplacental transfer of 25(OH)D in early pregnancy from mother to fetus.[@b5-ijwh-5-523],[@b13-ijwh-5-523] It is very obvious that maintaining optimum vitamin D nutrition during pregnancy is essential for prevention of hypovitaminosis D in the fetus and vitamin D deficiency at birth and in early infancy.[@b13-ijwh-5-523]
An estimated 1.6 billion people are anemic worldwide, and anemia is common during pregnancy. Approximately 50% of pregnant women have anemia (hemoglobin \<110 g/L) in resource-limited settings, compared with 12%--25% in developed regions.[@b14-ijwh-5-523] Anemia is a global public health problem affecting both developing and developed countries, with major consequences for human health as well as social and economic development. It occurs at all stages of the life cycle, but is more prevalent in pregnant women and young children. In 2002, iron-deficiency anemia (IDA) was considered to be among the most important contributing factors to the global burden of disease.[@b14-ijwh-5-523] Anemia is the result of a wide variety of causes that can be isolated, but more often coexist. Globally, the most significant contributor to the onset of anemia is iron deficiency, so "IDA" and "anemia" are often used synonymously, and the prevalence of anemia has often been used as a proxy for IDA. It is generally assumed that 50% of the cases of anemia are due to iron deficiency,[@b14-ijwh-5-523] but the proportion may vary among population groups and in different areas according to the local conditions. The main risk factors for IDA include a low intake of iron, poor absorption of iron from diets high in phytate or phenolic compounds, and period of life when iron requirements are especially high (ie, growth and pregnancy).
In recent years, the prevalence of vitamin D deficiency has increased, and the incidence of low vitamin D status has risen in the developing world and in the UK and other developed countries.[@b11-ijwh-5-523],[@b15-ijwh-5-523]--[@b17-ijwh-5-523] There is a high prevalence of vitamin D deficiency in pregnant women from non-Western countries in Northern Europe, and vitamin D deficiency during pregnancy is an ongoing epidemic.[@b18-ijwh-5-523]
At present, there is not enough evidence to evaluate the effectiveness of vitamin D in pregnancy, and therefore vitamin D supplementation is not routinely offered to all pregnant women in Qatar. Many studies have investigated vitamin D deficiency in pregnancy in different ethnic minority groups and the method of supplementation.[@b19-ijwh-5-523] There is evidence that an immediate dose is as effective as a daily dose, with no adverse effects.[@b13-ijwh-5-523]--[@b18-ijwh-5-523]
The objective of the study was to determine the impact of vitamin D deficiency on gestational diabetes mellitus (GDM), anemia, iron deficiency, and preeclampsia medical conditions among pregnant women.
Subjects and methods
====================
This was a cohort study conducted among pregnant Arab women in the third trimester from January 2011 to December 2011. The study was based on the logbook of the Women's Hospital, which registers all pregnant women visiting antenatal clinics at the Women's Hospital, Hamad Medical Corporation. The research assistants screened the logbook of the Women's Hospital during the study period and prepared a list of 2,487 pregnant Arab women above 24 weeks of gestation with any maternal complication. All the eligible women on the prepared list (n = 2,487) were approached, and consent was sought. Only 1,873 pregnant women (75.3%) participated in the study (refer to [Figure 1](#f1-ijwh-5-523){ref-type="fig"}). Nearly 25% of the approached subjects were excluded from the study, because some of them had incomplete medical records as well as incomplete questionnaires. A few of them did not give consent due to their busy schedule and lack of time to fill the questionnaire. Moreover, our study sample included 1,873 pregnant women, which is approximately 10% of total mothers that delivered in the year 2011. Each participant was provided with brief information about the study and was assured of strict confidentiality.
The study was approved by the Research Ethics Committee of Hamad Medical Corporation and by Institutional Review Board of Weill Cornell Medical College. All human studies are approved by the Research Ethics Committee, and therefore are performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. All the persons who agreed to participate in this study gave their informed consent prior to their inclusion in the study.
Data were collected through face-to-face interviews by qualified nurses using a validated questionnaire in the local language. The questionnaire covered sociodemographic characteristics of the pregnant women, family and medical histories, anthropometric and obstetric characteristics, and pregnancy outcomes. Maternal medical records were reviewed to collect detailed medical and clinical information. The questionnaire was pilot-tested on 100 randomly selected pregnant women for validity. The investigators made necessary corrections and modifications after considering minor differences and discrepancies that were found during the pilot study. Data on maternal complications of each patient were retrieved from the medical record file. In addition, follow-up data on pregnancy and neonatal outcome of each woman were obtained from the labor-room register immediately after delivery ([Figure 1](#f1-ijwh-5-523){ref-type="fig"}).
Laboratory investigations
-------------------------
### Blood collection and serum measurements of vitamin D
Serum concentrations of vitamin D metabolites were determined by radioimmunoassay kits from Immunodiagnostic Systems Ltd, Boldon, UK. A trained phlebotomist collected venous blood samples, separated serum, and stored at −70°C until analysis. Serum 25-hydroxyvitamin D (25\[OH\]D), a vitamin D metabolite, was measured using a commercially available kit (DiaSorin, Saluggia, Italy). The treated samples were then assayed using competitive-binding radioimmunoassay. Subjects were classified into four categories: (1) severe vitamin D deficiency, 25(OH)D \<10 ng/mL; (2) moderate deficiency, 25(OH)D 10--19 ng/mL; (3) mild deficiency, 25(OH)D 20--29 ng/mL; and (4) normal/optimal level, between 30 and 80 ng/mL,[@b13-ijwh-5-523],[@b15-ijwh-5-523] which is consistent with other studies.[@b18-ijwh-5-523],[@b19-ijwh-5-523] Other baseline biochemical parameters measured from the serum included vitamin D, calcium, albumin, bilirubin, magnesium, calcium, cholesterol, urea, triglyceride, and phosphorus and parathyroid hormone levels. On the basis of previous recommendations,[@b19-ijwh-5-523]--[@b21-ijwh-5-523] serum levels of these biochemical parameters were determined according to standard laboratory procedures. Furthermore, during the screening period, each patient provided a complete history, and a comprehensive examination was performed.
Preeclampsia was defined as gestational hypertension and proteinuria and return of all abnormalities to normal by 12 weeks postpartum. Gestational hypertension was defined according to WHO[@b20-ijwh-5-523] criteria as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg and/or higher for the first time after 20 weeks' gestation. According to the recommendations from the WHO Expert Committee,[@b21-ijwh-5-523] pregnant women were screened at 24--28 weeks for GDM, which was defined according to fasting venous blood glucose concentration ≥7.0 mmol/L and/or 2 hours post-oral glucose tolerance test venous blood glucose concentration ≥11.1 mmol/L. The WHO[@b22-ijwh-5-523] tabulation adopts the international definition for anemia for pregnant women of \<110 g/L. Anemia was defined as a hemoglobin level of \<11 g/dL for women.
Data were analyzed using SPSS version 20 (IBM, Armonk, NY, USA). Student's *t*-test was used to ascertain the significance of differences between mean values of two continuous variables and confirmed by the nonparametric Mann--Whitney test. Chi-squares analysis was performed to test for differences in proportions of categorical variables between two or more groups. In 2 × 2 tables, Fisher's exact test (two-tailed) replaced the chi-squared test if the assumptions underlying chi-squared were violated, namely in cases of small sample size and where the expected frequency was less than five in any of the cells. Pearson's correlation coefficient was used to evaluate the strength association between two continuous variables. All statistical tests were two-sided, and *P* \< 0.05 was considered statistically significant.
Results
=======
[Table 1](#t1-ijwh-5-523){ref-type="table"} shows the sociodemographic characteristics and lifestyle habits of the studied pregnant women according to vitamin D status. Nearly half of the studied pregnant women had vitamin D deficiency (48.4%). Of women at high risk of vitamin D deficiency, younger women below 30 years old (43.2%, *P* = 0.032), housewives (65.3%, *P* = 0.008), and those with low monthly household income (QR5,000--9,999) (49.2%, *P* = 0.03) were significantly more likely to have lower vitamin D compared with those of higher socioeconomic status. In terms of lifestyle habits, vitamin D deficiency was higher in pregnant women who had less exposure to sun (63.4% vs 59%, *P* = 0.05) and no daily physical activity (64.4% vs 59.0%, *P* = 0.05) compared to the vitamin D-sufficient group. Vitamin D supplement intake was significantly lower in vitamin D-deficient pregnant women (89.7%, *P* \< 0.001).
[Table 2](#t2-ijwh-5-523){ref-type="table"} shows the sociodemographic characteristics of studied pregnant women with the complications of GDM, anemia, iron deficiency, and preeclampsia. GDM (52.7% vs 47.3%), anemia (53.2% vs 46.8%), iron deficiency (55.6% vs 44.4%), and preeclampsia (51.9% vs 48.1%) were more frequent among pregnant Qatari women than in pregnant non-Qatari women. The pregnancy complications of GDM (52.7%), anemia (53.2%), iron deficiency (55.6%), and preeclampsia (51.9%) were higher in Qataris. Housewives had higher pregnancy complications -- GDM (66.2%), anemia (66.2%), iron deficiency (68.5%), and preeclampsia (58.1%) -- compared to working women. Also, the pregnant women with low monthly household income had a higher frequency of GDM (50.4%), anemia (46.3%), iron deficiency (51.9%), and preeclampsia (43.4%).
[Table 3](#t3-ijwh-5-523){ref-type="table"} shows the serum level of vitamin D among pregnant women with pregnancy complications. Severe deficiency was significantly higher in pregnant women with GDM (16.5% vs 11%), anemia (17.1% vs 11%), iron deficiency (18.5% vs 11.2%), and preeclampsia (19.8% vs 11.4%) when compared to the uncomplicated group. Also, moderate vitamin deficiency was significantly higher in pregnant women with GDM (38.1% vs 36.4%) and anemia (37.5% vs 36%) than in the uncomplicated group.
[Table 4](#t4-ijwh-5-523){ref-type="table"} shows the maternal and obstetric characteristics of pregnant women with pregnancy complications: GDM, anemia, iron deficiency, and preeclampsia. Obesity was higher among pregnant women with GDM (41.5%) and preeclampsia (41.1%). GDM (39.6%), anemia (45.8%), iron deficiency (40.1%), and preeclampsia (41.1%) were observed more in women, with four to six parity groups.
[Table 5](#t5-ijwh-5-523){ref-type="table"} shows the predictors for GDM, anemia, iron deficiency, and preeclampsia in studied pregnant women using multivariate logistic regression analysis. Vitamin D deficiency was a significant contributor for GDM (odds ratio \[OR\] 1.38, confidence interval \[CI\] 1.05--1.82; *P* = 0.019) and anemia (OR 1.87, CI 1.32--2.63; *P* \< 0.001). Previous low birth weight (OR 1.75, CI 1.20--256; *P* = 0.003) and consanguinity (second cousin) (OR 1.64, CI 1.22--2.20; *P* = 0.001) were the second major contributor for GDM and anemia. Vitamin D deficiency appeared as a significant contributor for iron deficiency (OR 1.84, CI 1.13--3.32; *P* = 0.020) and preeclampsia (OR 1.75, CI 1.16--2.58; *P* = 0.010) in pregnant women.
[Table 6](#t6-ijwh-5-523){ref-type="table"} shows the global variations and comparison in vitamin D status during pregnancy. Low vitamin D intake was higher in the US (24.8%),[@b40-ijwh-5-523] the People's Republic of China (24.2%),[@b38-ijwh-5-523] Canada (20.7%),[@b36-ijwh-5-523] and Australia (21%),[@b10-ijwh-5-523] which is in line with the present study rate (17.8%).
Discussion
==========
Vitamin D deficiency has been associated with several adverse health outcomes, including pregnancy outcomes, and is a public health issue worldwide. The present study demonstrated a greater risk of maternal complications in pregnant women with vitamin D deficiency compared with those with normal vitamin D levels. An increased risk of adverse pregnancy outcomes, such as GDM, anemia, iron deficiency and preeclampsia, was noted in the study sample of pregnant women, which is similar to the results observed by Kaludjerovic and Vieth.[@b22-ijwh-5-523] The mean serum levels of vitamin D in pregnant women with GDM (17.78 ± 7.7), anemia (17.65 ± 7.8), iron deficiency (16.01 ± 7.4), and preeclampsia (18.01 ± 8.6) revealed vitamin D insufficiency in pregnant women. It was reported[@b23-ijwh-5-523]--[@b24-ijwh-5-523] that vitamin D status in women with risk factors of maternal complications like GDM, anemia, and preeclampsia was significantly lower compared with women without risk factors. Even multivariate logistic regression analysis supported this finding that vitamin D deficiency was a significant contributor for GDM (OR 1.387, *P* = 0.019) and anemia (OR 1.87, *P*≤0.001).
Gestational diabetes is a result of pregnancy-induced insulin resistance and impaired compensating insulin secretion. Vitamin D improves insulin sensitivity by enhancing insulin responsiveness to glucose transport. In a meta-analysis review of various studies, Clifton-Bligh and McElduff[@b23-ijwh-5-523] reported that pregnant women with GDM had significantly lower 25(OH)D levels than the comparison group. In our study sample, severe vitamin deficiency was significantly higher in pregnant women with GDM (16.5%) than those without GDM (11%), which is in line with findings from the cross-sectional study of Maghbooli et al.[@b25-ijwh-5-523] More than half of the complicated pregnant women with GDM (54.6%) had vitamin deficiency. In the study of Zhang et al,[@b26-ijwh-5-523] 33% of GDM cases had vitamin D deficiency which was higher than in their comparable group (14%). On the contrary, no significant association was observed between 25(OH)D concentrations and GDM risk in the Indian population.[@b27-ijwh-5-523] Another study from the UK[@b28-ijwh-5-523] failed to find a significant association between vitamin D deficiency and risk of GDM. These study findings highlight the possibility that vitamin D sufficiency may be important to maternal health, vitamin D sufficiency may be important to maternal health; vitamin D supplementation may be a simple way to reduce the risk of adverse pregnancy outcome.
Gestational anemia is common in developing countries, where it affects more than 57% of pregnancies.[@b29-ijwh-5-523] Our study is one of the first to investigate the association of vitamin D and anemia. The present study demonstrated a significantly greater risk of anemia in pregnant women with severe vitamin D deficiency (17.1%) than that in normal groups (11%). More than half of the pregnant women with anemia had vitamin D insufficiency (54.6%), whereas it was lower in the normal group without anemia (47%). Sim et al[@b30-ijwh-5-523] reported a significant correlation between vitamin D deficiency and anemia. Anemia is regarded as a major risk factor for maternal and perinatal mortality and morbidity.
Similar to gestational diabetes and anemia, the data revealed a significant association between preeclampsia and severe vitamin D deficiency. Vitamin D deficiency was significantly higher in pregnant women with preeclampsia compared with the normal group (19.8% vs 11.4%), which is consistent with a meta-analysis[@b31-ijwh-5-523] showing a significant association between preeclampsia and 25(OH)D insufficiency. However, no significant association was observed in a prospective cohort study[@b32-ijwh-5-523] of pregnancies at high risk for preeclampsia in Canada. These study results revealed that maternal vitamin D deficiency was a strong, independent risk factor for preeclampsia. Vitamin D deficiency may also elevate blood pressure.
In the study sample, the risk of vitamin D deficiency was higher among young pregnant women below 30 years old (43.2%), housewives (65.3%), the educated (41.1%), and those with low monthly household income (\<QR10,000) (49.2%). In terms of lifestyle habits, vitamin D deficiency was higher among women with history of inadequate sun exposure (63.4%) and less physical activity (64.4%). It was reported[@b32-ijwh-5-523] that women with low socioeconomic status and inadequate sun exposure were more likely to have lower 25(OH)D concentrations than those of higher socioeconomic state and adequate sun exposure. In most countries, there is no monitoring of serum 25(OH)D~3~ levels during pregnancy for vitamin D deficiency. The vitamin D-deficiency epidemic during pregnancy is caused by a lack of adequate sunlight exposure needed to synthesize vitamin D~3~ in the skin, coupled with overall intakes that are too low to meet the increased demands of pregnancy. Previous studies of Bener et al[@b33-ijwh-5-523],[@b34-ijwh-5-523] support these findings that vitamin D deficiency was more prevalent in Qatar and the population had less exposure to sun, although it is a sun-enriched population. Similar socioeconomic status of vitamin D deficiency was observed in pregnant women with GDM, anemia, and preeclampsia. Pregnant women who developed these maternal complications were younger, educated, housewives, and with low monthly household income. These findings provide evidence suggesting that vitamin D deficiency or insufficiency is common during pregnancy, which leads to adverse pregnancy outcome. Similar to the results observed in Qatar (17.8%), vitamin D deficiency was higher in the US (24.8%),[@b40-ijwh-5-523] the People's Republic of China (24.2%),[@b38-ijwh-5-523] Canada (20.7%),[@b36-ijwh-5-523] and Australia (21%).[@b10-ijwh-5-523]
The current study supports the hypothesis that vitamin D insufficiency is associated with an increased risk of such maternal complications as GDM, anemia, and preeclampsia. The vitamin D-deficiency epidemic during pregnancy is caused by a lack of adequate sunlight exposure needed to synthesis vitamin D~3~ in skin, coupled with overall intakes that are too low to meet the increased demand of pregnancy. Women living in the Middle East generally cover their body, so they receive little vitamin D from solar UVB. Also, it is so hot during the summer that people tend to stay indoors. The study findings suggest that low levels of 25(OH)D may be a modifiable risk factor in pregnancy, and health-care providers should at least be encouraging pregnant women to take vitamin D.
Conclusion
==========
The study findings revealed that maternal vitamin D deficiency in pregnancy is significantly associated with elevated risk for GDM, anemia, and preeclampsia. The risk of vitamin D deficiency was higher in Qataris, housewives, and those with low monthly house hold income. Vitamin D-deficient pregnant women had less exposure to sunlight, no physical activity, and less vitamin D supplement intake. Vitamin D supplementation in early pregnancy may be a simple way to reduce the risk of these adverse pregnancy outcomes.
This work was generously supported and funded by the Qatar Foundation, grant NPRP 08-760-3-153. The project was partially supported and funded by the Qatar Diabetic Association, Qatar Foundation. We also would like to thank Hamad Medical Corporation (10146/10) and the Weill Cornell Medical College Qatar Institutional Review Board (2010--0021) for their ethical approval of this study.
**Disclosure**
The authors report no conflicts of interest in this work.
{#f1-ijwh-5-523}
######
Sociodemographic characteristics and lifestyle habits of the studied pregnant women according to vitamin D status (n = 1,873)
Variables Total, n (%) Vitamin D status
------------------------------------------------------------------------------ --------------- ------------------ ------------ ---------
**Nationality**
Qatari 907 (48.42) 426 (47.0) 481 (49.8) 0.22
Non-Qatari 966 (51.58) 481 (53.0) 485 (50.2)
**Age-group**
\<30 years 746 (39.83) 392 (43.2) 354 (36.6) 0.03
30--34 years 457 (24.40) 212 (23.4) 245 (25.4)
35--39 years 383 (20.45) 170 (18.7) 213 (22.0)
40--45 years 287 (15.32) 133 (14.7) 154 (15.9)
**Body mass index group (kg/m^2^)**
20--24.99 (normal) 854 (45.6) 411 (45.3) 443 (45.9) 0.93
25--30 (overweight) 539 (28.8) 260 (28.7) 279 (28.9)
\>30 (obese) 480 (25.6) 236 (26.0) 244 (25.3)
**Education level**
Illiterate 118 (6.30) 65 (7.2) 53 (5.5) 0.48
Primary 167 (8.92) 77 (8.5) 90 (9.3)
Intermediate 172 (9.18) 89 (9.8) 83 (8.6)
Secondary 631 (33.69) 303 (33.4) 328 (34.0)
University 785 (41.91) 373 (41.1) 412 (42.7)
**Occupation**
Housewife 1,164 (62.15) 592 (65.3) 572 (59.2) 0.01
Sedentary/professional 495 (26.43) 225 (24.8) 270 (28.0)
Manual 128 (6.83) 61 (6.7) 67 (6.9)
Business man 53 (2.83) 21 (2.3) 32 (3.3)
Army/police 33 (1.76) 8 (0.9) 25 (2.6)
**Monthly household income (QR)**[\*](#tfn1-ijwh-5-523){ref-type="table-fn"}
5000--9999 885 (47.25) 446 (49.2) 439 (45.4) 0.03
10,000--14,999 393 (20.98) 198 (21.8) 195 (20.2)
15,000--20,000 403 (21.52) 169 (18.6) 234 (24.2)
\>20,000 192 (10.25) 94 (10.4) 98 (10.1)
**Place of residence**
Urban 1,612 (86.07) 787 (86.8) 825 (85.4) 0.36
Rural 261 (13.93) 120 (13.2) 141 (14.6)
**Consanguineous marriages**
First cousin
Yes 803 (42.87) 399 (44.0) 404 (41.8) 0.34
No 1,070 (57.13) 508 (56.0) 562 (58.2)
Second cousin
Yes 629 (33.58) 319 (35.2) 310 (32.1) 0.16
No 1,244 (66.42) 588 (64.8) 656 (67.9)
**Smoking cigarettes/water pipe**
Yes 170 (9.08) 88 (9.7) 82 (8.5) 0.36
No 1,703 (90.92) 819 (90.3) 884 (91.5)
**Exposure to sun**
Yes 728 (38.9) 332 (36.6) 396 (41) 0.05
No 1,145 (61.1) 575 (63.4) 570 (59)
**Daily physical activity**
Yes 719 (38.76) 323 (35.6) 396 (41.0) 0.05
No 1,154 (61.24) 584 (64.4) 570 (59.0)
**Supplementary vitamin D**
Yes 609 (32.51) 93 (10.3) 516 (53.4) \<0.001
No 1,264 (67.49) 814 (89.7) 450 (46.6)
**Notes:**
US\$1 = QR3.65.
######
Sociodemographic characteristics of pregnant women with complications like gestational diabetes mellitus (GDM), anemia, iron deficiency, and preeclampsia (n = 1,873)
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Variable GDM (n = 260), n (%) Anemia (n = 216), n (%) Iron deficiency (n = 162), n (%) Preeclampsia (n = 129), n (%)
------------------------------------------------------------------- ------------------------------------------------------- ------------------------------------------------------- ------------------------------------------------------ ------------------------------------------------------
**Nationality**
Qatari 137 (52.7) 115 (53.2) 90 (55.6) 67 (51.9)
Non-Qatari 123 (47.3) 101 (46.8) 72 (44.4) 62 (48.1)
**Age-group**
\<30 years 116 (44.6) 108 (50.0)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 73 (45.1) 52 (40.3)
30--34 years 50 (19.2) 45 (20.8) 35 (21.6) 30 (23.3)
35--39 years 54 (20.8) 30 (13.9) 36 (22.2) 25 (19.4)
40--45 years 40 (15.4) 33 (15.3) 18 (11.1) 22 (17.1)
**Education**
Illiterate 19 (7.3) 18 (8.3) 13 (8.0) 8 (6.2)
Primary 27 (10.4) 15 (6.9) 16 (9.9) 7 (5.4)
Intermediate 24 (9.2) 23 (10.6) 15 (9.3) 8 (6.2)
Secondary 79 (30.4) 78 (36.1) 45 (27.8) 49 (38.0)
College/university 111 (42.7) 82 (38.0) 73 (45.1) 57 (44.2)
**Occupation**
Housewife 172 (66.2) 143 (66.2) 111 (68.5) 75 (58.1)
Sedentary/professional Clerk 57 (21.9)\ 51 (23.6)\ 40 (24.7)\ 33 (25.6)\
31 (11.9) 22 (10.2) 11 (6.8) 21 (16.3)
**Monthly income (QR)**[†](#tfn3-ijwh-5-523){ref-type="table-fn"}
\<10,000 131 (50.4) 100 (46.3) 84 (51.9) 56 (43.4)
10,000--14,999 53 (20.4) 49 (22.7) 29 (17.9) 30 (23.3)
15,000--20,000 58 (22.3) 46 (21.3) 37 (22.8) 26 (20.2)
\>20,000 18 (6.9) 21 (9.7) 12 (7.4) 17 (13.2)
**Place of residence**
Urban 233 (89.6) 190 (88.0) 135 (83.3) 118 (91.5)
Rural 27 (10.4) 26 (12.0) 27 (16.7) 11 (8.5)
**Consanguineous marriages**
First cousin
Yes 129 (49.6)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 113 (52.3)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 78 (48.1)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 57 (44.2)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"}
No 131 (50.4) 103 (47.7) 84 (51.9) 72 (55.8)
Second cousin
Yes 95 (36.5)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 75 (34.7)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 52 (32.1)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 40 (31)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"}
No 165 (63.5) 141 (65.3) 110 (67.9) 89 (69)
**Smoking cigarettes/water pipe**
Yes 37 (14.2)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 32 (14.8)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 26 (16.0)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"} 20 (15.5)[\*](#tfn2-ijwh-5-523){ref-type="table-fn"}
No 223 (85.8) 184 (85.2) 136 (84.0) 108 (83.7)
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Notes:**
*P* \< 0.05 when compared with uncomplicated group;
US\$1 = QR3.65.
######
Serum levels of vitamin D among pregnant women according to the status of pregnancy complications
Serum level of vitamin D Gestational diabetes mellitus Anemia Iron deficiency Preeclampsia
-------------------------- ------------------------------- ------------ ----------------- -------------- ----------- ------------ ----------- ------------
**Severe deficiency**
25(OH)D \<10 ng/mL 43 (16.5) 178 (11.0) 37 (17.1) 184 (11) 30 (18.5) 191 (11.2) 21 (19.8) 198 (11.4)
**Moderate deficiency**
25(OH)D 10--19 ng/mL 99 (38.1) 587 (36.4) 81 (37.5) 596 (36) 49 (30.2) 637 (37.2) 31 (29.2) 647 (37.1)
**Mild deficiency**
25(OH)D 20--29 ng/mL 72 (27.7) 616 (38.2) 70 (32.4) 618 (37.3) 57 (36.4) 629 (36.8) 50 (38.8) 638 (36.6)
**Optimal**
25(OH)D 30--80 ng/mL 46 (17.7) 232 (14.4) 28 (13) 259 (15.6) 26 (13.3) 254 (14.8) 17 (13.2) 261 (14.8)
*P*-value \<0.003 \<0.002 \<0.033 0.031
**Abbreviation:** 25(OH)D, serum 25-hydroxyvitamin D.
######
Maternal and obstetric characteristics of pregnant women with complications: gestational diabetes mellitus (GDM), anemia, iron deficiency, and preeclampsia (n = 1,873)
Variable GDM (n = 260), n (%) Anemia (n = 216), n (%) Iron deficiency (n = 162), n (%) Preeclampsia (n = 129), n (%)
--------------------------------- ------------------------------------------------------- ------------------------------------------------------- ------------------------------------------------------- ------------------------------------------------------
**Body mass index (kg/m^2^)**
Normal 95 (36.5)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 85 (39.4) 53 (32.7) 37 (28.7)
Overweight 57 (21.9) 62 (28.7) 52 (32.1) 39 (30.2)
Obese 108 (41.5)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 69 (31.9)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 57 (35.2)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 53 (41.1)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"}
**Parity group**
\<2 64 (24.6) 66 (30.6)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 35 (21.6) 31 (24)
2--3 72 (27.7) 43 (19.9) 50 (30.9) 39 (30.2)
4--6 103 (39.6) 99 (45.8) 65 (40.1) 53 (41.1)
\>6 21 (8.1) 8 (3.7) 12 (7.4) 6 (4.7)
**Previous abortion**
Yes 65 (25.0)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 48 (22.2)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 28 (17.3) 31 (24.0)
No 195 (75.0) 168 (77.8) 134 (82.7) 98 (76.0)
**Previous stillbirth**
Yes 29 (11.2) 20 (9.3) 12 (7.4) 12 (9.3)
No 231 (88.8) 196 (90.7) 150 (92.6) 117 (90.7)
**Antenatal care**
Yes 205 (78.8) 185 (85.6)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 140 (86.4)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 90 (69.8)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"}
No 55 (21.2) 31 (14.4) 22 (13.6) 39 (30.2)
**Gestational group**
28--31 weeks 2 (0.8) 2 (0.9) 5 (3.1) 1 (0.8)
32--35 weeks 7 (2.7) 9 (4.2) 10 (6.2) 6 (4.7)
Above 36 weeks 251 (96.5) 205 (94.9) 147 (90.7) 122 (94.6)
**Premature ruptured membrane**
Yes 35 (13.5)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 10 (4.6) 9 (5.6) 9 (7.0)
No 225 (86.5) 206 (95.4) 153 (94.4) 120 (93.0)
**History of preterm delivery**
Yes 44 (16.9)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 23 (10.6) 21 (13.05)[\*](#tfn5-ijwh-5-523){ref-type="table-fn"} 8 (6.2)
No 216 (83.1) 193 (89.4) 141 (87.0) 121 (93.8)
**Note:**
*P* \< 0.05 when compared with uncomplicated group.
######
Predictors for gestational diabetes mellitus, anemia, iron deficiency, and preeclampsia in studied pregnant women using multivariate logistic regression (n = 1,873)
Independent variable Odds ratio 95% confidence interval *P*-value
----------------------------------- ------------ ------------------------- -----------
**Gestational diabetes mellitus**
Premature ruptured membrane 3.2 (2.0--5.0) ≤0.001
Previous low birth weight 1.8 (1.2--2.6) ≤0.001
Previous stillbirth 1.5 (0.98--2.4) 0.06
Previous abortion 1.4 (1.0--2.0) 0.03
Vitamin D deficiency 1.4 (1.1--1.8) 0.02
Consanguinity (first cousin) 1.3 (1.0--1.8) 0.03
**Anemia**
Vitamin D deficiency 1.87 (1.3--2.6) ≤0.001
Consanguinity (second cousin) 1.64 (1.2--2.2) ≤0.001
Lack of vitamin D supplements 1.46 (1.0--2.1) 0.04
Less antenatal care 1.43 (0.95--2.2) 0.09
Nationality 1.33 (0.99--1.8) 0.05
Age-group 1.12 (0.98--1.3) 0.09
**Iron deficiency**
Gestational weeks 2.77 (1.78--4.4) ≤0.001
Previous low birth weight 2.17 (1.4--3.3) ≤0.001
Less antenatal care 1.93 (1.2--3.1) 0.01
Vitamin D deficiency 1.84 (1.1--3.3) 0.02
Nationality 1.52 (1.1--2.1) 0.01
Occupation 1.30 (1.0--1.6) 0.02
Age-group 1.23 (1.0--1.5) 0.02
**Preeclampsia**
Vitamin D deficiency 1.75 (1.2--2.6) 0.01
######
Vitamin D status during pregnancy or at delivery: global variations and comparisons
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Study Country Sample size Sun exposure/vitamin D intake Mean, ng/mL Mean, nmol/L
------------------------------------ ------------------------------------ ------------- -------------------------------------------------------------------- ----------------------------------------------- -------------------------------------------------
Dawodu and Akinbi[@b3-ijwh-5-523] United Arab Emirates/Arabs 192 Lack of sun exposure/low vitamin D intake 8.2 20.5
Jiang et al[@b6-ijwh-5-523] People's Republic of China/Chinese 152 Low sun exposure/low vitamin D intake 9.1\ 22.7\
12.7 31.8
Holmes et al[@b7-ijwh-5-523] UK/Caucasian 99 Low supplementation NK NK
Javaid et al[@b9-ijwh-5-523] UK/Caucasian 160 Low UV exposure/low rate of vitamin D supplementation in pregnancy NK NK
Bowyer et al[@b10-ijwh-5-523] Australia/mixed 971 Low sun exposure/sunscreen use/vitamin D intake 21[\*](#tfn6-ijwh-5-523){ref-type="table-fn"} 52.0[\*](#tfn6-ijwh-5-523){ref-type="table-fn"}
Hamilton et al[@b11-ijwh-5-523] USA/mixed 559 Low sun exposure 21.7 54.3
Sahu et al[@b12-ijwh-5-523] India/Indian 139 Inadequate exposure/low vitamin D/Ca intake 12.7 31.8
Viljakainen et al[@b16-ijwh-5-523] Finland/Caucasian 124 Low UVB exposure/inadequate vitamin D intake 16.4 41.0
Bassir et al[@b35-ijwh-5-523] Iran/Iranian 50 Lack of sun exposure/low vitamin D intake 5.1 12.8
Newhook et al[@b36-ijwh-5-523] Canada/Caucasian 50 Low UV exposure/low vitamin D supplementation 20.7/24.4 51.9/61.1
Ginde et al[@b37-ijwh-5-523] USA/mixed 928 Low outdoor activity/low vitamin D intake 26 65.0
Zhang et al[@b38-ijwh-5-523] People's Republic of China 953 Low sun exposure/low vitamin D intake 24.2/31 24.2
Parlea et al[@b39-ijwh-5-523] Canada 116 Low UV exposure/low vitamin D supplementation 24 24
Merewood et al[@b40-ijwh-5-523] USA 459 Low outdoor activity/low vitamin D intake 24.8 24.8
Burris et al[@b41-ijwh-5-523] USA 1,087 Low outdoor activity/low vitamin D intake 23.6 59
Nicolaidou et al[@b42-ijwh-5-523] Greece 123 Lack of sun exposure/low vitamin D intake 6.6--8.2 16.4--20.4
Halicioglu et al[@b43-ijwh-5-523] Turkey 258 Lack of sun exposure/low vitamin D intake 11.5 11.5
Molla et al[@b44-ijwh-5-523] Kuwait 128 Low UVB exposure/low vitamin D/Ca intake 13.3 33.3
Bener et al (present study) Qatar 767 Lack of sun exposure/low vitamin D intake 17.8 44.5
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Notes:**
Median; to convert serum 25-hydroxyvitamin D to ng/mL, divide by 2.5.
**Abbreviations:** NK, not known; UVB, ultraviolet B.
|
The work is exhibited in a gallery space as a presentation of my own outlook on two polarities; the cultural scene and the government of the Republic of Croatia. By dissecting the work one can observe that there are several components which are bound to each other and, as such, indivisible. The visual impression is limited to only two achromatic colors, two polarities; white and black color, stressing the above stated relationship which I often bring into question in my works. Although whiteness prevails in the work (the sheet of the tent), one can notice that it is actually supported by the black foundation (the poles), of almost negligible quantity, but without which the tent would not be stable, and be what it is. The front entrance into the tent has been sewn shut from the outside with black thread, once again of almost negligible blackness compared to the white sheet of the tent. The black label (Made in Croatia / Property of the Ministry of Culture) reminds us that we are, on more than one level, dependent on the Mother State, as our upbringing and practice tell us. Can a loveless child be a contended child? In other words, can a cultural scene exist without the state and that relationship which becomes more questionable year after year?! The state creates the generally accepted social conditions, considering culture to be a fringe element of national interests, a burden, something redundant, or it does not consider it at all when all is said and done. The snubbing attitude of the state towards culture raises questions such as whether this society needs anything cultural, should we nurture culture, rest our hopes on it, or simply leave it to be its own purpose, and prescribe it in such doses which would ensure that the tent occupants do not protest as they repeat their mantra: “Shut up, things will get better!” On the other hand, is there a limit after which pleasure becomes pain, and I ask myself whether we had already passed it?
Is the state or culture the one with a problem? Or maybe better days will arrive when art will, (un)fortunately, disappear from people’s lives in the Republic of Croatia.
|
Socket sclerosis: a rare complication obstructing orthodontic tooth movement.
Socket sclerosis is usually asymptomatic and does not require any treatment. The only potential complication arises during orthodontic treatment, wherein a sclerosed socket of the premolars may be an obstacle in closing spaces by movement of teeth through the extraction space. This article demonstrates the problems encountered during the orthodontic treatment of a 20-year-old woman with socket sclerosis and the treatment strategy employed to overcome the same.
|
If you are not satisfied with your purchase on www.salomon.com, you can return it and receive a full refund (excluding return costs) by completing the return procedure on our online platform within 30 days of receiving the product. The items must be returned within the stated period, and must be intact and complete. They must not have been used (or washed or worn) and must be in their original packaging in good condition.
If you are returning goods because the product is damaged or is not the product you ordered, please contact customer services for a refund of your shipping costs, at 1 833 550 2736 from 7 a.m. to 5 p.m MST Monday to Friday.
Do you have any questions ?
Our product experts are available to help you find the product that best fits your needs and to answer all your questions from Monday to Friday, 7am to 5pm
|
You only need to add some assumptions to your expressions, since for some set
of values the Log[] is not defined
For example use
FullSimplify[Log[x^n] - n*Log[x], Assumptions -> x > 0 && n > 0]
and it will return 0 as expected
yehuda
On Sunday 30 January 2005 10:18, Cyrus Erik Eierud wrote:
> Please Help!
>
> Thanks for all great answers I've already found here!
> My problem is that I can not simplify what to me seems as a very
> simple equality task. This is what I want Mathematica to return zero
> for:
>
> in[1]:= FullSimplify[Log[x^n] - n*Log[x]]
>
> out[1]= -n Log[x] + Log[x^n])
>
> I have used Simplify to check equalities, but the one above (and many
> other equations similar to the one above) just don't simplify. Am I
> doing anything wrong or does anyone know of a better way to check
> equalities?
>
> Appreciate any help,
> Cyrus Eierud, Student
> cyruserik at tele2.se
|
1. Field of the Invention
The present invention relates to a multiple transistor consisting of a vertical transistor and a lateral transistor, and also to a method of manufacturing the multiple transistor.
2. Description of the Prior Art
In general, where a plurality of vertical transistors, each having a different current gains (h.sub.FE) are formed within a semiconductor wafer or chip by the use of a diffusion process, the h.sub.FE s of of the respective transistors do not considerably differ, even when the emitter area is varied using the same pattern. Therefore, unless the number of diffusion steps is increased to make the base width and the emitter impurity concentration different, transistors having different h.sub.FE s cannot be formed. An increase in the number of diffusion steps, however, is unpreferable due to a decrease in the yield of finished products and an increase in cost.
For such reasons, it has been impossible to manufacture transistors having arbitrary h.sub.FE s within a semiconductor wafer or chip under the same diffusion conditions, and hence, the degree of freedom in design for h.sub.FE has been low.
|
Q:
Can str_replace be safely used on a UTF-8 encoded string if it's only given valid UTF-8 encoded strings as arguments?
PHP's str_replace() was intended only for ANSI strings and as such can mangle UTF-8 strings. However, given that it's binary-safe would it work properly if it was only given valid UTF-8 strings as arguments?
Edit: I'm not looking for a replacement function, I would just like to know if this hypothesis is correct.
A:
Yes. UTF-8 is deliberately designed to allow this and other similar non-Unicode-aware processing.
In UTF-8, any non-ASCII byte sequence representing a valid character always begins with a byte in the range \xC0-\xFF. This byte may not appear anywhere else in the sequence, so you can't make a valid UTF-8 sequence that matches part of a character.
This is not the case for older multibyte encodings, where different parts of a byte sequence are indistinguishable. This caused a lot of problems, for example trying to replace an ASCII backslash in a Shift-JIS string (where byte \x5C might be the second byte of a character sequence representing something else).
A:
It's correct because UTF-8 multibyte characters are exclusively non-ASCII (128+ byte value) characters beginning with a byte that defines how many bytes follow, so you can't accidentally end up matching a part of one UTF-8 multibyte character with another.
To visualise (abstractly):
a for an ASCII character
2x for a 2-byte character
3xx for a 3-byte character
4xxx for a 4-byte character
If you're matching, say, a2x3xx (a bytes in ASCII range), since a < x, and 2x cannot be a subset of 3xx or 4xxx, et cetera, you can be safe that your UTF-8 will match correctly, given the prerequisite that all strings are definitely valid UTF-8.
Edit: See bobince's answer for a less abstract explanation.
|
Q:
Mysqli to PDO - If else statement to try catch
I'm trying to convert some mysqli code to PDO.
Before I had this:
$updateTaskQuery = "UPDATE `task` SET `user_id` = {$query['user_id']}, `status_id` = {$query['status_id']} WHERE `id` = {$query['task_id']}";
$updateTask = mysqli_query($mysqli, $updateTaskQuery);
//### Check for error
if($mysqliError = mysqli_error($mysqli)) {
echo json_encode(array('error' => 'Update Task MySQLi Error: '.$mysqliError));
exit;
} else {
echo json_encode(array('success' => true));
exit;
}
So far, I have convertet it to this:
$sql = $db->prepare ( "UPDATE `task` SET `user_id` = {$query['user_id']}, `status_id` = {$query['status_id']} WHERE `id` = {$query['task_id']}" );
$sql->execute ();
$updateTask = $sql->fetchAll ( PDO::FETCH_ASSOC );
try {
$updateTask;
} catch ( PDOException $ex ) {
//handle
}
My question here is, how can I include my else statement in the new code?
UPDATE: Wrong, but working code
try {
// Update task
$query = $db->prepare ( "UPDATE `task` SET `user_id` = {$query['user_id']}, `status_id` = {$query['status_id']} WHERE `id` = {$query['task_id']}" );
$query->execute();
echo json_encode ( array (
'success' => true
) );
} catch ( PDOException $e ) {
// catch a pdo error
echo json_encode ( array (
'error' => 'Update Task PDO Error: ' . $e->getMessage (),
'error_trace' => $e->getTraceAsString ()
) );
}
A:
you have a lot wrong here, I will see If i can help - I haven't "tested" this but it should be really close.
//your using json so set the correct content headers
header('Content-Type: application/json');
//try only works on code you "try" anything outside isn't included in the try block
try{
//connect
$PDO = new PDO(
'mysql:host=' . $_dbHost . ';' . 'dbname=' . $_dbName,
$_dbUser,
$_dbPass
);
//set exception error mode
$PDO->setAttribute(PDO::ATTR_ERRMODE, PDO::ERRMODE_EXCEPTION);
//set fetch assoc array as default
$PDO->setAttribute(PDO::ATTR_DEFAULT_FETCH_MODE, PDO::FETCH_ASSOC);
//query with named placeholders ( even pdo can get sql injection when you dont use placeholders )
$sql = 'UPDATE `task` SET `user_id` = :user_id, `status_id` = :status_id WHERE `id` = :id';
//prepare query
$stmt = $PDO->prepare( $sql );
//execute with data
$stmt->execute(array(
':user_id' => $query['user_id'],
':status_id' => $query['status_id'],
':id' => $query['id'],
));
echo json_encode(array(
'success' => true,
'results' => $stmt->fetchAll()
)
);
}catch( PDOException $e ){
//catch a pdo error
echo json_encode(array(
'error' => 'Update Task PDO Error: '.$e->getMessage(),
'error_trace' => $e->getTraceAsString()
)
);
}catch( Exception $e ){
//the beauty of exceptions ~ catch some other non-pdo exception
echo json_encode(array(
'error' => 'Runtime Error: '.$e->getMessage(),
'error_trace' => $e->getTraceAsString()
)
);
}
///more code can go here
To answer your question, you don't need the else statement any longer. Everything in the try block with run until it throws an exception, then it will land in the appropriate catch block and run that code. In that block you can get the error messages by referencing the variable you set for the exception class instance to use, in this case I set $e. This is a common convention to use $e just as it is to use $i for iteration on a loop. Generally I wont use short variables like this except in cases where the variable is not a integral part of your code. In this case $e is not something I expect to use outside of the catch.
I also included the proper content header. This will help javascript libraries such as jQuery parse the JSON properly when they return the data.
I love these kinds of questions, so I hope my explanation has helped you to understand PDO and Exceptions a bit more. Your on the right track using them.
One last note is you can run more code after the try catch blocks.
To explain my comment ( about wrong code that works ), when you don't sanitize input in your SQL you get nasty things like this. Assume we have this query:
$sql = "UPDATE
`task`
SET
`user_id` = {$query['user_id']},
`status_id` = {$query['status_id']}
WHERE
`id` = {$query['task_id']}
";
The problem here is if someone enters a chunk of sql into one of your inputs such as
$query['task_id'] = '0; DROP table task;';
What this does is complete your query and drop your table!
$sql = "UPDATE
`task`
SET
`user_id` = 1,
`status_id` = 2
WHERE
`id` = 0;
DROP table task;
";
While I am by no means and expert at SQL injection, as I never have done it, that's the gist of it. There are a lot of things they can do worse then dropping tables, such as creating database users. Accessing files on the system, such as password files for system users or other confidential information. Not to mention a whole plethora of second stage attacks such as adding malicious javascript. Which would be printed on the screen and allow them to do XSS ( cross site scripting ) type attacks on site visitors etc...
Trust me even if you think you don't have anything worth "stealing" on your site. It's better to learn the right way then the hard way on something like this. Sorry if I'm a bit harsh on this but it is a very important concept to learn if you plan to do any "serious" coding.
Just because something doesn't work strait off, such as a code "sample" that I provided ( I did explicitly say I didn't test it ). Doesn't mean there is a fundamental problem with it ( it was a simple copy paste error ). The "wrong" code that works is fundamentally flawed.
|
Archive
Court rejects 10 Commandments bid
The US Supreme Court has rejected a long-disputed bid by Alabama's top judge to display a monument inscribed with the 10 Commandments at the state judicial building.
04 Nov 2003 00:22 GMT
Judge Roy Moore remains suspended
The high court refused to hear two appeals by Judge Roy Moore on Monday, who was suspended in August.
Moore suspension from work came after his refusal to comply with a federal judge's order to remove the granite monument because it violated constitutional church-state separation.
The judge, who was elected as Alabama's chief justice in 2000, has been accused of violating judicial ethics.
He is fighting his suspension, and a hearing has been set for 12 November in Montgomery, the state capital.
Court order removal
The 2.3 tonne stone marker was placed in the rotunda of the judicial building in Montgomery by Moore and his followers in 2001.
The monument was removed on 27 August after eight associate justices of the Alabama Supreme Court decided to comply with an order by US District Court Judge Myron Thompson.
It has been locked in a closet since then.
Moore, who has strong support from fundamentalist Christian groups for his championing of public displays of the 10 Commandments, vowed on Monday to continue his struggle in the court of public opinion.
Fundamentalist
"The United States Supreme Court may turn a deafening ear to the people of Alabama, but they will not be able to avoid the thundering roar that's going to be heard across this country," Moore said in a news conference outside Montgomery.
Moore's supporters also pledged to place the issue on the national agenda.
Civil liberties groups, who have accused Moore of imposing his personal religious views on others and using his office to promote the Christian faith, are equally adamant about resisting such attempts.
The controversial monument isunder lock and key
A number of civil liberties groups sued in 2001 on behalf of three Alabama lawyers who said they had been offended by the monument's display.
Thompson and then a US appeals court in Atlanta ruled the monument must be removed.In one appeal, Moore's lawyers told the Supreme Court the federal judge lacked jurisdiction to issue an injunction to Alabama's chief justice requiring the monument's removal.
In the second appeal, his lawyers urged the high court to overrule its long-standing test dating back to 1971 on what represents an unconstitutional government endorsement of religion.
According to the original constitutional definitions, they said this monument would not represent a government "establishment of religion" forbidden by the First Amendment.
Controversy over?
One group involved in the lawsuit, Americans United for Separation of Church and State, said the high court's action meant Moore has lost his final appeal and the legal controversy is over.
"This is the end of the legal line for Roy Moore," said the Rev. Barry Lynn, the group's executive director.
"It is time for Moore to face facts: he's on the wrong side of the Constitution. Religious symbols belong in our homes and houses of worship, not our courthouses."
In August, the Supreme Court rejected a request by Moore for a last-minute stay to prevent the monument's removal.
The high court in recent years has let stand other US appeals court rulings that have barred the display of large granite monuments engraved with the Ten Commandments on the grounds of state capitals.
|
Parameters
Name
Type
Description
Title
String
Name of the playlist to get (use "" to get root of all playlists)
Property description
Retrieves SDBPlaylist object of the given playlist name. Probably most often used parameter is an empty string, the root (virtual) playlist node is returned then, you can use its ChildPlaylists property to get all the first level playlists.
|
Reinfeldt added that the Titans have “stressed to him the importance that he be here soon.
Titans coach Jeff Fisher said Britt is close by and if a deal is reached could arrive in time to practice Saturday.
“It looks like things are falling into place around him, I’m optimistic,” Fisher said. “Before practice I visited with Mike, and he felt like things were moving, so I’d say we have a decent chance of having him out here tomorrow.”
Meanwhile, center Kevin Mawae began training camp on the active PUP (physically unable to perform) list and said he hopes to return in two to three weeks.
“I don’t know. I’d say a couple to three weeks. Hopefully, I’ll get some time in before the preseason is up. I don’t have a timeline or a deadline or anything like that. My goal is opening day,” Mawae said.
Mawae said the six-month recovery time from the triceps tendon surgery was a success and now he is involved in the rehab process to get the muscle strong again.
“Once you get through the recovery time, you’ve got to build the muscle back up and build the muscle endurance back to where it needs to be,” Mawae said.
The only other Titans player to miss practice on the first day was kicker Rob Bironas, who was backed down because of a minor muscle pull. Fisher said the injury was not related to the groin injury that kept Bironas out for a good portion of camp and preseason last year.
“He’s just a little sore. We’ll hold him out for a couple of days. He’ll kick. There’s nothing to be concerned about,” Fisher said.
|
Q:
VS2015 adding dependency to project.json doesn't show Autocomplete
I am following along to a video and the user types in 'static' in the project dependencies section and gets Autocomplete for "Microsoft.AspNet.StaticFiles": "1.0.0-rc1-final" in the project.json file dependencies section but as I am typing along I don't get any Autocomplete. Its the same issue when I try to add "Microsoft.AspNet.Mvc"...I don't get the Autocomplete, so I don't even know what version to manually type in. Is this user error?
"dependencies": {
"Microsoft.AspNet.IISPlatformHandler": "1.0.0-rc1-final",
"Microsoft.AspNet.Server.Kestrel": "1.0.0-rc1-final",
"static"
},
expected Autocomplete to provide this:
"dependencies": {
"Microsoft.AspNet.IISPlatformHandler": "1.0.0-rc1-final",
"Microsoft.AspNet.Server.Kestrel": "1.0.0-rc1-final",
"Microsoft.AspNet.StaticFiles": "1.0.0-rc1-final"
},
This is what I get with the AutoComplete right now for EntityFramework:
Update:
I am following along code from a Lynda.com Up and running with ASP.Net 5 video series, and the lesson video shows this as autocomplete:
But when I follow along, I don't get the same:
A:
The solution was to make sure package sources are enabled/checked in Visual Studio under Tools -> Options -> NuGet Package Manager -> Package Sources.
My Machine-wide Packages/Microsoft and .NET was not checked (which is kinda strange because I did do a default install) but checking that solved my issue.
|
Quantitative stability analyses of multiwall carbon nanotube nanofluids following water/ice phase change cycling.
Multiwall carbon nanotube nanofluids are regularly investigated for phase change enhancement between liquid and solid states owing to their improved heat transfer properties. The potential applications are numerous, the most notable being latent heat thermal energy storage, but the success of all nanofluid-assisted technologies hinges greatly on the ability of nanoparticles to remain stably dispersed after repeated phase change cycles. In this report, the stability of aqueous nanofluids made from oxygen-functionalized multiwall carbon nanotubes (f-MWCNTs) was profiled over the course of 20 freeze/thaw cycles. Sonication was used after each cycle to re-disperse clusters formed from the crystallization process. This study offers a quantitative evaluation of f-MWCNT-nanofluid stability as a result of phase change through optical characterization of concentration and particle size. It also provides insight into the integrity of the surface functionalities through zeta potential and XPS analyses. Concentration and particle size measurements showed moderate and consistent recoverability of f-MWCNT dispersion following ultrasonication. XPS measurements of solid-state MWCNTs exposed to freeze/thaw cycling in water, and zeta potential analyses of the nanofluids indicate that the surface oxygen content is preserved throughout phase change and over repeated cycles. These results suggest a resilience of oxygen-functionalized MWCNTs to the freezing and thawing of water, which is ideal for their utilization as phase change enhancers.
|
HomeAway.co.uk, the home of Holiday-Rentals®, is part of the HomeAway family. As the world leader in holiday rentals, we offer the largest selection of properties for any travel occasion and every budget. We’re committed to helping families and friends find the perfect holiday rental to create unforgettable travel experiences together.
3BR Direct Oceanfront Condo
3BR Direct Oceanfront Condo
Shalimar is an eight-story, oceanfront, condominium complex and is the perfect place for your family's summer vacation or for a golf outing in the spring or fall. The Myrtle Beach area has over 100 golf courses from casual to challenging.
Shalimar is located in the Cherry Grove Section of North Myrtle Beach, South Carolina. Cherry Grove is known for its relaxing family atmosphere and large beaches. It is a great place to relax and watch the pelicans and dolphins. The fishing pier and shopping are only a short walk away.
Our three-bedroom, two-bath condominium is on the eighth floor which allows for magnificent, unobstructed ocean views. There is an excellent open floor plan that includes a fully-stocked kitchen, dining room, and living room. You can enjoy an ocean view from the kitchen, living room, master bedroom, and the second bedroom. The third bedroom has a view of the town and marsh. The huge balcony is accessible from the kitchen, living room, and master bedroom. We ask that you refrain from smoking inside the condominium.
The master bedroom has a king-size bed and a whirlpool tub. The second bedroom has two twin beds and the third bedroom has a queen-size bed.
Complimentary high-speed wireless internet.
During the peak season summer months, the condominium is available by the week only (Saturday to Saturday). Monthly rates available during the winter.
Pets, motorcycles, and trailers are not permitted.
Keywords: Condominium
Overview
Details
condo
1,200
sq. ft.
People
6
Bedrooms
3
Bathrooms
2
Minimum Stay
3 - 7 nights
About the property
Condo, 3 Bedrooms, 2 Baths, (Sleeps 6)
Shalimar is an eight-story, oceanfront, condominium complex and is the perfect place for your family's summer vacation or for a golf outing in the spring or fall. The Myrtle Beach area has over 100 golf courses from casual to challenging.
Shalimar is located in the Cherry Grove Section of North Myrtle Beach, South Carolina. Cherry Grove is known for its relaxing family atmosphere and large beaches. It is a great place to relax and watch the pelicans and dolphins. The fishing pier and shopping are only a short walk away.
Our three-bedroom, two-bath condominium is on the eighth floor which allows for magnificent, unobstructed ocean views. There is an excellent open floor plan that includes a fully-stocked kitchen, dining room, and living room. You can enjoy an ocean view from the kitchen, living room, master bedroom, and the second bedroom. The third bedroom has a view of the town and marsh. The huge balcony is accessible from the kitchen, living room, and master bedroom. We ask that you refrain from smoking inside the condominium.
The master bedroom has a king-size bed and a whirlpool tub. The second bedroom has two twin beds and the third bedroom has a queen-size bed.
Complimentary high-speed wireless internet.
During the peak season summer months, the condominium is available by the week only (Saturday to Saturday). Monthly rates available during the winter.
Great condo!
I had the duty of renting three condos for a large family vacation. This is the condo the my mother-in-law and sister-in-laws family stayed. My mother-in-law doesn't normally like condos and usually rents houses and she absolutely LOVED it! She is already talking about renting the same ones for next year. My sister-in-law says that the kitchen was stocked, condo was exceptionally clean and the balcony was wonderful with a constant breeze. She was very satisfied with this condo and is looking forward to vacation next year. This is the first time that we have rented thru VRBO and I can't tell you how easy it was to deal with all three owners. This particular unit you pick up everything at Cherry Grove Realty and my sister-in-law stated that they were WONDERFUL! We have always had dealings with a larger realty there and we always seem to have some issue but, this year was definitely a different experience. It was our most relaxing vacation in 19 years.
better than expected...
This unit in Shalimar was nicer than expected-the pictures did not do justice for the view and we were pleasantly surprised by the size of the condo. You cannot beat the view of the ocean from the kitchen window, the living room and the master bedroom. When we reserved, the owners said they were coming to stay after we left and were going to do some updates and minor repairs. This should make any future guest's stay even better. The condo is in a quiet area of Myrtle Beach, compared to the downtown area. We would definitely stay at this condo again!! Amanda Johnson
Review Submitted: May 4, 2011
Date of Stay: Apr 2011
Source: VRBO, from HomeAway
Owner's Response: We are glad you enjoyed your stay. We did go down right after you and installed a new glasstop stove in the kitchen. We also had the couch cushions re-stuffed and put some new furniture in the twin room.
Ideal location with fantastic views from living and master bedroom
A very clean and comfortable condo steps to the beach through a front sunning area and small pool, disappointingly unheated. Very quiet, even though several other units in the building are dog-friendly, barking was never a problem. Construction next door could have detracted from the enjoyment were the weather warm enough to spend more time on the balcony.
We especially appreciated the brief walk to the fishing pier and slightly longer one to the large grocery store containing all that would be needed for a week's stay. Views are spectacular but if you must see the sunrise you'll have to either go out on the beach or rent a unit on the other side of the building!
The living area was roomy enough for three of us, but it's questionable how six people could fit on the two huge overstuffed loveseats which had little support and I found to be most uncomfortable. Lighting needs improvement...small bulb in lamp between loveseats was inadequate for reading. Floor lamp had high wattage bulb with inadequate shade.
Kitchen was very well furnished. Trash compactor didn't work. Refrigerator had water leaking onto the shelf continuously.
Beds were very comfortable, with adequate blankets. We brought our own linens and saved a few dollars.
Jacuzzi was a real plus for my visitors.
Wi-Fi reception was excellent! For one who forgot the cell phone charger, having a phone for local and toll-free calls was handy!
Review Submitted: Apr 8, 2011
Date of Stay: Apr 2011
Source: VRBO, from HomeAway
Owner's Response: We are delighted that you enjoyed your stay and thank you for your comments. We appreciate all comments so we can make future guests' experiences even better. It is true that the sunrise is obstructed from our balcony. On the other hand, the sunsets are spectacular! We are going to the condo this month and will do our best to address the couches and the lighting situation. We have already repaired the leak in the refrigerator (it was a blocked defrost tube). We are sprucing up the kitchen a bit with a new glass top stove when we get down there. Unfortunately, the trash compactor has never worked and we don't plan on fixing it. Our plan is to remove it in the near future when we put a new floor in.
Our goal is to offer our guests a first class rental experience. We hope that you will stay with us again!
|
Q:
Matching from a starting delimiter to an end delimiter (Regex/pattern matching)
I am trying to match for a certain block of text.
The format of the text I want to match is
<pevz:url>https://some.server.com/arbitraryFoo.jpeg</pevz:url>
where only <pevz:url> and </pevz:url> are known.
My naive try was to match with
<pevz:url>*([0-9a-zA-Z:/._-])<\/pevz:url>
but that didn't work. I am using gedit to match with the default search and replace (no advanced-find).
How can I match for the whole string?
Best regards,
Joe Cocker
A:
You can try:
<pevz:url>(.*?)<\/pevz:url>
or
<pevz:url>([^>]+)<\/pevz:url>
Regex Demo
|
/*
* Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
*
* Licensed under the Apache License, Version 2.0 (the "License").
* You may not use this file except in compliance with the License.
* A copy of the License is located at
*
* http://aws.amazon.com/apache2.0
*
* or in the "license" file accompanying this file. This file is distributed
* on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either
* express or implied. See the License for the specific language governing
* permissions and limitations under the License.
*/
package software.amazon.smithy.diff.evaluators;
import static org.hamcrest.MatcherAssert.assertThat;
import static org.hamcrest.Matchers.equalTo;
import java.util.List;
import org.junit.jupiter.api.Test;
import software.amazon.smithy.diff.ModelDiff;
import software.amazon.smithy.model.Model;
import software.amazon.smithy.model.shapes.ResourceShape;
import software.amazon.smithy.model.shapes.StringShape;
import software.amazon.smithy.model.validation.ValidationEvent;
public class ChangedResourceIdentifiersTest {
@Test
public void detectsChangedIdentifiers() {
StringShape id = StringShape.builder().id("foo.baz#Id").build();
ResourceShape r1 = ResourceShape.builder().id("foo.baz#R").addIdentifier("a", id).build();
ResourceShape r2 = ResourceShape.builder().id("foo.baz#R").addIdentifier("b", id).build();
Model modelA = Model.assembler().addShapes(r1, id).assemble().unwrap();
Model modelB = Model.assembler().addShapes(r2, id).assemble().unwrap();
List<ValidationEvent> events = ModelDiff.compare(modelA, modelB);
assertThat(TestHelper.findEvents(events, "ChangedResourceIdentifiers").size(), equalTo(1));
}
}
|
A bit of history
The League of Legends Masters Series (LMS) made huge strides in 2015, going from perennial group-stage dropout to a realistic threat, making it out of groups at three straight international tournaments. In all of these instances, the team was generally not favored, but for the first time in what seems like forever, a Taiwanese representative seems like it might defy that trend.
Most see Flash Wolves as a team that should progress to the knockout round at this year's Mid-Season Invitational. Editor's Picks Nuovo Gaming launches Oceanic esports organization, signs iaguz and Tomflr
Pack leader - Silent Wolf still inspires Fox players after a decade
Making varsity: Collegiate esports primed for growth and exposure 2 Related
The LMS and Flash Wolves have both come a long way in the year and a half since Taiwan, Hong Kong and Macao seceded from the Southeast Asian Garena Premier League (GPL). The GPL had its time in the sun in Season 2, as the Taipei Assassins procured that season's World Championship title and third place at the famed double-elimination IPL 5. However, it lacked any significant depth past the Taipei Assassins, and Taiwanese teams still dominated the GPL. There was a glimmer of hope for the region when resources an influx of resources made for vast improvements, but that proved to be merely a honeymoon period.
Those newfound resources would dried up due to Riot's seeding system, which only accounts for the GPL and not the more competitive Taiwanese League of Legends Nova League (LNL). The league's skill ceiling began to stagnate as a result of the lack of competition in the GPL, caused by region-based slots keeping the more competitive Taiwanese teams away from another World Championship.
Flash Wolves was even denied the opportunity to play in the GPL, despite being stronger than numerous Southeast Asian teams. The eventual implementation of the LMS in 2015 would prove to be a real boon for Taiwanese League of Legends and Flash Wolves, as it sought to redeem itself after being rejected by the GPL at the end of 2014.
Flash Wolves did more than redeem itself -- it became the best team in the LMS and posted a stunning 19-2 record on its way to earning the first seed in its first LMS split. A freshly age-eligible Huang "Maple" Yi-Tang and new addition from solo-queue, a Rengar god, Hung "Karsa" Hau-Hsuan, catapulted the young team to the top. Unfortunately, the team was met with the unexpected opposition from a surging ahq e-Sports Club in the 2014 Spring LMS Finals and was promptly dismantled 3-1.
After all that time on top, it was put down with hardly a whimper in response.
Since then, Flash Wolves struggled to defeat ahq e-Sports Club, unable to overtake it the entirety of the following Summer Split. However, Flash Wolves did manage to keep pace and attend 2015 World Championships alongside its fellow countrymen, making it out of groups with ahq.
Now, several months removed from the World Championship, Flash Wolves is the one that has come out on top in the LMS Spring Split, having defeated first seed, ahq e-Sports Club, 3-0 in the LMS Spring Finals.
Growth in Spring Split
Karsa is all smiles after defeating Pain Gaming during the League of Legends World Championships 2015. Provided by Riot Games
It wasn't easy for Flash Wolves. It still struggled against its rival during the regular season, extending its losing streak against ahq to 11 before finally taking a game at the end of the season. Its 0-2 opener against ahq was disappointing to say the least, as ahq's star top laner, Chen "Ziv" Yi took down opposing top laner, Yu "MMD" Li-Hung, in the 1v1 and snowballed the victory on his own merit.
Maple had his own standout performance during the series, easily out-laning Wong "Chawy" Xing-Lei and Liu "Westdoor" Shu-Wei, but it wasn't enough to keep the team afloat. Yet again, Flash Wolves would be unable to deal with ahq, with the skill discrepancy too large.
Among other things, the team finally made attempts to improve its line-up, with the retirement of Chou "Steak" Lu-Hsi and attempted inclusion of Huang "Breeze" Chien-Yuan at AD carry.
It's well-known that the core of Flash Wolves' roster at the Season 5 World Championship dated back to the original Season 3, lightning-in-a-bottle, Gama Bears. Moving on from years of synergy including Hsiung "NL" Wen-An and Steak was not so simple. While the team was theoretically improving in its respective positions, Hu "SwordArt" Shuo-Jie and Karsa were expected to lead for the first time while Maple was the only consistent factor in Flash Wolves' games.
Disconnect plagued Flash Wolves early on, as it attempted to assimilate its new players. MMD and Breeze were clear improvements individually, as they performed well, but it frankly didn't matter. Flash Wolves lost and fell to its lowest low, losing 0-2 to newly promoted team, XGamers, in a stunning upset. It looked the worst it ever had, with Karsa fumbling numerous ganks and SwordArt being downright invisible.
It was tough enough for Karsa and SwordArt to adjust to primary shot-calling duties for the first time, much less assimilating new members into the roster. That hardly stopped the Wolves though, as the 0-2 loss to XG seemingly set off a new initiative. Finally settled on a starting roster featuring NL over Breeze, the team hoped to stabilize its play. The puppies subsequently picked it up and only lost three more games in the Spring Split as the roster began to develop a new identity for itself in 2016.
Flash Wolves began to put the "Flash" in its team title, becoming one of the stronger early game teams in the league. Karsa and SwordArt finally found their footing as the best one-two punch roaming duo in the league. They became known for forcing tower dives and driving just about any advantage home, leading the league with an average win time of 30:42. Meanwhile, the team's first place counterparts, ahq, was playing a much slower game and limping through the early game, pulling out wins with superior team-fighting and late game macro.
The style began to stagnate for ahq while Flash Wolves innovated its play, continuously taking steps to overcome its longtime rival.
Dethroning ahq and moving on to MSI
SwordArt when he was on the Flash Wolves. Provided by Riot Games
Flash Wolves eventually made its claim to second place and was the second seed in the gauntlet, only having to win one series before having the opportunity to face ahq e-Sports Club. Flash Wolves has a history of choking in gauntlet situations, but this time it was poised and prepared, taking a 3-0 set against Machi in the second match. The Wolves looked terrifying and for the first time in a long time, it looked ready to tear apart ahq.
Despite this, it was still coming in with a historical 5-10 record against the team. Considering ahq's dominant regular season, Flash Wolves was still very much seen as underdogs.
It didn't take long for it to prove that perception wrong. It dominated ahq in nearly every facet of the game. SwordArt and Karsa controlled the pace of the series, stifling any invasion into their jungle and rendering Xue "Mountain" Zhao-Hong ineffective in his aggressive invades.
Karsa flexed his ability to play carry junglers like no other in Taiwan, as he outclassed Mountain multiple times on Elise, Graves, and Nidalee. Maple proved to be amazing in this series as well, completely outclassing Chawy and Westdoor, putting together a KDA of 38 while dying only once in the series. Flash Wolves, as a team, stacked 50 kills to ahq's mere 16 in a dominating 3-0 route, with all wins being 35 minutes or less.
After a year of chasing ahq and being stuck in its shadow, Flash Wolves is finally the king of the LMS and is looking forward to MSI.
The wolves are coming into the Mid Season Invitational proving it's the best the region has to offer. The players are notably on hot streaks, with Maple having the best split of his career and with Karsa taking over the current jungle meta incredibly well, having always favored carry junglers over tanks, with Nidalee and Lee Sin being his most famous champions. SwordArt has even leveled up his game, roaming and creating picks on Alistar and his beloved Thresh.
The team has a decent profile for MSI, with the meta favoring a combination of tank top laners and carry junglers, with the main carry force coming from the mid lane. MMD has been known to be unstable, but in a more forgiving tank meta, it should be of little issue internationally. Maple has also had a fair amount of strong performances internationally and shown he can duke it out with some of the best.
The Flash Wolves is not only set up with the possibility to break out of groups, but potentially have its best chance ever at being a finalist at an international event. But Royal Never Give Up, G2 Esports, and Counter Logic Gaming won't be put away so easily so it's up to Flash Wolves to show if it can do the necessary damage.
|
india
Updated: Feb 12, 2019 17:00 IST
Alleging that he was stopped from boarding a plane from Lucknow to Allahabad, Samajwadi Party chief Akhilesh Yadav lashed out at the Yogi Adityanath government on Tuesday. He said this government believes in “roko, thoko neeti” (policy of stop and encounter).
Yadav claimed that he was prevented from participating in a function at the Allahabad University, where he was scheduled to attend an event of the students’ union. In December last year, Samajwadi Party backed candidate Uday Yadav had won the union poll.
Yadav said that he had informed the Yogi government in December about his visit to Allahabad University. “We sent the first programme on December 27…A detailed programme was sent on February 2 so that so that the administration could make necessary arrangements in the view of the Kumbh,” Yadav said at a press conference in Lucknow.
“But this government believes in ‘roko, thoko neeti’. It stopped me from boarding the plane. This government is scared of students,” he said adding, “I was also invited by a seer at Kumbh over lunch but this government led by a yogi has problem with that also.”
“I was prevented from boarding the airplane without any written orders. Currently detained at Lucknow airport,” the former chief minister tweeted earlier.
Read: #NeverAgain let the BJP’s two and a half men rule us: Akhilesh to voters
“It is clear how frightened the govt is by the oath ceremony of a student leader. The BJP knows that the youth of our great country will not tolerate this injustice anymore!” Akhilesh posted.
Coming out in support of Akhilesh Yadav, BSP chief Mayawati slammed the BJP government asking if it is “so scared” of the her party’s alliance with the SP. She said, “Is BJP government so scared of SP-BSP alliance that they have started banning political events? They seem to be using Kumbh for political agenda. Not allowing Akhilesh Yadav to go to Prayagraj proves that.”
The SP and the BSP announced their alliance for the Lok Sabha elections in January. The two parties had joined hands last year to hand over an embarrassing defeat to the BJP in Lok Sabha bypolls after the party formed its government with huge majority in the state.
The Samajwadi Party leader also posted a series of photographs from the Lucknow airport showing an official standing at the stairs of the aircraft apparently blocking his path. Akhilesh Yadav was seen in another photo surrounded by a sea of people, including policemen.
Officials from the police and administration were already present when Akhilesh reached the airport to board the plane for Prayagraj. The moment Akhilesh stepped out the car, the officials made a ring around him.
“Do not touch me. Keep your hands off me,” he said.
Akhilesh said the officials did not have any written orders and failed to explain the reason behind their action.
I was stopped at the airport without any written order. The officials could not clear the issue despite being asked. The only motive behind stopping me from going to the students union programme was to suppress the socialist thoughts amongst the youth,” Akhilesh tweeted in Hindi.
Read: Alliance, ambition and the allure of realpolitik
Yogi Adityanath said that Akhilesh’s visit would have sparked a row in Prayagraj.
“Samajwadi Party should do away with its negative activities. At Prayagraj, Kumbh Mela is going on and Akhilesh’s visit could have created chaos there. Therefore, his visit was cancelled. The party is known for creating the ruckus,” he told newspersons in Lucknow.
Yogi Adityanath and the Bharatiya Janta Party slammed Akhilesh’s allegation that his government had stopped him from attending a swearing-in ceremony of the students union at Allahabad University. A leader of the Samajwadi Party Chatra Sabha, the student’s wing of the Samajwadi Party, holds the post of president of the university’s students’ union.
“Barely 10 days back Akhilesh Yadav had visited Kumbh and taken bath at Sangam. He has been detained to ensure that the situation remains peaceful as Kumbh is on in Prayagraj,” Adityanath said in Lucknow soon Akhilesh accused his government of vendetta.
“The Samajwadi Party should stop indulging in negative activities,” Adityanath said.
BJP leader Chandramohan said the party cadre had checked and exposed what he called “Yadav’s lie.” He said, “It is clear that the Samajwadi Party is frustrated. The BJP is going to resist such tactics.”
The Samajwadi Party also released the videos of Akhilesh Yadav’s detention at the airport. The Samajwadi Party workers, who were waiting for him at the Prayagraj airport, created a ruckus.
Samajwadi Party legislators also tried to raise the issue in the Uttar Pradesh assembly during Question Hour.
Anticipating law and order problem at the Allahabad University due to animosity between rival student groups, the university administration has not permitted any political programme to be held on campus and intimated them accordingly.
Accordingly, the administration in Prayagraj has informed the chief minister’s office and Lucknow’s district magistrate, said principal secretary (home) Arvind Kumar.
|
All relevant data are within the paper.
Introduction {#sec001}
============
Schistosomiasis is considered a neglected parasitic disease and affects over 200 million people worldwide.This disease is caused by trematodes of the genus *Schistosoma* (Weiland, 1858) \[[@pone.0195519.ref001]\] and by the species *Schistosoma mansoni* in Brazil (Sambon 1907). In Brazil, *Biomphalaria glabrata* (Say, 1818) is one of the most important intermediate host species of *S*. *mansoni* due to its wide geographic distribution and high infection rates. The occurrence and spread of schistosomiasis are related to poor environmental conditions, with a high prevalence of human cases usually found in people live in unfavorable socioeconomic conditions \[[@pone.0195519.ref002],[@pone.0195519.ref003],[@pone.0195519.ref004]\].
Human and mollusk infections occur in freshwater bodies contaminated with the feces of people infected with schistosomes. Absent or inadequate sanitation, human cultural habits, and the presence of the *Biomphalaria* mollusk contribute to the persistence of the parasite\'s life cycle and consequently to the disease\'s geographic spread \[[@pone.0195519.ref005],[@pone.0195519.ref006],[@pone.0195519.ref007]\]. In Brazil, *S*. *mansoni* infection is considered a public health problem in 19 states. As a result, Brazil is considered the Latin American country with the greatest transmission foci, with approximately 30 million people at risk of contracting the disease and approximately 4--6 million people already infected \[[@pone.0195519.ref008],[@pone.0195519.ref009],[@pone.0195519.ref010]\].
Initially, schistosomiasis was typically endemic in rural areas; however, due to human migration, transmission also occurs in large Brazilian urban centers, with the northeast of the country having the highest prevalence rates \[[@pone.0195519.ref009], [@pone.0195519.ref011]\]. For instance, the state of Sergipe had a prevalence of 13.6% (14,471 positive cases/106,287 people analyzed) in 2005, 11.2% (16,196/145,069) in 2006, 11.8% (10,220/86,824) in 2007, and 10.6% (8,329/78,859) in 2008 based on parasitological examinations performed by the Brazilian Schistosomiasis Combat Program (Programa de Combate a Esquistossomose--PCE). In each year, 100% of the municipalities analyzed tested positive for *S*. *mansoni* infections \[[@pone.0195519.ref012]\]. Moreover, the tropical climate contributes to the establishment of schistosomiasis in this region because it provides conditions suitable for the presence of the mollusks that encourage the transmission of this endemic disease when associated with poor sanitary conditions \[[@pone.0195519.ref013]\].
Neglected tropical diseases, such as schistosomiasis, require prevention and control measures and elimination efforts; therefore, an understanding of the transmission factors in a particular area should be obtained \[[@pone.0195519.ref014],[@pone.0195519.ref015]\]. These factors include the characteristics of the environment and the sociocultural processes that influence the transmission process. Moreover, implementation of health education interventions, preventive chemotherapy directed at the entire community, and access to sanitation and safe water are essential to interrupt the transmission and control cycle of schistosomiasis \[[@pone.0195519.ref016]\]. Furthermore, identifying the breeding grounds of the *S*. *mansoni* mollusk intermediate host and performing an evaluation of its ecological and biological aspects are critical for investigations of the transmission risk rates and the implementation of control strategies in locations where the disease is established \[[@pone.0195519.ref017],[@pone.0195519.ref018]\].
The aim of this study was to evaluate potential schistosomiasis foci based on studies on the prevalence of *S*. *mansoni* infection in mollusks and individuals who lived near these foci, the influence of rainfall and environmental contamination due to the lack of a proper sanitation system, and sociocultural aspects of this endemic disease's transmission process in an urban community in northeastern Brazil.
Materials and methods {#sec002}
=====================
Study site and population {#sec003}
-------------------------
The study was conducted in a community in the city of Nossa Senhora do Socorro, which is located in the state of Sergipe. The city has approximately 155,334 inhabitants and a land area of 158 km^2^ in northeastern Brazil. The community named Parque dos Farois has a population of approximately 15,000 according to local health service information. The study area has a biome characterized by Atlantic forest and a hot and humid tropical climate and is located on the shores of the sub-basin of the Poxim River ([Fig 1](#pone.0195519.g001){ref-type="fig"}). The main environmental issues affecting the study area are riparian deforestation, trash in the open, disordered urban occupation, and deficiencies in the water supply and sanitary network \[[@pone.0195519.ref019]\].
{#pone.0195519.g001}
Ethical aspects {#sec004}
---------------
The study received institutional approval from the Ethics Committee on Human Research from Tiradentes University under protocol log number 28689514.4.0000.5371. All participants were invited to sign the free and informed consent term. In this document, participants must have the freedom to leave the study at any time. All questions are clarified to ensure that there is no doubt about your participation without study. In cases of illiteracy, a thumbprint and a signature of a witness were collected. The parents or guardians signed the consent form authorizing the participation of minors under 18 years of age. By signing the form, the participants agreed to participate in the interview through questionnaires and to provide a fecal sample for the parasitological analysis. Participants under 18 years of age were assisted by parents or guardians during the interview. The results were communicated to the participants, and adults and children under 18 years of age infected with any type of helminth were referred to the local health unit for treatment with 50 mg/kg body weight of praziquantel (Farmanguinhos® 600 mg; manufactured by the Institute of Technology in Pharmaceuticals/Farmanguinhos with registration by the Oswaldo Cruz Foundation). The treatment accordance with according to the Brazilian guidelines of the Control Program of Schistosomiasis (PCE).
Rainfall data {#sec005}
-------------
The data related to rainfall were obtained from the Department of Water Resources of the State of Sergipe (Secretaria de Recursos Hídricos do Estado de Sergipe---SRH-SE) for monthly correlations with the prevalence and abundance of the collected mollusks.
Mollusk collection and cercarial emission {#sec006}
-----------------------------------------
Mollusks were collected at five georeferenced points identified as P1 (10°55′230′′S 37°8′161′′W), P2 (10°55′231′′S 37°8′135′′W), P3 (10°55′326′′S 37°8′00′′W), P4 (10°55′18.37′′S 37°7′56.98′′W), and P5 (10°55′16.04′′S 37°7′51.80′′W), which were distributed in streams located near residences. The points were selected based on work that developed for the identification of mollusk transmitters of schistosomiasis by health workers operating in the Schistosomiasis Control Program (PCE) operated by the state government.
Mollusks were collected with the aid of sieves and tongs. Each person conducted collections for a 1-hour period monthly between August 2013 and August 2014. *S*. *mansoni* infection was evaluated by individual exposure of the mollusks to artificial light at a distance of 30 cm for 4 hours. This procedure provided a temperature between 28 and 30°C, which was capable of stimulating cercarial emission \[[@pone.0195519.ref020], [@pone.0195519.ref021]\].
Questionnaire survey {#sec007}
--------------------
To evaluate human infection with *S*. *mansoni*, a non-probabilistic, cross-sectional study was performed for accessibility in a 0.64-km^2^ area where mollusk collections were conducted concomitantly during the study period. The residents of all households located approximately 4.5 km from this area were visited and invited to participate in the study regardless of their ages gender, and whether or not they had a diagnosis of *S*. *mansoni* infection.
A questionnaire was used to interview the participants face to face. The questionnaire consisted of multiple choice questions or binary answers, such as affirmation and negation questions (yes or no), and was used to obtain information about individual patterns divided into the following categories: family identification data (address, age, gender, place of birth, and residence time in the area), economic data (family income), and environmental/cultural aspects (water supply, garbage disposal, sewage, and forms of contact with the river/stream water, among other information) ([S1 Appendix](#pone.0195519.s001){ref-type="supplementary-material"}).
Human parasitological diagnosis {#sec008}
-------------------------------
Following the principles of non-probabilistic sampling for convenience, all individuals who participated in the interviews received a plastic container for the collection of feces. The containers were collected, identified, and transferred to the laboratory for the parasitological analysis. One sample was collected from each individual in the morning in 2014. Two slides were prepared using the qualitative spontaneous sedimentation technique and the quantitative Kato-Katz technique. The Kato-Katz technique was also used to determine the parasitic load of the individuals for *S*. *mansoni*. The parasite load was defined by the number of eggs found on the slides and multiplied by the constant 24 and thus the number of eggs per gram of faeces (opg) was obtained, and was categorized as low light (1--99 epg), moderate (100--399 epg), or heavy (≥400 epg) \[[@pone.0195519.ref016], [@pone.0195519.ref022]\]. Individuals who had *S*. *mansoni* eggs on at least one of the two analyzed slides were considered positive. Other parasites found in the analyses were also included in this study.
Microbiological analysis {#sec009}
------------------------
To determine the presence of total and thermotolerant coliforms, 100 ml of water was collected from points P1, P2, P3, P4, and P5 in both the dry and rainy seasons of 2014. The multiple-tube quantitative method was applied for the microbiological evaluation to determine the "most probable number" (NMP) of target microorganisms in the sample through the distribution of aliquots in a series of tubes containing differential culture media for the growth of the target microorganisms \[[@pone.0195519.ref023]\]. The sodium lauryl sulfate, Brilliant Green, and EC broths were used for the analysis. In the sodium lauryl sulfate broth, the presumed presence of coliforms was evidenced by the formation of gas in the Durham tubes, which was produced by the fermentation of lactose present in the environment. The Brilliant Green broth was used for confirmative evidence of total coliforms in the confirmed positive tubes inoculated in the sodium lauryl sulfate broth \[[@pone.0195519.ref024]\].
The total and thermotolerant coliform levels were analyzed and compared to the limits set by resolution 357 of the National Environmental Council (Conselho Nacional do Meio Ambiente---CONAMA) from March 2005, which provided the designation \"Class 3 fresh water\" to designate water suitable for human consumption.
Statistical analysis {#sec010}
--------------------
The software STATISTICA 7.1 was used to perform the Spearman test correlation (rs) was used between two quantitative variables. This test was applied between the average of the monthly rainfall and the abundance and infection of mollusks. O software BioEstat 5.3 was used to perform the for the analysis of the independent variables and their associations with people with or without schistosomiasis, odds ratios were calculated (OR) with 95% confidence intervals (CIs) to verify the influence on the occurrence of a certain event between the parasitized and non-parasitized groups and Student's paired t-test was used to compare the microbiological parameters of the NMP to the values established by the CONAMA resolution. A 5% level of significance was adopted for all analyses \[[@pone.0195519.ref025]\].
Results {#sec011}
=======
Malacological study {#sec012}
-------------------
[Fig 2](#pone.0195519.g002){ref-type="fig"} shows the distribution of the collected mollusks. A total of 10,270 *B*. *glabrata* mollusks were collected during the study period. There was a negative correlation \[(rs) = -0.03, n = 10.270, p\> 0.05\] between a monthly average rainfall and an abundance of *B*. *glabrata* occurring in the previous season (one month before). Mollusks were collected 2,487 (IC 95%: 0.024--0.038) in the year 2013 and in the year 2014 7,883 (95% CI: 0,101--0,115).
{#pone.0195519.g002}
After exposure to the light stimulus to release the cercariae, 8.8% (n = 912) (95% CI: 0.083--0.094) of the mollusks were found to be positive for *S*. *mansoni*. However, the prevalence varied between 0 and 34.5% during the study period. No positive mollusks were found between November 2013 and March 2014 (with the exception of December); however, starting in April, mollusks positive for *S*. *mansoni* infection were found coinciding with the rainy season ([Fig 3](#pone.0195519.g003){ref-type="fig"}). A correlation was observed between rainfall and *B*. *glabrata* infection during the study period \[(rs) = 0.042, n = 10.270 p\<0.05\].
{#pone.0195519.g003}
Human infection {#sec013}
---------------
Of the 232 stool samples collected from individuals living close to the mollusk collection points, 66.4% had at least one intestinal parasite. *Schistosoma mansoni* infection had the second highest prevalence at 16.4% when the spontaneous sedimentation and Kato-Katz techniques were combined ([Table 1](#pone.0195519.t001){ref-type="table"}). The *S*. *mansoni* parasitic load of the infected population was 54.9, with a range from 24 to 144 eggs per gram (epg) of feces.
10.1371/journal.pone.0195519.t001
###### Prevalence of *S*. *mansoni* and intestinal parasitic infections determined by the spontaneous sedimentation and Kato-Katz methods.
{#pone.0195519.t001g}
Infeccion by Number (%)
------------------------------- ------------ ----------- ----------- ----------- ----------- ------------
***S*. *mansoni*** 31 (13.4) 7 (3.1) 22 (24.7) 16 (24.6) 38 (16.4) 0.12--0.22
***A*. *lumbricoides*** 30 (12.9) 34 (14.6) 40 (44.9) 24 (36.9) 64 (27.6) 0.22--0.24
***T*. *trichiura*** 10 (4.3) 11 (4.7) 13 (14.6) 8 (13.3) 21 (9.1) 0.06--0.13
**Hookworm** 10 (4.3) 0 (0) 7 (7.8) 3 (4.6) 10 (4.3) 0.02--0.08
***E*. *vermicularis*** 2 (0.8) 4 (1.7) 3 (3.4) 3 (4.6) 6 (2.6) 0.01--0.06
***H*. *nana*** 2 (0.8) 0 (0) 0 (0) 2 (3.1) 2 (0.8) 0.00--0.03
***E*. *coli*** 10 (4.3) \- 3 (3.4) 7 (10.7) 10 (4.3) 0.02--0.08
***E*. *histolytica/dispar*** 1 (0.4) \- 1 (1.2) 0 (0) 1 (0.4) 0.00--0.02
***G*. *lamblia*** 2 (0.8) \- 0 (0) 2 (3.1) 2 (0.8) 0.00--0.03
\*Combined result indicates any people that was positive either by Spontaneus Sedimentation or Kato-Katz.
\#Confidence Intervals 95.0% (C.I.95%).
Individuals infected with other helminths were encountered, particularly *Ascaris lumbricoides* with a prevalence of 27.6%. The commensal *Entamoeba coli* was among the most prevalent protozoans ([Table 1](#pone.0195519.t001){ref-type="table"}).
Regarding the age group, the majority of individuals infected with *S*. *mansoni* were children and adolescents in the 8- to 17-year-old age group and were represented by both genders (18.4% males and 23.7% females) (95% CI: 0.279--0.578). No differences were observed in the prevalence rates between men (10.5%) and women (7.9%) in the 29- to 49-year-old age group (95% CI: 0.191--0.475). Only males were infected in the age group of 18--28 years (18.4%) (95% CI: 0.092--0.334). However, only males were infected (7.9%) in the 50 years and above age group (95% CI: 0.027--0.208). Generally, *S*. *mansoni* infection was more prevalent in males, with the exception of the 40- to 49-year-old age group (95% CI: 0.058--0.273) ([Fig 4](#pone.0195519.g004){ref-type="fig"}).
{#pone.0195519.g004}
Questionnaire survey {#sec014}
--------------------
### Demographic and socioeconomic variables {#sec015}
Regarding provenance, 65.8% of the residents infected with *S*. *mansoni* came from the state of Sergipe, 13.1% from Alagoas, and 7.9% from Pernambuco. In [Table 2](#pone.0195519.t002){ref-type="table"} it is possible to check the residence time of infected residents, and 50% reported living in place of 2 to 10 years and 36.8% live 11 to 20 years. The level of education of individuals infected with *S*. *mansoni* stood basically on incomplete primary education, with 78.9% of the population that low education level. The socioeconomic approach showed that 89.5% of the population had a low wage band to a minimum income.
10.1371/journal.pone.0195519.t002
###### Residence time of individuals infected with Schistosoma mansoni in a community of Nossa Senhora do Socorro municipality, Sergipe, Brazil---2014.
{#pone.0195519.t002g}
Residence time Individuals
------------------- ------------- ------
**2 a 10 years** 19 50
**11 a 20 years** 14 36.8
**\> 20 years** 5 13.2
**Total** 38 100
\*N = Number
### Risk factors for *S*. *mansoni* infection {#sec016}
The odds ratio analysis for each of the variables tested showed that male residents were 3.86 times more likely to be infected with *S*. *mansoni*. The remaining analyzed variables showed no significant risk of infection ([Table 3](#pone.0195519.t003){ref-type="table"}). However, the area has poor sanitation, with residential waste dumped directly onto the streets through plumbing pipes improvised by the residents themselves, leading to the presence of open sewers with streams and puddles formed by both rainwater and residential sewage. These streams and puddles contained mollusks infected with *S*. *mansoni*.
10.1371/journal.pone.0195519.t003
###### Association of environmental and demographic factors with *Schistosoma mansoni* infection in residents of the Nossa Senhora do Socorro municipality, 2014.
{#pone.0195519.t003g}
Risk Factors *S*. *mansoni* OR[\*](#t003fn001){ref-type="table-fn"} 95.0% C.I.[^\#^](#t003fn002){ref-type="table-fn"} for OR P-Value
----------------------------- ---------------- ----------------------------------------- ---------------------------------------------------------- --------- ------------ ----------
**Potable water** Yes 10 81 0.84 0.35--2.00 0.873
No 14 96
**Garbage collection** Yes 22 153 0.62 0.16--2.36 0.7534
No 3 13
**Sewerage** Yes 13 92 0.93 0.40--2.15 0.9633
No 12 79
**Rudimentary cesspool** Yes 15 122 0.72 0.27--1.88 0.6816
No 7 41
**Contact (water streams)** Yes 13 79 1.22 0.51--2.89 0.8048
No 11 82
**Contact (river water)** Yes 15 93 1.19 0.50--2.81 0.8500
No 10 74
**Habit (Barefoot)** Yes 14 86 0.96 0.41--2.25 0.8975
No 11 65
**Sex** M 29 131 3.86 1.76--8.44 \<0.0007
F 9 157
\*OR = Odds Ratio
\#Confidence Intervals 95.0% (C.I.95%)
### Microbiological water analysis {#sec017}
The microbiological analysis results revealed different levels of contamination by total and thermotolerant coliforms in the samples from the five mollusk collection points. During the dry season, the total coliforms ranged from 2.1 x 10^4^ to 7.5 x 10^4^, and the fecal coliforms/thermotolerant coliforms ranged from 1.4 x 10^3^ and 4.6 x 10^3^. During the rainy season, the total coliforms ranged from 1.7 x 10^4^ to 4.7 x 10^4^, and the fecal coliforms/thermotolerant coliforms ranged from 1.4 x 10^3^ to 5.4 x 10^3^. During both the dry and rainy seasons, the thermotolerant coliform concentration was in accordance with the CONAMA resolution and showed no significant differences when the total mean was compared with the limits set by CONAMA. Conversely, the total coliform levels were above the allowed limit, and the paired t-test revealed total mean values significantly above those set by CONAMA in both evaluation periods: dry season---p \< 0.0322, mean 3.54, and standard deviation ± 2.24; rainy season---p \< 0.0094, mean 3.12, and standard deviation ± 1.24.
Discussion {#sec018}
==========
Epidemiological studies of schistosomiasis are important for elucidating the dynamics of transmission in endemic areas because each site has biological, ecological, social, and economic characteristics that may affect this process \[[@pone.0195519.ref026]\]. The urban area studied presents a social and environmental scenario that is conducive to the formation/maintenance of potential schistosomiasis transmission foci. The presence of mollusks in the water bodies, the individual cultural habits of each resident, and the precarious nature of the basic sanitation are essential for understanding the results obtained in this study.
In the scenario of schistosomiasis transmission in Brazil, *B*. *glabrata* mollusks are important actors because of their wide geographic distribution and the high rates of S. mansoni infection reported by other authors \[[@pone.0195519.ref011],[@pone.0195519.ref012]\]. In the study area, the parasitological analysis revealed the presence of *B*. *glabrata* mollusks releasing *S*. *mansoni* cercariae, (8.8%), which can be considered as a sign of the presence of transmission focus (\[[@pone.0195519.ref027]\]) in view of the fact that infected residents were also found (16.4%).
On the other hand, the Schistosoma transmission has an important seasonal component that is related to the distribution and density of the intermediate host and determined in the rainy periods by the increase of the rivers flow and the temporary water pools \[[@pone.0195519.ref028]\]. Although this study revealed no significant association between the mollusk density and rainfall, the latter factor appeared to influence these parameters; indeed, the density of the mollusks present and the mollusks shedding eliminating *S*. *mansoni* cercariae increased as the rainfall time increased. This correlation has been reported in other studies conducted in Brazil, which is the main endemic area for mansonic schistosomiasis in the Americas because the hot and humid tropical climate is conducive to the reproduction and maintenance of the host mollusks \[[@pone.0195519.ref008],[@pone.0195519.ref029],[@pone.0195519.ref030]\]. Before the rainy season, an important measure for the control of schistosomiasis is to maintain the cleanliness of the urban environment, as this can promote a significant reduction in the number of mollusks and their potential dispersion during the rainy season.
The identification of schistosomiasis transmission foci is crucial for the development of control and health management plans in these communities \[[@pone.0195519.ref031]\]. Based on the location of mollusks infected with *S*. *mansoni*, we performed a parasitic evaluation of the community close to the mollusk collection sites. The results showed prevalence greater than 15%; thus, the area can be considered highly prevalent even though the intensity of the infections was low, with less than 100 eggs per gram of stool \[[@pone.0195519.ref016]\]. Similar to reports in other studies \[[@pone.0195519.ref032]\], this fall in the infection rate is the result of control measures based only on treatment without the implementation of any other major interventions in the health infrastructure \[[@pone.0195519.ref033]\].
A cross-sectional study conducted in the same area but based on data from the Schistosomiasis Control Program (Programa de Controle da Esquistossomose---PCE) from 2003--2008 found that the municipality of Nossa Senhora do Socorro was considered to have medium endemicity, with a mean prevalence of 13.98% \[[@pone.0195519.ref031]\]. These results show that the epidemiological status of schistosomiasis has undergone no major changes in improvements in the population over the last 10 years despite the efforts of government control programs. Other soil-borne parasitic infections were also identified in the residents of the study area, showing that the population was exposed to various infectious agents probably due to the poor sanitary conditions in the area.
The highest prevalence encountered in this study occurred in children and adolescents, which corroborated the results from other studies that identified these age groups as the most affected due to the lack of hygiene, more frequent contact of young people with contaminated water, and their recreational activities. The greatest number of positive cases in Brazil has been found in the 6- to 20-year-old age group \[[@pone.0195519.ref006], [@pone.0195519.ref033], [@pone.0195519.ref034]\].
Together with the state of Sergipe, other states of the Brazilian Northeast are considered endemic for schistosomiasis, which reinforces the attention to the spread of the disease in the region and also to other regions of Brazil and the world as a result of migratory processes \[[@pone.0195519.ref006],[@pone.0195519.ref009],[@pone.0195519.ref032],[@pone.0195519.ref033]\]. Many of the individuals infected with *S*. *mansoni* in this study were born in other Northeastern states such as Alagoas and Pernambuco which may also contribute to the spread and establishment of endemic disease in the study area. Notably, all positive individuals had a low education level and family income, reinforcing the fact that schistosomiasis is strongly linked to a low socioeconomic status.
Several studies have reported that schistosomiasis is a disease that affects males more frequently \[[@pone.0195519.ref035], [@pone.0195519.ref036]\], which was also observed in our study, which showed that these have a significantly higher chance of acquiring the infection. This result may be related to the greater involvement of men in activities in rivers and streams near the community, such as fishing, river bathing for leisure and other activities. Similar results were found in the literature for studies conducted in the northeastern region of Brazil that showed higher prevalence rates for *S*. *mansoni* in males; the studies also suggested that the reason for these higher prevalence rates was possibly because males attended health services less frequently \[[@pone.0195519.ref037], [@pone.0195519.ref038], [@pone.0195519.ref039]\].
Although there was no significant association between schistosomiasis and demographic, social, environmental and cultural variables, certainly constant contact in areas with poor sanitation and in which infected intermediary hosts were detected may have contributed to the transmission of *S*. *mansoni* in the locality. Additionally, contact with contaminated water can also be accidental in the rainy season, when floods and overflowing streams facilitate human contact with the parasite \[[@pone.0195519.ref040], [@pone.0195519.ref041]\]. The same variables, sewage destination and water supply were analyzed in a study carried out in the State of Alagoas/ razil, in which no significant associations with *S*. *mansoni* infection were found \[[@pone.0195519.ref006]\].
Microbiological water quality evaluation is an essential tool for the prevention of waterborne infections. In addition to indicating the fecal contamination rate, this analysis is relevant because it focuses on confirming the presence of human waste in the water from which *B*. *glabrata* was collected. This finding is of interest because mollusk and human infections are strongly linked to contamination of the aquatic environment by coliforms \[[@pone.0195519.ref042]\]. The results obtained in this research revealed that the freshwater bodies were contaminated with total and thermotolerant coliforms, which certainly confirmed sewage discharge in that area and increased the likelihood that the *S*. *mansoni* eggs came from infected individuals in the community. Coliform bacteria have been extensively used as fecal pollution indicators in the evaluation of the sanitary conditions of water bodies because these microorganisms are specific to the human intestinal tract \[[@pone.0195519.ref032]\].
This study reveals worrying data regarding schistosomiasis in the community, including high prevalence of infection, presence of *S*. *mansoni* infected mollusks in streams located in front of residents\' homes. These findings point to the need to implement comprehensive control measures aimed at controlling the population of mollusks because residents share the urban space with the residents, associated with improved sanitation conditions so that the sewage is not dumped directly into the environment.
The situation of the urban community under study shows a new epidemiological situation observed in many Brazilian cities, the disorderly growth of cities often caused by rural migration, which gives rise to pockets of poverty in which residents live without adequate sanitation conditions being exposed to diverse diseases such as schistosomiasis. The Government\'s program of disease control in Brazil has worked and undoubtedly achieved good overall results, however this new urban transmission profile must be investigated and recognized in its particularities since each place has different socioenvironmental characteristics.
In conclusion, the prevalence of schistosomiasis in the studied community is related to the aforementioned conditions that contribute to the maintenance of the endemic cycle. These results indicate that understanding the peculiarities of endemic areas is important for defining specific prevention and control strategies. Government intervention is necessary in the actions of improvements in sanitation and the treatment of infected residents with an aim of disrupting the biological cycle of *S*. *mansoni* to attempt to minimize the risks of transmission and the emergence of new cases.
On the other hand, the recognition by the population of the need for health care is fundamental to reverse situations such as the one observed in the present research and that certainly go through the implementation of educational actions. Individuals infected with *S*. *mansoni* and/or other parasites were referred to the local health clinic for treatment, which does not guarantee that they will be reinfected in the same place. Finally, the data presented in this research show the reality of several urban communities concentrated in the periphery of many Brazilian cities and which clearly show that schistosomiasis in addition to other health situations could be minimized in a relevant way if, for example, adequate basic sanitation was offered to these populations.
Supporting information {#sec019}
======================
###### Questionnaire used in interviews with residents to obtain the results of socioeconomic and demographic variables.
In addition to risk factors for *S*. *mansoni* infection.
(PDF)
######
Click here for additional data file.
We thank the Fundação de Apoio a Pesquisa e Inovação Tecnológica---Fapitec for the grant award.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Current address: Federal Institute of São Paulo, Avaré, São Paulo, Brazil.
|
Diffusion tensor characteristics of gyrencephaly using high resolution diffusion MRI in vivo at 7T.
Gyrification of the human cerebral cortex allows for the surface expansion that accommodates many more cortical neurons in comparison to other mammals. For neuroimaging, however, it forms a feature that complicates analysis. For example, it has long been established that cortical layers do not occupy the same depth in gyri and sulci. Recently, in vivo diffusion imaging has provided insights into the fibre architecture of the cortex, usually showing radial tensor orientations. This makes it relevant to investigate whether cortical diffusion tensor metrics depend on the gyral pattern. High-resolution (1mm isotropic) diffusion weighted MRI of the medial wall of the hemispheres was performed at 7 T. Diffusion data were resampled to surfaces in the cortex and underlying white matter, where the cortical surfaces obeyed the equivolume principle for cortical laminae over the cortical curvature. Diffusion tensor metrics were averaged over bins of curvature to obtain maps of characteristic patterns in the gyrus. Diffusivity, anisotropy and radiality varied with curvature. Radiality was maximal in intermediate layers of the cortex next to the crown of the gyrus, not in white matter or on the crown. In the fundus, the deep cortical layers had tangential tensor orientations. In the white matter, tensor orientation changed from radial on the crown to tangential under the banks and fundus. White matter anisotropy gradually increased from the crown to the fundus. The characteristic pattern in the gyrus demonstrated here is in accordance with ex vivo diffusion MR microscopy and histological studies. The results indicate the necessity of taking into account the gyral pattern when cortical diffusion data is analysed. Additionally, the data suggest a confound for tractography approaches when reaching the gyrus, resulting in a possible bias towards the gyral crown. The implications for mechanisms that could drive cortical folding are discussed.
|
---
bibliography:
- 'bibliography.bib'
---
[**On Robustness Analysis of Stochastic Biochemical Systems by Probabilistic Model Checking** ]{}\
Luboš Brim, Milan Česka, Sven Dražan, David Šafránek$^{\ast}$\
Systems Biology Laboratory, Faculty of Informatics, Masaryk University, Brno, Czech Republic\
$\ast$ E-mail: [email protected]
Abstract {#abstract .unnumbered}
========
This report proposes a novel framework for a rigorous robustness analysis of stochastic biochemical systems. The technique is based on probabilistic model checking. We adapt the general definition of robustness introduced by Kitano to the class of stochastic systems modelled as continuous time Markov Chains in order to extensively analyse and compare robustness of biological models with uncertain parameters. The framework utilises novel computational methods that enable to effectively evaluate the robustness of models with respect to quantitative temporal properties and parameters such as reaction rate constants and initial conditions.
The framework is applied to gene regulation as an example of a central biological mechanism where intrinsic and extrinsic stochasticity plays crucial role due to low numbers of DNA and RNA molecules. Using our methods we have obtained a comprehensive and precise analysis of stochastic dynamics under parameter uncertainty. Furthermore, we apply our framework to compare several variants of two-component signalling networks from the perspective of robustness with respect to intrinsic noise caused by low populations of signalling components. We succeeded to extend previous studies performed on deterministic models (ODE) and show that stochasticity may significantly affect obtained predictions. Our case studies demonstrate that the framework can provide deeper insight into the role of key parameters in maintaining the system functionality and thus it significantly contributes to formal methods in computational systems biology.
Introduction
============
Robustness is one of the fundamental features of biological systems. According to Kitano [@Kitano2004] *“robustness is a property that allows a system to maintain its functions against internal and external perturbations”*. To formally analyse robustness, we must thus precisely identify model of a biological system and define formally the notions of a system’s function and its perturbations. In this paper, we propose a novel framework for robustness analysis of stochastic biochemical systems. To this end, inspected systems are described by means of stochastic biochemical kinetics models, system functionality is defined by its logical properties, and system perturbation is modelled as a change in stochastic kinetic parameters or initial conditions of the model.
Processes occurring inside living cells exhibit dynamic behaviours that can be observed and classified as carrying out a certain function – maintaining stable concentrations, responding to a change of the environment, growing etc. Kinetic models with parameters are used to formally capture cell dynamics. To observe and analyse a dynamic behaviour on a kinetic model, all its numerical parameters must be instantiated to a specific value. This poses a challenge since the precise values of all parameters (kinetic constants, initial concentrations, environmental conditions etc.) may not be known, may be known but without a given accuracy of measurement or may in principle form an interval instead of being a single value (e.g. non-homogeneous cell populations, different structural conformations of a molecule leading to multiple kinetic rates etc.). This implies that the behaviour of a kinetic model for a given single parametric instantiation and its derived functionality may not provide an adequate result and it is therefore unavoidable to take into account possible uncertainties, variance and inhomogeneities.
The concept of robustness addresses this aspect of functional evaluation by considering a weighted average of all behaviour across a *space of perturbations* each altering the model parameters (hence its behaviour) in a particular way and having a certain probability of occurrence. A general definition of robustness was introduced by Kitano [@Kitano2007]: $$R^{\mathcal{S}}_{\mathcal{A},\mathbf{P}} = \int_{\mathbf{P}}{\psi(p)D^{\mathcal{S}}_{\mathcal{A}}(p)dp}$$ where $\mathcal{S}$ is the system, $\mathcal{A}$ is the function under scrutiny, $\mathbf{P}$ is the space of all perturbations, $\psi(p)$ is the probability of the perturbation $p\in \mathbf{P}$ and $D^{\mathcal{S}}_{\mathcal{A}}(p)$ is an *evaluation function* stating how much the function $\mathcal{A}$ is preserved under a perturbation *p* in the system $\mathcal{S}$.
For the macroscopic view as provided by the deterministic modelling framework based on ordinary differential equations (ODEs), the concept of robustness has been widely studied. There exist many mature analytic techniques based on static analysis as well as dynamic numerical methods for effective robustness analysis of ODE models. In circumstances of low molecular/cellular numbers such as in signalling [@ueda2007stochastic], immunity reactions or gene regulation [@Gillespieetal05], intrinsic and extrinsic noise plays an important role and thus these processes are more faithfully modelled stochastically. However, the existing methods and tools are not adequate for rigorous and effective analysis of stochastic models with uncertain parameters. In order to bridge this gap we adapt the concept of robustness to stochastic systems.
The main challenge of the adaptation lies in the interpretation of the evaluation function $D^{\mathcal{S}}_{\mathcal{A}}(p)$. We discuss several definitions of the evaluation function that give us different options how to quantify the ability of the system to preserve the inspected functionality under a parameters perturbation. We show how absolute and relative robustness of the stochastic systems can be captured and analysed using our framework.
Semantics of stochastic biochemical kinetics models can be defined by *Continuous Time Markov Chains* (CTMCs) where the evolution of the probability density vector describing the population of particular species is given by the chemical master equation (CME) [@Gillespie1977]. A function of a system in the biological sense is any intuitively understandable behaviour (e.g., stability of *ERK* signal effector population in high concentration observed in a given time horizon). In order to define the robustness of a system formally we need to make precise the intuitive and informal concept of functionality. Our framework builds on the formal methods where the functionality of a system is expressed indirectly by its logical properties. This leads to a more abstract approach emphasising the most relevant aspects of a system function and suppressing less important technicalities. We use stochastic temporal logics, namely the bounded time fragment of *Continuous Stochastic Logic (CSL)* [@Aziz1996csl] further extended with *rewards* [@Kwiatkowska2006rewards] (e.g., $\mathsf{P}_{\geq 0.9}[ \mathsf{G}^{[t_1,t_2]} (ERK > \textit{high})]$). To broaden the scope of possibly captured functionality we extend CSL with a class of *post-processing functions* defined over probability density vectors. We show that the bounded fragment of CSL with rewards and post-processing functions can adequately capture many biological relevant behaviours that are recognisable in finite time intervals.
Our framework is based on probabilistic model checking techniques that compute the probability with which a given CTMC satisfies a given CSL formula. The computation can be conducted using Monte Carlo based methods such as Gillespie’s stochastic stimulation algorithm [@Gillespie1977] or numerical methods such as uniformisation [@Stewart2009uniformization]. Although Monte Carlo based methods (often denoted as statistical model checking) can produce detailed simulations for stochastically evolving biochemical systems, computing a statistical description of their dynamics that is necessary for evaluating $D^{\mathcal{S}}_{\mathcal{A}}(p)$, such as the probability density, mean, or variance, requires a large number of individual simulations. Moreover, if $\mathcal{A}$ describes a behaviour that occurs rarely, the evaluation of $D^{\mathcal{S}}_{\mathcal{A}}(p)$ requires an extremely large number of simulations to be performed to obtain sufficient accuracy. As was shown in [@munsky2006finite] in such situations numerical methods are substantially more efficient. Since rigorous stochastic robustness analyses may require to compute precise probabilities of all behaviours, we build our framework on probabilistic numerical methods.
To analyse the robustness of the CTMC $\mathcal{C}$ with respect to the CSL formula $\Phi$ over the space of perturbations $\mathbf{P}$, which can be discrete but still very large or continuous and thus infinite, one needs to efficiently compute or approximate the evaluation function $D^{\mathcal{C}}_{\Phi}$, i.e, the values $D^{\mathcal{C}}_{\Phi}(p)$ for all $p\in \mathbf{P}$. One of the possible approaches (recently used in [@Bartocci2013]) is to effectively sample the perturbation space $\mathbf{P}$ and use standard statistical or numerical methods to obtain values in grid points. These values can be afterwards interpolated linearly or polynomially. Using adaptive grid refinement such an approach provides an arbitrary degree of precision. A disadvantage of this method is the fact that the obtained result is an approximation not providing any minimal and maximal upper bounds. Therefore such an approach can neglect sharp changes or discontinuities in the landscape of the evaluation function $D^{\mathcal{C}}_{\Phi}$.
It is worth noting that the evaluation function can be discontinuous or may change its value rapidly on a very small perturbation interval in situations when the given CSL formula contains nested probability operators. In particular, this is inevitable to formulate hypothesis requiring the detailed temporal program [@Alon2004] of the biological system (e.g., temporal ordering of events). The actual shape of the evaluation function arises from the combination of such a formula and the particular model. Especially, high sensitivity of a model to the perturbed parameter can intensify rapid changes in the evaluation function. An example of a formula with a nested probability operator is mentioned in Section \[sec:functionality\].
To evaluate the function $D^{\mathcal{C}}_{\Phi}$ we employ in our framework another approach that is based upon our min-max approximation method recently published in [@CAV2013]. The method guarantees strict upper and lower estimates of $D^{\mathcal{C}}_{\Phi}(p)$ without neglecting any sharp changes or discontinuities. This method exploits numerical techniques for probabilistic model checking, can provide arbitrary degree of precision and thus can be considered as an orthogonal approach to the adaptive grid refinement. The framework further extends the min-max approximation to a more general class of stochastic biochemical models (i.e., incorporation of stochastic Hill kinetics) and a more general class of quantitative properties (i.e., including post-processing functions) and allows us to compute the robustness of such systems. In our framework we provide the user not only a numerical value giving the robustness of the system but possibly also a landscape visualisation of the evaluation function.
We demonstrate the applicability of the proposed method by means of two biological case studies – a model predicting dynamics of a gene regulatory circuit controlling the $G_1/S$ phase transition in the cell cycle of mammalian cells, and two models representing different topologies of a general two-component signalling mechanism present in procaryotic cells. Both cases are examples of cellular processes where stochasticity plays a crucial role especially because of low numbers of molecules involved.
The former case study exploits the usability of the method to analyse bistability (and its robustness) in the stochastic framework and thus provides a stochastic analysis analogy to the study presented in [@Swatetal04] under the deterministic (ODE) setting. Robustness is employed to characterise parameterisations of the model with respect to the tendency of the molecule population to choose one of the possible steady states irreversibly deciding whether the cell will or will not commit to *S*-phase. The results show that intrinsic and extrinsic noise caused by randomness in protein-DNA binding/unbinding events and other processes controlling the chemical affinity of involved molecules can significantly affect the cell decision. In our model, the intrinsic noise of chemical reactions is inherently captured by stochastic mass action kinetics whereas the extrinsic noise is considered by means of parameter uncertainty.
The latter case study focuses on analysing the effect of intrinsic noise on the signalling pathway functionality. In particular, two topologically different variants of a two-component signalling pathway are exploited for different levels of input signal and different levels of intrinsic noise appearing in transcription of the two signalling components. The considered topologies have been compared in the previous study presented by Steuer et al. [@steuer2011robust] where robustness has been analysed in the setting of deterministic (ODE) models. Here the signalling mechanism is remodelled in the stochastic setting and robustness is employed to quantify under which circumstances the individual topologies are less amenable to intrinsic noise of the underlying protein transcription mechanism. The results show that the stochastic approach can uncover facts unpredictable in the deterministic setting.
Formal analysis of complex stochastic biological systems employing both the numerical and the statistical methods generally suffers from extremely high computational demands. These computational demands are even more critical if we need to analyse systems with uncertain parameters which is also the case of our framework. However, our framework has been designed in order to be adapted to high performance computing platforms (e.g. multi-core workstations and massively parallel general-purpose graphic processing units) and also to be successfully combined with existing acceleration methods, see e.g. [@munsky2006finite; @Henzinger2009; @Didier]. Although the acceleration is a subject of our future research (inspired by our previous results [@BBS10]), we already employ the fact that the min-max approximation method can be efficiently parallelised. In the second case study where the analysis of the inspected perturbation space requires an extensive numerical computation, we utilise a high performance multi-core workstation to achieve the acceleration. Fundamentally different approaches to overcome the time complexity of stochastic analyses of complex biological systems build on a moment closure computation and on a fluid approximation, see e.g. [@Verena2013; @Bortolussi2012]. These approaches are briefly discussed in the related work.
The main contributions of this paper can be summarised in the following way:
1. The adaptation of the general concept of robustness of Kitano [@Kitano2007] to the class of stochastic systems modelled by CTMCs. The key step of the adaptation is a definition of the evaluation function that reflects the quantitative aspects of stochastic models and their behaviours. We discuss several definitions of the function allowing for different ways of capturing stochastic robustness.
2. Introduction of a novel framework based on formal methods to evaluate robustness of the stochastic system with respect to the functionality given by a stochastic temporal property and to *perturbations in reaction rate parameters and initial conditions*. The framework significantly extends the min-max approximation method published in [@CAV2013], namely with the support for Hill kinetics and post-processing functions.
3. Demonstration of the fact that our concept of robustness can capture and quantify the ability of the stochastic systems to maintain their functionality. We apply our framework to two biologically relevant case studies. Namely, it is the gene regulation of mammalian cell cycle where we explore the impact of stochasticity in low molecule numbers to bistability of a regulatory circuit controlling G1/S transition and analysis of noise behaviour in different topologies of two-component signalling systems. The case studies show that our framework provides deeper understanding of how the validity of an inspected hypotheses depends on reaction rate parameters and initial conditions.
Related work
------------
The discussion on related work can be roughly divided into two parts. First, we summarise the existing methods for parameter exploration and robustness analysis of stochastic models. Second, we briefly mention the methods and tools allowing for robustness analysis of ODE models.
In the field of stochastic models, parameter estimation methods and the concept of robustness are not as established yet as in the case of ODE models. We have recently published a method [@CAV2013] where the CSL model checking techniques are extended in order to systematically explore the parameters of stochastic biochemical kinetic models. In [@Verena2012] a CTMC is explored with respect to a property formalised as a deterministic timed automaton (DTA). It extends [@Verena2011] to parameter estimation with respect to the acceptance of the DTA. Most approaches to parameter estimation [@Reinker; @Verena2011; @Petzold2012] rely on approximating the maximum likelihood. Their advantage is the possibility to analyse infinite state spaces [@Verena2011] (employing dynamic state space truncation with numerically computed likelihood) or even models with no prior knowledge of parameter ranges [@Petzold2012] (using Monte-Carlo optimisation for computing the likelihood). In [@Verena2013] the moment closure approach is considered to capture the distribution of highly populated species in combination with discrete stochastic description for low populated species. The method is able to cope with multi-modal distributions appearing in multi-stable systems. The method introduced in [@Bortolussi2012] exploits fluid (limit) approximation techniques and in that way enables an alternative approach to CSL model checking of stochastic models. Despite the computational efficiency, a shared disadvantage of all the mentioned methods is that they rely on approximations applicable only to models that include highly populated species. This is not the case of, e.g., gene regulation dynamics.
Approaches based on Markov Chain Monte-Carlo sampling and Bayesian inference [@Wilkinson; @Clarke; @Cago] can be extended to sample-based approximation of the evaluation function, but at the price of undesired inaccuracy and high computational demands [@Bernardini; @Paolo]. Compared to these methods, our method provides the upper and lower bounds of the result which makes it more reliable and precise but at the price of higher computational demands. The most relevant contribution to this domain has been recently introduced by Bartocci et al. [@Bartocci2013]. To our best knowledge this is the only related work addressing robustness of stochastic biochemical systems. The work is based on the idea to directly adapt the concept of behaviour oriented robustness to stochastic models. Individual simulated trajectories of the CTMC are locally analysed with respect to a formula of Signal Temporal Logic (STL), a linear-time temporal logic interpreted on simulated time sequences. For each simulated trajectory, the so-called satisfaction degree representing the distance from being (un)satisfied is computed, thus resulting into a randomly sampled distribution of the satisfaction degree. This distribution thus gives modellers another source of information in addition to probability of formula satisfaction (percentage of valid trajectories in the sampled set). In comparison, our method directly (and exactly) computes the probability of formula satisfaction for a different kind of temporal logic – the branching-time CSL logic. This allows to express more intricate properties that require branching time, e.g., multi-stability. On the other hand, our method as conceptually based on transient analysis does not allow to compute local analysis of individual trajectories, i.e., to obtain a satisfaction degree would require non-trivial elaboration at the level of numerical algorithms.
In the domain of ODE models, there exist several analytic methods for effective analysis under parameter uncertainty. They build on static analysis (stoichiometric analysis, flux balance analysis) as well as dynamic numerical methods (simulation, monitoring by temporal formulae, sensitivity analysis) implemented in tools (e.g. [@Hoops2006copasi; @Loew2001vcell; @Fages2004biocham]). Robustness analysis with respect to functionality specified in terms of temporal formulae has been introduced recently [@Fainekos2009robustness; @Rizketal09]. There exist two major approaches how to define and analyse robustness. If only parameters of the model are perturbed, we speak of a *behaviour oriented approach* to robustness. This approach has been explored by Fainekos & Pappas [@Fainekos2009robustness], further extended by A. Donzé et al. [@Donze2010robust] and implemented in the toolbox Breach [@Donze2010breach]. Another option could be to perturb the model structure i.e. the reaction topology, as this is done in many gene knock-out biological experiments. Such changes are in principle discrete and the problem of robustness computation for such perturbations would reduce to solving many individual instances of the same problem for each discrete topology. However identifying model behaviour shared among individual perturbations can lead to more efficient analysis [@Barnat12coloredmc].
Yet another way to look at perturbations is from the perspective of property uncertainty. If the system is considered fixed and all parameters exactly known, the uncertainty then lies in the property of interest. For a specific property such as “The concentration of X repeatedly rises above 10 and drops below 5 within the first 20 minutes” where all three numerical constants can be altered, we explore how much would they have to be altered in order to affect the property validity in the given model. This approach has been adopted for ODEs by F. Fages et al. [@Rizketal09] and implemented in the tool BIOCHAM [@Fages2004biocham]. When only parameters of the property are perturbed, it is the case of a *property oriented approach* to robustness.
Methods {#sec:methods}
=======
Methodology Overview
--------------------
In this paper we propose a formal framework that allows to analyse the robustness of stochastic biochemical systems with respect to a space of perturbing parameters. The framework consists of the following objects:
- *a finite state stochastic biochemical system given by a set of chemical species participating in a set of chemical reactions*
Each of the reaction is associated with a stochastic rate function that for a fixed stochastic rate constant returns the rate of the reaction. To formalise such system we use a population based finite state continuous time Markov Chain (CTMC), i.e, a state of the CTMC is given by populations of particular species and the evolution of the CTMC is driven by the chemical master equation (CME) [@Gillespie1977; @Didier].
- *a perturbation space defined by a Cartesian product of uncertain stochastic rate constants given as value intervals with minimal and maximal bounds*
Additionally, the perturbation space may also be expanded by initial conditions of the system (i.e, interval for the size of a population of a particular species) encoded in the initial state of the CTMC. The given stochastic system and the perturbation space induce a *set of parameterised* CTMCs.
- *set of paths that describe the evolution of a fully instantiated stochastic system (i.e., in which all stochastic rate constants and the initial state are specified) over time*
For a state of the system and a finite time there is a unique probability measure of all paths starting in that state that defines *probability distribution* over states occupied by the system at the given time. Each perturbation from a given perturbation space possibly leads to a different probability distribution.
- *stochastic temporal property interpreted over the paths and states of CTMC enabling to specify an a priori given quantitative hypothesis about the system*
We primarily focus on the *bounded time fragment of Continuous Stochastic Logic* (CSL) [@Aziz1996csl] further extended with rewards [@Kwiatkowska2006rewards]. For most cases of biochemical stochastic systems the bounded time restriction is adequate since a typical behaviour is recognisable in finite time. Additionally, we also consider properties given by a class of post-processing functions defined over probability distributions at the given finite time.
The main goal of our framework is to analyse how the validity of an *a priori* given hypothesis expressed as a temporal property depends on uncertain parameters of the inspected stochastic system. For this purpose we adapt the general definition of robustness [@Kitano2007] to the class of stochastic systems. While the concept of robustness is well established for deterministic systems [@Donze11rob-behaviour; @Rizk09rob-property], it has not been adequately addressed for stochastic systems. The key difference is the fact that evolution of a stochastic system is given by a set of paths in contrast to a single trajectory as in the case of a deterministic system. Hence a stochastic system at the given time is described by a probability distribution over states of the corresponding CTMC in contrast to the single state representation of a deterministic system. Therefore, the definition of robustness for stochastic systems requires a more sophisticated interpretation of the evaluation function that determines how the quantitative temporal property is preserved under a perturbation of the system’s parameters.
Similarly to Kitano, we define robustness of stochastic systems as the integral of an evaluation function. In our case the evaluation function $D^{\mathcal{C}_p}_{\Phi}$ for each parameter point *p* from the inspected perturbation space $\mathbf{P}$ returns the quantitative model checking result for the respective CTMC $\mathcal{C}_p$ and the given property $\Phi$. We show how robustness can be effectively over/under-approximated for a class of quantitative temporal properties using new techniques for model checking of parameterised CTMCs. Moreover, if the property can be expressed using only the bounded time fragment of CSL with rewards (i.e., without post-processing functions) we can extend the approach to global quantitative model checking techniques. They enable to compute the model checking result for all states of a CTMC with the same price as for a single state and thus to analyse the perturbation of initial conditions in a much more effective way. Finally, we demonstrate how robustness can capture and quantify the ability of a stochastic system to maintain its functionality described by such class of properties.
Since the inspected perturbation space is in principle dense the set of parameterised CTMCs to be explored is infinite. It is thus not possible to compute the model checking result for each CTMC individually. The straightforward approach to overcome this problem could be to sample points from the perturbation space and use existing model checking techniques for fully instantiated CTMCs. That way we can obtain precise model checking results in the grid points and then interpolate them linearly or polynomially. Although an adaptive grid refinement could provide an arbitrary degree of precision, it does not guarantee strict lower and upper bounds. Hence such an approach could neglect sharp changes or discontinuities of the evaluation function. Since we want to guarantee strict bounds of obtained results, we extend our previously published method [@CAV2013]. This method allows to compute strict minimal and maximal bounds on the quantitative model checking results for all CTMCs $\left\{ \mathcal{C}_p \mid p \in \mathbf{P} \right\}$ for a given perturbation space $\mathbf{P}$.
Models
------
The formalism used to model a biochemical system is essential since it not only dictates the possible behaviours that may or may not be captured, but also determines the means of detecting them. ODEs enable the study of large ensembles of molecules in population count and species diversity since they abstract from the individualistic properties of each molecule such as position or its stochastic behaviour and take as its variables only concentrations of each species. Stochastic models such as CTMCs abstract positions of molecules but maintain their individual reactions. Even more detailed models such as Brownian dynamics which keep track of positions but abstract from the geometry and orientation of each molecule could be used. However as the amount of information about each individual molecule increases the computational complexity of proving some property to hold over all the behaviours of a model becomes quickly infeasible even for small models.
In our framework we focus on stochastic biochemical systems that can be formalised as a finite state system $\mathcal{S}$ defined by a set of *N* *chemical species* in a well stirred volume with fixed size and fixed temperature participating in *M* *chemical reactions*. The number $X_i$ of molecules of each species $S_i$ has a specific bound and each reaction is of the form $u_1 S_1 + \ldots + u_N S_N \longrightarrow v_1 S_1 + \ldots + v_N S_N$ where $u_i, v_i \in \mathbb{N}_0$ represent *stoichiometric coefficients*.
A *state* of a system in time $t$ is the vector $\mathbf{X}(t) = (X_1(t),
X_2(t), \ldots, X_N(t) )$. When a single reaction with index $r \in \{1, \ldots,
M\}$ with vectors of stoichiometric coefficients $U_r$ and $V_r$ occurs the state changes from $\mathbf{X}$ to $\mathbf{X}' = \mathbf{X} - U_r + V_r$, which we denote as $\mathbf{X} \stackrel{r}{\rightarrow} \mathbf{X}'$. For such reaction to happen in a state $\mathbf{X}$ all reactants have to be in sufficient numbers and the state $\mathbf{X}'$ must reflect all species bounds. The *reachable state space* of $\mathcal{S}$, denoted as $\mathbb{S}$, is the set of all states reachable by a finite sequence of reactions from *an initial state* $\mathbf{X}_0$. The set of indices of all reactions changing the state $\mathbf{X}_i$ to the state $\mathbf{X}_j$ is denoted as $\mathsf{reac}(\mathbf{X}_i,\mathbf{X}_j) =
\{r \mid \mathbf{X}_i \stackrel{r}{\longrightarrow} \mathbf{X}_j \}
$. Henceforward the reactions will be referred directly by their indices.
According to [@Gillespie1977; @Didier] the behaviour of a stochastic system $\mathcal{S}$ can be described by the CTMC $\mathcal{C} = (\mathbb{S},
\mathbf{X}_0, \mathbf{R})$ where the transition matrix $\mathbf{R}(\mathbf{X}_i,\mathbf{X}_j)$ gives the probability of a transition from $\mathbf{X}_i$ to $\mathbf{X}_j$. Formally, the transition matrix is defined as: $$\mathbf{R}(\mathbf{X}_i,\mathbf{X}_j) \stackrel{def}{=} \sum_{r \in \mathsf{reac}(\mathbf{X}_i,\mathbf{X}_j) } f_r(\mathbf{k}_r, \mathbf{X}_i)$$ where $f_r$ is a *stochastic rate function* and $\mathbf{k}_r$ is a vector of all numerical parameters occurring in $f_r$ such as a *stochastic rate constant* $k_r$, stoichiometry exponents, Hill coefficients etc.
In case of mass action kinetics the stochastic rate function has the simple form of a polynomial of reacting species populations. That is $ f_r(\mathbf{k}_r,
\mathbf{X}_i) = k_r \cdot C_{r,i}$ where $C_{r,i} \stackrel{def}{=} \prod_{l =
1}^{N}\binom{\mathbf{X}_{i,l}}{u_l}$ corresponds to the population dependent term such that $\mathbf{X}_{i,l}$ is the *l*th component of the state $\mathbf{X}_{i}$ and $u_l$ is the stoichiometric coefficient of the reactant $S_l$ in reaction $r$. However, sometimes the mass action kinetics is not sufficient, especially, when the reactions are not elementary but are rather an abstraction of several reactions with unknown precise dynamics (e.g. gene transcription) or if including all elementary reactions would cause the analysis to be computationally infeasible. In such cases dynamics are typically approximated by Hill functions [@Hill1910], a quasi-steady-state approximation [@Madsen2012] of the law of mass conservation. For sake of simplicity of our presentation we will further assume that for each reaction $r$ the vector $\mathbf{k}_r$ is one-dimensional and thus $\mathbf{k}_r = k_r$, the proposed methods can however be directly used also for multi-dimensional vectors of constants. To comply with standard notation in the area of CTMC analysis henceforward the states $\mathbf{X}_i \in \mathbb{S}$ will be denoted as $s_i$.
The probability of a transition from state $s_i$ to $s_j$ occurring within *t* time units is $1 - e^{-\mathbf{R}(s_i,s_j) \cdot t}$, if such a transition cannot occur then $\mathbf{R}(s_i,s_j) = 0$. The time before any transition from $s_i$ occurs is exponentially distributed with an overall *exit rate* $E(s_i)$ defined as $E(s_i) = \sum_{s_j \in
\mathbb{S}}{\mathbf{R}(s_i,s_j)}$. A path $\omega$ of CTMC $\mathcal{C}$ is a non-empty sequence $\omega=s_0,t_0,s_1,t_1\ldots$ where $\mathbf{R}(s_i,s_j)>0$ and $t_i\in \mathbb{R}_{\geq 0}$ is the amount of time spent in the state $s_i$ for all $i \geq 0$. For all $s\in \mathbb{S}$ we denote by $Path^{\mathcal{C}}(s)$ the set of all paths of $\mathcal{C}$ starting in state $s$. There exists the unique probability measure on $Path^{\mathcal{C}}(s)$ defined, e.g., in [@Kwiatkowska2007]. Intuitively, any subset of $Path^{\mathcal{C}}(s)$ has the unique probability that can be effectively computed. For the CTMC $\mathcal{C}$ the transient state distribution $\pi^{\mathcal{C},s,t}$ gives for all states $s'\in \mathbb{S}$ the transient probability $\pi^{\mathcal{C},s,t}(s')$ defined as the probability, having started in the state *s*, of being in state $s'$ at the finite time *t*.
Perturbations
-------------
In our approach we have focused on the behavioural approach for stochastic systems and thus we will now define a set of perturbed stochastic systems and their CTMCs. Let each stochastic rate constant $k_r$ have a value interval $[k_r^{\bot},k_r^{\top}]$ with minimal and maximal bounds expressing an *uncertainty range* or *variance* of its value. A *perturbation space* $\mathbf{P}$ induced by a set of stochastic rate constants $k_r$ is defined as the Cartesian product of the individual value intervals $\mathbf{P} =
\prod_{r=1}^{M}[k_r^{\bot},k_r^{\top}]$. A single *perturbation point* $p
\in \mathbf{P}$ is an *M*-tuple holding a single value of each rate constant, i.e., $p = (k_{1_p}, \ldots , k_{M_p})$.
A stochastic system $\mathcal{S}_p$ with its stochastic rate constants set to the point $p\in \mathbf{P}$ is represented by a CTMC $\mathcal{C}_p =
(\mathbb{S}, s_0, \mathbf{R}_p)$ where transition matrix $\mathbf{R}_p$ is defined as: $$\mathbf{R}_p(s_i,s_j) \stackrel{def}{=} \sum_{r \in \mathsf{reac}(s_i,s_j) } f_r(k_{r_p}, s_i)$$ A *set of parameterised* CTMCs induced by the perturbation space $\mathbf{P}$ is defined as $\mathbf{C} = \{ \mathcal{C}_p \mid p \in \mathbf{P} \}$.
Additionally, we consider the perturbation of initial conditions of the stochastic system that are represented by different initial states of the corresponding CTMC. In this case we extend the perturbation space such that a single perturbation point $p\in \mathbf{P}^e = \mathbb{I} \times \mathbf{P}$ where $\mathbb{I} \subseteq \mathbb{S}$ is an *M*+1-tuple holding a single value of an initial state and a single value of each rate constant, i.e., $p =
(s_p,k_{1_p}, \ldots , k_{M_p})$ and CTMC $\mathcal{C}_p =(\mathbb{S}, s_p,
\mathbf{R}_p)$.
Functionality {#sec:functionality}
-------------
To be able to automatically analyse a system’s function $\mathcal{A}$ under scrutiny there must be a formal way of expressing a function of a system. A function of a system in the biological sense is any intuitively understandable behaviour such as response, homoeostasis, reproduction, respiration or growth. It can be a high level concept such as chemotaxis as well as a low level one e.g. reaching of a state with a given number of molecules of a specific species.
The inspected function can usually be described by a *property* that is understood as an abstraction of a system’s behaviour expressed in some temporal logic and given as a formula of that logic. Unlike the intuitive concept of a biological function mentioned above, a property may be formally verified over a formal model of a system and proven to hold or to be violated. Since the concept of robustness builds on the notion of a function that can be measured, we focus on a quantitative logic for stochastic systems. We use *continuous stochastic logic* (CSL) [@Aziz1996csl; @Baier2003mcctmc] extended with *reward* operators [@Kwiatkowska2006rewards]. Reward operators allow us to further broaden the scope of possibly captured behaviour. They enable to express properties such as the probability of a system being in the specified set of states over a time interval or the probability that a particular reaction has occurred.
Full CSL with rewards can express properties concerning a system in near future as well as the infinite steady state situation. In this paper we focus only on the *bounded time fragment of CSL*. This fragment allows us to speak only about behaviour within a finite time horizon. For most cases of biochemical stochastic systems, such as intracellular reaction cascades or multi-cellular signalling, the bounded time restriction is adequate since a typical behaviour is recognisable within finite time intervals [@KwiatkowskaBIO].
As we show later on, there exist several biologically relevant properties that cannot be directly expressed by CSL with rewards. Therefore, we employ a class of post-processing functions to specify and analyse robustness of stochastic systems with respect to such properties. The key idea of these functions is to process and aggregate the transient state distribution at the given finite time.
Let $\mathcal{C} =\left(\mathbb{S}, s_0, \mathbf{R}, L \right)$ be a labelled CTMC such that *L* is a labelling function which assigns to each state $s \in
\mathbb{S}$ the set *L(s)* of atomic propositions that are valid in state $s$. We consider the specification of the inspected property using the bounded time fragment of CSL with rewards and post-processing functions. The syntax of this logic is defined in the following way. A state formula $\Phi$ is given as $$\Phi::= \mathsf{true} \mid a \mid \neg\Phi \mid \Phi \wedge \Phi \mid
\mathsf{P}_{\sim p}[\phi] \mid \mathsf{R}_{\sim r}[\mathsf{C}^{\leq t}] \mid
\mathsf{R}_{\sim r}[\mathsf{I}^{=t}] \mid \mathsf{E}_{\sim r}[\mathsf{I}^{=t}]$$ where $\phi$ is a path formula given as $\phi::= \mathsf{X}\mbox{~}\Phi \mid$ $\Phi\mbox{~}\mathsf{U}^{I}\mbox{~}\Phi$, *a* is an atomic proposition, $\sim \in \left\{ <, \leq, \geq, > \right\}$, $p \in [0,1]$ is a probability, $r
\in \mathbb{R}_{\geq 0}$ is an expected reward and $I = [a,b]$ is a bounded time interval such that $a,b \in \mathbb{R}_{\geq 0} \wedge a \leq b$. Path operators $\mathsf{G}$ (always) and $\mathsf{F}$ (eventually) are derived in the standard way using the operator $\mathsf{U}$. In order to specify properties containing rewards ($\mathsf{R}_{\sim r}[\mathsf{C}^{\leq t}]$ is the *cumulative reward* acquired up to time *t*, $\mathsf{R}_{\sim
r}[\mathsf{I}^{= t}]$ is the *instantaneous reward* in time *t*) the CTMC $\mathcal{C}$ is enhanced with reward (cost) structures. Two types of reward structures can be used, *a state reward* and *a transition reward*. For sake of simplicity, we consider in this paper only state rewards, however, the proposed methods can be easily extended to transition rewards as well. The state reward $\rho(s)$ defines the rate with which a reward is acquired in state $s\in \mathbb{S}$. A reward of $t \cdot \rho(s)$ is acquired if $\mathcal{C}$ remains in state *s* for *t* time units.
Since the function $\rho$ has to be defined before the actual analysis of the CTMC, the rewards for particular states have to be known prior to the specification of the property. This fact limits the class of properties that can be expressed using such structures. For example, noise expressed by a *mean quadratic deviation* (*mqd*) of the population probability distribution of a species at a given time cannot be specified using CSL with rewards. To compute the *mqd* we need to know the mean of the distribution to be able to obtain the corresponding coefficients and encode them into state rewards.
To overcome this problem we introduce the abstract state operator $\mathsf{E}_{\sim r}[\mathsf{I}^{=t}]$ which *evaluates* the state distribution $\pi^{\mathcal{C},s_0,t}$ at the given time instant *t* by a user provided real-valued *post-processing* function $Post(\pi^{\mathcal{C},s_0,t})$ and compares it to $\sim r$. At the end of this section we show how to define $Post$ in order to specify biologically relevant properties such as noise using the *mqd*. The *mqd* is also used in the second case study to analyse a noise in different variants of signalling pathways.
The formal semantics of the bounded fragment of CSL with rewards and post-processing functions is defined similarly as the semantics of full CSL and thus we refer the readers to original papers. The key part of the semantics is given by the definition of the satisfaction relation $\vDash$. It specifies when a state $s$ satisfies the state formula $\Phi$ (denoted as $s \vDash \Phi$) and when a path $\omega$ satisfies the path formula $\phi$ (denoted as $\omega
\vDash \phi$). The informal definition of $\vDash$ is as follows:
- $s \vDash \mathsf{E}_{\sim r}[\mathsf{I}^{=t}]$ iff $Post(\pi^{\mathcal{C},s,t})$ satisfies $\sim r$.
- $s \vDash \mathsf{P}_{\sim p}[\phi]$ iff the probability of all paths $\omega \in Path^{\mathcal{C}}(s) $ that satisfy the path formula $\phi$ (denoted as $Prob^{\mathcal{C}}(s,\phi)$) satisfies $\sim p$, where
- $\omega$ satisfies $\mathsf{X}\mbox{~}\Phi$ iff the second state on $\omega$ satisfies $\Phi$
- $\omega$ satisfies $\Phi\mbox{~}\mathsf{U}^{I}\mbox{~}\Psi$ iff there exists time instant $t\in I$ such that the state on $\omega$ occupied at $t$ satisfies $\Psi$ and all states on $\omega$ occupied before $t'\in [0,t)$ satisfy $\Phi$
- $s \vDash \mathsf{R}_{\sim r}[\mathsf{C}^{\leq t}]$ iff the sum of expected rewards over $Path^{\mathcal{C}}(s)$ *cumulated* until $t$ time units (denoted as $Exp^{\mathcal{C}}(s,\mathsf{X}_{\mathsf{C}^{\leq t}})$) satisfies $\sim r$
- $s \vDash \mathsf{R}_{\sim r}[\mathsf{I}^{= t}]$ iff the sum of expected rewards over all paths $\omega \in Path^{\mathcal{C}}(s)$ at time *t* (denoted as $Exp^{\mathcal{C}}(s,\mathsf{X}_ {\mathsf{I}^{= t}})$) satisfies $\sim\!r$.
A set $Sat_{\mathcal{C}}(\Phi)=\{s \in \mathbb{S} \mid s \vDash \Phi \}$ denotes the set of states that satisfy $\Phi$.
Note that the syntax and semantics can be easily extended with “quantitative” formulae in the form $\Phi ::= \mathsf{P}_{= ?}[\phi] \mid
\mathsf{R}_{= ?}[\mathsf{C}^{\leq t}] \mid \mathsf{R}_{= ?}[\mathsf{I}^{= t}]
\mid \mathsf{E}_{=?}[\mathsf{I}^{=t}]$, i.e., the topmost operator of the formula $\Phi$ returns a quantitative result, as used, e.g., in PRISM [@KNP11]. In this case the result of a decision procedure is not in the form of a boolean yes/no answer but the actual numerical value of the probability $Prob^{\mathcal{C}}(s,\phi)$, the expected reward $Exp^{\mathcal{C}}(s,\mathsf{X})$ for $\mathsf{X} \in \{ \mathsf{X}_
{\mathsf{I}^{= t}}, \mathsf{X}_ {\mathsf{C}^{\leq t}}\}$ or the value of $Post^{\mathcal{C}}(s,t)$. The computation of a numerical value is of the same complexity class as the computation of a result to be compared leading to a boolean answer, although in some cases the comparison may be carried out on less precise or preliminary results. As we will show the quantitative result is much more suitable for robustness analysis.
To demonstrate that the bounded time fragment of CSL with rewards and post-processing functions can adequately capture relevant biological behaviours and thus be successfully used in the robustness analysis of stochastic biochemical systems, we list several formalisations of such behaviours.
- *stochastic reachability* - $\mathsf{P}_{\geq 0.8}[ \mathsf{F}^{[5,10]} ( A \geq 3)]$ expresses the property “The probability that the population of *A* exceeds 3 between 5 and 10 time units is at least $80\%$”.
- *stochastic stability* - $\mathsf{P}_{=?}[ \mathsf{G}^{[0,5]} ( A \geq 1 \wedge A \leq 3 )]$ represents the quantitative property “What is the probability that the population of *A* remains between 1 and 3 during the first 5 time units?”
- *stochastic temporal ordering of events* - $\mathsf{P}_{<0.2}[(A \le 2)\ \mathsf{U}^{[2,3]}\ \mathsf{P}_{\geq 0.95} [ ( 2 < A
\leq 5)\ \mathsf{U}^{[0,10]} (A > 5)]]$ expresses the stochastic version of the following temporal pattern: “Species A is initially kept below 2 until it reaches 5 and finally exceeds 5.” The formula quantifies both the time constrains of the events and the probability that the events occur. It expresses that “The probability that the system has following probabilistic temporal pattern is less that $20\%$: the population of *A* is initially kept below 2 until the system between 2 and 3 times units reaches the states satisfying the subformula $\mathsf{P}_{\geq 0.95} [ ( 2 < A \leq 5)\ \mathsf{U}^{[0,10]} (A > 5)]]$." The subformula specifies the states where “The probability that the population of *A* remains greater than 2 and less or equal 5 until it exceeds 5 within 10 time units, is greater than $95\%$."
- *cumulative reward property* - $\mathsf{R}_{<5}[\mathsf{C}^{\leq 100}]$, where $\forall s\in \mathbb{S}\ \rho(s) = 1$ if $0 \leq A \leq 3$ in *s*, captures the property that “The overall time spent in states with population of *A* between 0 and 3 within the first 100 time units, is less than 5 time units”, which can also be understood as “The probability of the system being in a state with population of *A* between 0 and 3 within the first 100 time units is less then 5%”.
- *noise as mean quadratic deviation* - $\mathsf{E}_{<10}[\mathsf{I}^{=100}]$, where the post-processing function is defined as $Post(\pi) = \sum_{s \in \mathbb{S}}{\lvert s(A) - mean(\pi,A)\rvert^2 \cdot \pi(s)}$, $s(A)$ gives the population of *A* in state *s* and $mean(\pi,A)$ is the mean of the distribution $\pi$ defined as $mean(\pi,A) = \sum_{s \in \mathbb{S}}{s(A) \cdot \pi(s)}$. This property states that “The mean quadratic deviation of the distribution of species *A* at time instant $t=100$ must be less then 10”.
The $\mathsf{E}$ operator could in principle be extended to allow for intervals and be interpreted as an integral of a user-provided post-processing function over the given time interval. This could lead e.g. to the noise over time interval which is more natural then an instantaneous noise, however the computation complexity of such an operator would be very large.
![[**Running example.**]{} The example model contains one species *X* with the population bounded to 40, two reactions: production of *X* ($\emptyset \rightarrow X$ with rate $k_1$), degradation of *X* ($X \rightarrow \emptyset$ with rate $k_2 \cdot [X]$, $k_2 = 0.01$) and initial population of *X* is 15. The corresponding CTMC has 41 states (initial state $s_0$ corresponds to state with initial population). The inspected formula $\Phi$ represents the quantitative property “What is the probability that the population of $X$ is between 15 and 20 at time 1000?” The perturbation space $\mathbf{P}$ is given by the interval of the rate $k_1 \in [0.1,0.3]$. On the right, there are depicted three transient distributions at time 1000 for three different values of $k_1$ and the resulting probability for the formula $\Phi$ obtained as the sum of probabilities in states with populations from 15 to 20.[]{data-label="fig:decomposition"}](decomposition){width="\textwidth"}
Robustness
----------
Let us recap the general definition of Kitano [@Kitano2007] to show how it can be interpreted and how we propose to use it in the context of stochastic systems.
$$R^{\mathcal{S}}_{\mathcal{A},\mathbf{P}} = \int_{\mathbf{P}}{\psi(p)D^{\mathcal{S}}_{\mathcal{A}}(p)dp} \hspace{1cm} D^{\mathcal{S}}_{\mathcal{A}}(p) = \left\{
\begin{array}{cl}
0 & p \in \mathbf{B} \subset \mathbf{P} \\
f_{\mathcal{A}}(p)/f_{\mathcal{A}}(0) & p \in \mathbf{P} \setminus \mathbf{B}
\end{array} \right.$$
### Functionality evaluation
Kitano proposed that the evaluation function $D^{\mathcal{S}}_{\mathcal{A}}(p)$ stating how much the functionality $\mathcal{A}$ is preserved in perturbation *p* should be defined using a subspace $\mathbf{B}$ of all perturbations where the system’s function is completely missing and the rest $\mathbf{P} \setminus \mathbf{B}$ where the functions’ viability is somehow altered. This definition is meaningful e.g. in cases where the perturbation would lead to a system not having the function at all (speed of reproduction of a dead cell) or in cases where a plain measurement would provide a function’s value, however, in reality the system would lack the function altogether (inside temperature during homoeostasis experiment in conditions when an organism loses thermal control and has temperature of environment). These examples have in common that the information about a system lacking its function is provided from outside because if it could be deducible from the system’s state alone it could be incorporated into the evaluation function $D^{\mathcal{S}}_{\mathcal{A}}(p)$ itself.
For perturbations $p \in \mathbf{P} \setminus \mathbf{B}$ where the system maintains its function at least partially, Kitano proposes to express the evaluation function $D^{\mathcal{S}}_{\mathcal{A}}(p) = f_{\mathcal{A}}(p)/f_{\mathcal{A}}(0)$ relatively to the ground unperturbed state $f_{\mathcal{A}}(0)$. This is meaningful e.g. for naturally living systems where the ground state is measurable and is considered as an optimal performance state. Such a definition could then enable the comparison of a common property of different species. For example, a reproduction rate for a mouse and a sequoia tree with respect to perturbations of their environment. If a mouse has 20 offsprings per year in base temperature and 22 offsprings for a 2 degrees Kelvin rise then the evaluation function $D^{\mathcal{S}_M}_{\mathcal{A}}(+2 K) = 22/20 = 1.1$. While if a sequoia has 1000 seedlings in ground temperature and 1200 for a 2 degrees Kelvin rise then $D^{\mathcal{S}_S}_{\mathcal{A}}(+2 K) = 1200/1000 = 1.2$.
We can see that the relativistic nature of Kitano’s definition enables comparison of otherwise incomparable organisms and their robustness to perturbations. In our example, the sequoia is more robust to the single perturbation of temperature by $+2K$ than the considered species of mice. However, in cases when no ground state is given the absolute value can be more adequate. The next subsection shows that robustness in stochastic systems can be defined in several different ways providing both the absolute and relative interpretations.
### Robustness in Stochastic systems
Let $\mathcal{S}$ be a stochastic system with CTMC $\mathcal{C} = \left(\mathbb{S}, s_0, \mathbf{R}, L \right)$, let $\mathbf{P}$ be a space of perturbations to the stochastic kinetic constants of $\mathcal{C}$ and let $\Phi$ be a formula of the bounded time fragment of CSL with rewards and evaluation functions formalising the system’s function $\mathcal{A}$. Since the evaluation of $\Phi$ is inherently dependent on the initial conditions of the system that are encoded using the initial state $s_0$, we consider the evaluation function in the form $D^{\mathcal{C},s_0}_{\Phi}$.
In cases where the set of perturbed stochastic kinetic constants $\mathbf{P}$ is actually extended by initial conditions to $\mathbf{P}^e$, then for a single perturbation point $p = (s_p, k_{1_p}, \ldots , k_{M_p}) \in \mathbf{P}^e$ we consider the initial state $s_0$ of $\mathcal{C}$ to be substituted by $s_p$ in all subsequent expressions, otherwise it remains the original $s_0$.
Let us first define an auxiliary *Eval* function which is then used in the definition of $D^{\mathcal{C},s_0}_{\Phi}$: $$Eval_{\Phi}^{\mathcal{C}}( s_0) = \left\{
\begin{array}{cl}
Prob^{\mathcal{C}}(s_0,\phi) & \mbox{if~} \Phi \equiv \mathsf{P}_{\star}[\phi] \\[0.3em]
Exp^{\mathcal{C}}(s_0,\mathsf{X}_{\mathsf{C}^{\leq t}}) & \mbox{if~} \Phi \equiv \mathsf{R}_{\star}[\mathsf{C}^{\leq t}] \\[0.3em]
Exp^{\mathcal{C}}(s_0,\mathsf{X}_{\mathsf{I}^{= t}}) & \mbox{if~} \Phi \equiv \mathsf{R}_{\star}[\mathsf{I}^{= t}] \\[0.3em]
Post(\pi^{\mathcal{C},s_0,t}) & \mbox{if~} \Phi \equiv \mathsf{E}_{\star}[\mathsf{I}^{= t}]
\end{array} \right.
\label{eg:eval}$$ where $\star \in \{=?, \sim\! r\}$. Given these specifications the evaluation function $D^{\mathcal{C},s_0}_{\Phi}$ can be restated in several different ways:
$$\begin{aligned}
D^{\mathcal{C},s_0}_{\Phi}(p) &= \left\{
\begin{array}{clcl}
\hspace{1cm} 0 \hspace{1cm} & p \in \mathbf{B} \subset \mathbf{P} & \vee & Eval_{\Phi}^{\mathcal{C}_p}(s_0) \nsim r \\[0.5mm]
1 & p \in \mathbf{P} \setminus \mathbf{B} & \wedge & Eval_{\Phi}^{\mathcal{C}_p}(s_0) \sim r
\end{array} \right.
\label{eq:sem_boolean}\\
D^{\mathcal{C},s_0}_{\Phi}(p) &= \left\{
\begin{array}{cl}
\hspace{1cm} 0 \hspace{1cm} & p \in \mathbf{B} \subset \mathbf{P} \\[0.8mm]
\frac{Eval_{\Phi}^{\mathcal{C}_p}(s_0)}{r} & \mbox{else if~} \sim \in \{ \geq, > \} \\[0.8mm]
\frac{r}{ Eval_{\Phi}^{\mathcal{C}_p}(s_0)} & \mbox{else if~} \sim \in \{ \leq, < \}
\end{array} \right.
\label{eq:sem_relative}\\
D^{\mathcal{C},s_0}_{\Phi}(p) &= \left\{
\begin{array}{cl}
\hspace{1cm} 0 \hspace{1cm} & p \in \mathbf{B} \subset \mathbf{P} \\[0.5mm]
Eval_{\Phi}^{\mathcal{C}_p}(s_0) & \mbox{else}
\end{array} \right.
\label{eq:sem_absolute}\\
D^{\mathcal{C},s_0}_{\Phi}(p) &= \left\{
\begin{array}{cl}
\hspace{1cm} 0 \hspace{1cm} & p \in \mathbf{B} \subset \mathbf{P} \\[0.8mm]
\lvert Eval_{\Phi}^{\mathcal{C}_p}(s_0) - X \rvert^2 & \mbox{else, ~} X = agr \{Eval_{\Phi}^{\mathcal{C}_p}(s_0) \mid \mathcal{C}_p \in \mathbf{P} \} \wedge agr\in \{min,max,avg \}
\end{array} \right.
\label{eq:sem_relative2} \end{aligned}$$
The first definition of the evaluation function (\[eq:sem\_boolean\]) is possible for the specification where the topmost operator of the formula $\Phi$ includes the threshold $r$ (i.e. $\star = \sim \! r$). Because $D^{\mathcal{C},s_0}_{\Phi}(p)$ returns a qualitative result robustness $R^{\mathcal{C}}_{\Phi,\mathbf{P}}$ specifies the measure of all perturbations in $\mathbf{P}$ for which the property holds in a strictly boolean sense – it is the fraction of $\mathbf{P}$ where the property is valid. This definition can be used, e.g., in the property $\Phi_A = \mathsf{P}_{\geq 0.8}[ \mathsf{F}^{[0,5]} (X > 300)]$ which specifies that in $80\%$ of cases the population of X is larger than 300 within 5 seconds. For this property and a model with a parameter $k \in [0,10]$ the robustness gives us the fraction of the parametric interval $[0,10]$ for which the model satisfies $\Phi_A$.
In the second definition (\[eq:sem\_relative\]) $D^{\mathcal{C},s_0}_{\Phi}(p)$ returns the quantitative value that is relative to the threshold *r*. Therefore, robustness can be interpreted as the average relative validity of the property over $\mathbf{P}$. If *r* corresponds to the validity of $\Phi$ in conditions considered natural for the inspected system $\mathcal{S}$ (i.e, to the unperturbed state) then this interpretation complies with the original definition of Kitano. Let us consider the same property $\Phi_A$ and the same parametric space $k \in [0,10]$. If in $60\%$ of model behaviours the population of X is larger than 300 within 5 seconds than the robustness is 0.6/0.8 = 0.75. If the probability is different in each *k* then the robustness gives us the average value that meets our expectations.
The third definition (\[eq:sem\_absolute\]) is possible for specifications using the quantitative semantics of formula $\Phi$ (i.e. $\star = ?$). The robustness gives the mean validity over all $\mathbf{P}$ regardless of any probability threshold *r*. This interpretation is convenient when there are no *a priori* assumptions about the system expected behaviour.
Finally, to express the fact that the system behaviour remains the same (with respect to the evaluation function) across the space of perturbations we introduce the fourth definition (\[eq:sem\_relative2\]). It uses an aggregation function to compute a mean value and then express the variance from the mean. This definition enables us to compare models which have same numerical values of robustness in the sense of definition (\[eq:sem\_absolute\]) but which achieve the average value with very different landscapes of evaluation function.
While the last three definitions require the precise computation of the probability value in every $p \in \mathbf{P}$, the first definition is amenable to approximate solutions. In this case it suffices to ensure that the probability is larger or smaller then *r*. In many cases it can be achieved without computing the precise value and thus statistical model checking techniques can be efficiently used. In both case studies we use definition (\[eq:sem\_absolute\]), since we do not consider any ground unperturbed state. We assume $\mathbf{B}$ to be an empty set and expect all the lack of functionality $\mathcal{A}$ to be fully expressible in terms of the property $\Phi$.
Robustness computation
----------------------
Now we look how robustness $R^{\mathcal{C}}_{\Phi,\mathbf{P}^e}$ can be efficiently computed by using the evaluation function $D^{\mathcal{C},s_0}_{\Phi}$. Let us first consider the case where the space of perturbations $\mathbf{P}$ does not contain different initial states.
As will be shown in the next section the computation of $Eval^{\mathcal{C}_p}_{\Phi}(s_0)$ even for a single perturbation point $p$ is rather complex, therefore a computation of the integral over the whole space of perturbations is not possible in an explicit sense. Instead a way to approximate the upper and lower bounds $R_{\Phi,\mathbf{P},\top}^{\mathcal{C}}$ and $R_{\Phi,\mathbf{P},\bot}^{\mathcal{C}}$ is introduced enabling the approximation of the value of the integral as $$\begin{array}{rcl}
R_{\Phi,\mathbf{P}}^{\mathcal{C}} &\stackrel{def}{=}& \displaystyle\int_{\mathbf{P}}{\psi(p)D^{\mathcal{C}}_{\Phi}(p)dp} \\[0.6em]
R_{\Phi,\mathbf{P}}^{\mathcal{C}} &\simeq& \displaystyle\frac{1}{2}\left(
R_{\Phi,\mathbf{P},\top}^{\mathcal{C}} + R_{\Phi,\mathbf{P},\bot}^{\mathcal{C}} \right) \pm Err_{\Phi,\mathbf{P}}^{\mathcal{C}} \hspace{1cm}
Err_{\Phi,\mathbf{P}}^{\mathcal{C}} = \displaystyle\frac{1}{2}\left( R_{\Phi,\mathbf{P},\top}^{\mathcal{C}} - R_{\Phi,\mathbf{P},\bot}^{\mathcal{C}} \right)
\end{array}$$
The computation of $R_{\Phi,\mathbf{P},\top}^{\mathcal{C}}$ and $R_{\Phi,\mathbf{P},\bot}^{\mathcal{C}}$ is due to the approximation of the upper $D^{C}_{\Phi,\mathbf{P},\top}$ and lower $D^{\mathcal{C}}_{\Phi,\mathbf{P},\bot}$ bounds for values of the evaluation function $D^{\mathcal{C}}_{\Phi}(p)$ over $\mathbf{P}$ $$D^{\mathcal{C}}_{\Phi,\mathbf{P},\top} \geq max \left\{ D^{\mathcal{C}}_{\Phi}(p) \mid p \in \mathbf{P} \right\}\hspace{1cm} D^{\mathcal{C}}_{\Phi,\mathbf{P},\bot} \leq min \left\{ D^{\mathcal{C}}_{\Phi}(p) \mid p \in \mathbf{P} \right\}$$ Because such an approximation would be too course for most cases a finite decomposition of the perturbation space $\mathbf{P}$ into perturbation subspaces $\mathbf{P} = \mathbf{P}_1 \cup \ldots \cup \mathbf{P}_n$ is used which then under the assumption of equal probability of all perturbations gives better robustness bounds. Hence we get that: $$R_{\Phi,\mathbf{P},\top}^{\mathcal{C}} = \sum_{i = 1}^{n}{\frac{|\mathbf{P}_i|}{|\mathbf{P}|} \cdot D_{\Phi,\mathbf{P}_i,\top}^{\mathcal{C}}} \hspace{1cm}
R_{\Phi,\mathbf{P},\bot}^{\mathcal{C}} = \sum_{i = 1}^{n}{\frac{|\mathbf{P}_i|}{|\mathbf{P}|} \cdot D_{\Phi,\mathbf{P}_i,\bot}^{\mathcal{C}}}
\label{eq:robustness_sum}$$
Let us now consider the case in which the space of perturbations is extended with initial states $\mathbf{P}^e = \mathbb{I} \times \mathbf{P}$ where $\mathbb{I} \subseteq \mathbb{S}$ and $\mathbf{P}$ is non-singular, for this case the integral defining robustness is actually a finite sum of integrals: $$R_{\Phi,\mathbf{P}^e}^{\mathcal{C}} \stackrel{def}{=} \displaystyle\sum_{s \in \mathbb{I}}{\frac{1}{|\mathbb{I}|} \displaystyle\int_{p \in \mathbf{P}}{\psi(p)D^{\mathcal{C}}_{\Phi}(p)dp}} = \frac{1}{|\mathbb{I}|} \displaystyle\sum_{s \in \mathbb{I}}{ R_{\Phi,\mathbf{P}}^{\mathcal{C}} }$$ where $\psi(p)$ gives the probability of perturbation *p* with respect to $\mathbf{P}$. This expression is valid for uniform distributions of the initial states over the whole space of perturbations $\mathbf{P}^e$, however, it can be straightforwardly modified for non-uniform distributions. Using the expression the robustness computation for perturbations containing a single initial state can be easily extended to perturbations containing different initial states. Moreover, in Section \[sec:GMC\], we show that for most properties the model checking procedure (utilised in the robustness computation) returns results for an arbitrary set of initial states $\mathbb{I} \subseteq \mathbb{S}$ with the same time complexity as for a single state.
The accuracy of the approximation can be further improved using the *piece-wise linear approximation* of robustness. This concept is illustrated in Figure \[fig:pla\_approx\]. Since the spaces $\mathbf{P}_i$ and $\mathbf{P}_{i+1}$ have a common point *p* (in a general *n* dimensional perturbation space $2^n$ subspaces intersect in a single point *p*), we can use this to obtain a more precise range of values for the value of the property $\Phi$ in *p* as $$D_{\Phi,p,\top}^{\mathcal{C}} = min\left\{ D_{\Phi,\mathbf{P}_i,\top}^{\mathcal{C}} \mid p \in \mathbf{P}_i \right\} \text{ and } D_{\Phi,p,\bot}^{\mathcal{C}} = max\left\{ D_{\Phi,\mathbf{P}_i,\bot}^{\mathcal{C}} \mid p \in \mathbf{P}_i \right\}.$$
Under the assumption that the value of a property does not change rapidly over sufficiently small subspaces $\mathbf{P}_i$ the resulting upper and lower bound of robustness can then be computed from linear interpolation of grid points *p*. The decision in which cases such an assumption is acceptable is up to user since there is in general no efficient way of resolving this situation. In such a case the overall piecewise linear approximation of robustness will usually have a higher precision albeit without the guarantee of strict upper and lower bounds.
![[**Piecewise linear approximation of robustness.**]{} An improved approximation is shown in dark green, it is computed by linearly interpolating grid points in which the upper and lower bounds of a property may be computed more precisely as the minimum resp. maximum of the values from all parameter subintervals sharing boundary grid points. The obtained result is more precise that the original robustness (in light pink) albeit without the conservative guarantee on bounds.[]{data-label="fig:pla_approx"}](0017_pla_reduced){width="\textwidth"}
To understand how $D_{\Phi,\mathbf{P}_i,\top}^{\mathcal{C}}$ and $D_{\Phi,\mathbf{P}_i,\bot}^{\mathcal{C}}$ can be efficiently computed first the methods for transient analysis and global CSL model checking based on *uniformisation* are revisited [@Baier2000mcviatran; @Kwiatkowska2007]. Afterwards we present the *min-max approximation* [@CAV2013] that allows us to approximate the quantitative model checking result for continuous sets of parameterised CTMCs. The key idea is to employ a method called *parameterised uniformization* – a modification of the standard uniformization technique presented in [@CAV2013]. Finally, we show how to control the *approximation error* in order to obtain the required error bound.
### Transient analysis
The aim of transient analysis is to compute a transient probability distribution. Given an initial distribution $\pi^{\mathcal{C}, s_0, 0}$ (i.e. $\pi^{\mathcal{C}, s_0, 0}(s) = 1$ if $s_0 = s$, and 0, otherwise) at time 0 of a CTMC $\mathcal{C} = \left( \mathbb{S}, s_0, \mathbf{R}\right)$ what will the transient state distribution $\pi^{\mathcal{C},s_0,t}$ look like in some future yet finite time $t \in \mathbb{R}_{\geq 0}$.
Transient analysis of a CTMC may be efficiently carried out by a standard technique called *uniformization* [@Kwiatkowska2007]. The transient probability in time *t* is obtained as a sum of expressions giving the state distributions after *i* discrete reaction steps of the respective *uniformized* discrete time Markov chain (DTMC) weighted by the *i*th probability of the Poisson process. It is the probability of *i* such steps occurring in time *t*, assuming the delays between steps of the CTMC $\mathcal{C}$ are exponentially distributed with *rate* $q$. Formally, for the rate $q$ satisfying $q \geq max\{E^{\mathcal{C}}(s) \mid s\in \mathbb{S}\}$ (*E* is the exit rate of state *s*) the uniformized DTMC $\mathsf{unif}(\mathcal{C})$ is defined as $\mathsf{unif}(\mathcal{C}) = \left( \mathbb{S}, s_0, \mathbf{Q}^{\mathsf{unif}(\mathcal{C})} \right)$ where $$\mathbf{Q}^{\mathsf{unif}(\mathcal{C})}(s,s') = \left\{
\begin{array}{cl}
\frac{\mathbf{R}(s,s')}{q} & \mbox{if~~} s \neq s' \\
1-\sum_{s'' \neq s}{\frac{\mathbf{R}(s,s'')}{q}} & \mbox{otherwise.}
\end{array} \right.$$ and the $i$th Poisson probability in time $t$ is given as $\gamma_{i,q \cdot t}= e^{-q \cdot t} \cdot \frac{\left(q\cdot t\right)^i}{i!}$. The transient probability can be computed as follows: $$\pi^{\mathcal{C}, s_0, t} = \sum^{\infty}_{i=0}{\gamma_{i,q
\cdot t} \cdot \pi^{\mathcal{C},s_0,0} \cdot (
\mathbf{Q}^{\mathsf{unif}(\mathcal{C})})^i} \approx
\sum^{R_{\epsilon}}_{i=L_{\epsilon}}{\gamma_{i,q \cdot t} \cdot
\pi^{\mathcal{C},s_0,0} \cdot ( \mathbf{Q}^{\mathsf{unif}(\mathcal{C})})^i}.$$ Although the sum is in general infinite, for a given precision $\epsilon$ the lower and upper bounds $L_{\epsilon}, R_{\epsilon}$ can be estimated by using techniques such as of Fox and Glynn [@FoxGlynn1988] which also allow for efficient solutions of the Poisson process. In order to make the computation of uniformization feasible the matrix-matrix multiplication is reduced to a vector-matrix multiplication, i.e., $$\pi^{\mathcal{C}, s_0, 0} \cdot
(\mathbf{Q}^{\mathsf{unif}(\mathcal{C})})^i = ( \pi^{\mathcal{C}, s_0, 0} \cdot
(\mathbf{Q}^{\mathsf{unif}(\mathcal{C})} )^{i-1} ) \cdot
\mathbf{Q}^{\mathsf{unif}(\mathcal{C})}.$$
Standard uniformization can be intractable when the system under study is too complex, i.e., contains more than in order of $10^7$ states and the upper estimate $R_{\epsilon}$, denoting the number of vector-matrix multiplications as iterations, is high (more than in order of $10^6$). Therefore, many approximation techniques have been studied in order to reduce the state space and to lower the number of iterations $R_{\epsilon}$. State space reductions are based on the observation that in many cases (especially in biochemical systems) a significant amount of the probability mass in a given time is localized in a manageable set of states. Thus neglecting states with insignificant probability can dramatically reduce the state space while the resulting approximation of the transient probability is still sufficient. Methods allowing efficiently state-space reduction are based on finite projection techniques [@munsky2006finite; @Henzinger2009] and dynamic state space truncation [@Didier].
Since the number of iterations $R_{\epsilon}$ inherently depends on the uniformization rate *q* that has to be greater then the maximal exit rate of all the states of the system, a variant of standard uniformization, so-called *adaptive* uniformization [@Moorsel94], has been proposed. It uses a uniformization rate that adapts depending on the set of states the system can occupy at a given time, i.e, after a particular number of reactions. In many cases, a significantly smaller rate *q* can be used and thus the number of iterations $R_{\epsilon}$ can be significantly reduced during some parts of the computation. Moreover, adaptive uniformization can be successfully combined with reduction techniques mentioned above [@Didier]. The downside of adaptive uniformization is that the Poisson process has to be replaced with a general *birth* process which is more expensive to solve. See, e.g [@Moorsel94], for more details.
For sake of simplicity, we present our methods for the computation of $D_{\Phi,\mathbf{P}_i,\top}^{\mathcal{C}}$ and $D_{\Phi,\mathbf{P}_i,\bot}^{\mathcal{C}}$ using standard uniformization. However, our method can be successfully combined with the aforementioned techniques.
### Global CSL Model Checking {#sec:GMC}
The aim of the global model checking technique is to efficiently compute for any CSL formula $\Phi$ the values $Eval_{\Phi}^{\mathcal{C}}(s)$ for all states $s \in \mathbb{S}$. On the other hand, the goal of local model checking technique is to compute $Eval_{\Phi}^{\mathcal{C}}(s)$ for a single state $s \in \mathbb{S}$. The crucial advantage of the global approach is the fact that it has the same asymptotic and also practical complexity as the local approach. Therefore, the global model checking technique is much more suitable for robustness analysis over perturbations of initial conditions that are encoded as the initial state of the corresponding CTMC.
Global model checking returns the vector of size $|\mathbb{S}|$ such that the *i*th position contains the model checking result provided that $s_i$ is the initial state. Let $\mathcal{C} =\left(\mathbb{S}, \mathbf{R}, L \right)$ be a labelled CTMC where the initial state is not specified. The crucial part of this method is to compute the vector of probabilities $\overline{Prob}^{\mathcal{C}, \phi}$ for any path formula $\phi$ and the vector of expected rewards $\overline{Exp}^{\mathcal{C}, \mathsf{X}}$ for $\mathsf{X} \in \{ \mathsf{X}_ {\mathsf{I}^{= t}}, \mathsf{X}_ {\mathsf{C}^{\leq t}}\}$ such that for all $s\in \mathbb{S}$ the following holds: $$\overline{Prob}^{\mathcal{C}, \phi}(s) = Prob^{\mathcal{C}}(s,\phi) \wedge \overline{Exp}^{\mathcal{C}, \mathsf{X}}(s) = Exp^{\mathcal{C}}(s,\mathsf{X})$$
In local model checking the computation of $Prob^{\mathcal{C}}(s,\phi)$ and $Exp^{\mathcal{C}}(s,\mathsf{X})$ is reduced to the computation of the transient probability distribution $\pi^{\mathcal{C},s,t}$, see [@Baier2000mcviatran; @Kwiatkowska2007] for more details. Thus, for different initial states $s$ we have to compute the corresponding transient probability distributions separately. The key idea of the global model checking method is to use *backward transient analysis*. The result of backward transient analysis is the vector $\tau^{\mathcal{C},\mathbb{A},t}$ such that for arbitrary set of states $\mathbb{A}$, the value $\tau^{\mathcal{C},\mathbb{A},t}(s)$ is the probability that $\mathbb{A}$ is reached from $s$ at the time *t*. Without going into details the vector $\tau^{\mathcal{C},\mathbb{A},t}$ can be computed in a very similar way using the uniformized DTMC $\mathsf{unif}(\mathcal{C})$ as in the case of vector $\pi^{\mathcal{C},s,t}$. Only vector-matrix multiplications is replaced by matrix-transposed-vector multiplication and $\tau^{\mathcal{C},\mathbb{A},0}(s) = 1$ if $s\in\mathbb{A}$, and $0$, otherwise.
The global model checking technique can not be used if $\Phi$ includes the operator $\mathsf{E}_{\sim r}[I^{=t}]$. In such a case we have to compute the value $Post(\pi^{\mathcal{C},s,t})$. Hence the local model checking technique has to be employed, i.e., we first compute the vector $\pi^{\mathcal{C},s,t}$ and then apply the user specified function *Post*.
Now we briefly show how the vector $\overline{Prob}^{\mathcal{C}, \phi}$ is computed using backward transient analysis. Since the definition of next operator $\mathsf{X}\ \Phi$ does not rely on any real time aspects of CTMCs, its evaluation stems from the probability of the next reaction that can be easily obtained from the transition matrix $\mathbf{R}$. The evaluation of the until operator $\Phi_1 \mathsf{U}^{I} \Phi_2$ depends on the form of the interval $I$ and is separately solved for the cases of $I=[0,t_1]$ and $I=[t_1, t_2]$ where $t_1, t_2 \in \mathbb{R}_{\geq 0}$. It is based on a modification of the uniformized infinitesimal generator matrix $\mathbf{Q}^{\mathsf{unif}}$ where certain states are made absorbing. This means that all outgoing transitions are ignored in dependence on the validity of $\Phi_1$ and $\Phi_2$ in these states.
For any CSL formula $\Phi$, let $\mathcal{C}[\Phi] = \left( \mathbb{S}, s_0, \mathbf{R}[\Phi], L \right)$, where $\mathbf{R}[\Phi](s,s') = \mathbf{R}(s,s')$, if $s \vDash \Phi$, and $0$, otherwise. The formula $\phi = \Phi_1 \ \mathsf{U}^{[0,t]} \ \Phi_2$ can be evaluated using the vector $\tau^{\mathcal{C},\mathbb{A},t}$ in the following way: $$\overline{Prob}^{\mathcal{C}, \phi} = \tau^{\mathcal{C}[\neg \Phi_1 \wedge \Phi_2], \mathbb{A}, t} \mbox{~where~} s \in \mathbb{A} \mbox{~iff~} s \vDash \Phi_2.$$ For the formula $\phi = \Phi_1 \ \mathsf{U}^{[t_1,t_2]} \ \Phi_2$ the evaluation is split into two parts: staying in states satisfying $\Phi_1$ until time $t_1$ and reaching a state satisfying $\Phi_2$, while remaining in states satisfying $\Phi_1$, within time $t_2-t_1$. The formula $\phi$ can be evaluated using the vector $\tau^{\mathcal{C},\overline{v},t}$ that takes a vector $\overline{v}$ instead of a set $\mathbb{A}$ (i.e., $\tau^{\mathcal{C}, \overline{v}, 0} = \overline{v}$) in the following way: $$\overline{Prob}^{\mathcal{C}, \phi} = \tau^{\mathcal{C}[\neg \Phi_1], \overline{v}, t_1} \mbox{~where~} \overline{v} = \tau^{\mathcal{C}[\neg \Phi_1 \wedge \Phi_2], \mathbb{A}, t_2-t_1} \mbox{~and~} s \in \mathbb{A} \mbox{~iff~} s \vDash \Phi_2.$$
The backward transient analysis can be also used in the case of reward computation. Since operator $\mathsf{R}_{\sim p}[\mathsf{I}^{= t}]$ expresses the expected reward at time $t$, the vector $\overline{Exp}^{\mathcal{C}, \mathsf{X}_{\mathsf{I}^{= t}}}$ can be computed as follows: $$\overline{Exp}^{\mathcal{C}, \mathsf{X}_{\mathsf{I}^{= t}}} = \tau^{\mathcal{C}, \overline{v}, t} \mbox{~where~} \overline{v} = \rho \mbox{~such that~} \rho \mbox{~is the given state reward structure.}$$ For evaluation of the operator $\mathsf{R}_{\sim p}[\mathsf{C}^{\leq t}]$ we have to use *mixed Poisson probabilities* (see, e.g., [@Kwiatkowska2006rewards; @Kwiatkowska2007]) in the backward transient analysis. It means that during the uniformization the Poisson probabilities $\gamma_{i,q \cdot t}$ are replaced by the mixed Poisson probabilities $\bar{\gamma}_{i,q \cdot t}$ that can be computed as: $$\bar{\gamma}_{i,q \cdot t} = \frac{1}{q} \cdot \left( 1- \sum_{j = 1}^{i} \gamma_{j,q \cdot t} \right) \mbox{.}$$ Using the given state reward structure $\rho$ we can compute the vector $\overline{Exp}^{\mathcal{C}, \mathsf{X}_ {\mathsf{C}^{\leq t}}}$ in the following way:
$$\overline{Exp}^{\mathcal{C}, \mathsf{X}_ {\mathsf{C}^{\leq t}}} = \tau^{\mathcal{C}, \overline{v}, t} \mbox{~where~} \overline{v} = \rho \mbox{~and the mixed Poisson probabilities~} \bar{\gamma}_{i,q \cdot t} \mbox{~are used.}$$
To recap the overall method of stochastic model checking of CTMCs over CSL formulae we present the methods from an abstract perspective. The evaluation of a structured formula $\Phi$ proceeds by bottom-up evaluation of a set of atomic propositions, probabilistic or expected reward inequalities and their boolean combinations. This evaluation gives us a discrete set of states that are further used in the following computation. The process continues up the formula until the root is reached. The final verdict is reported either in the form of a boolean yes/no answer or as the actual numerical value of the probability or the expected reward. This process can be easily extended for the operator $\mathsf{E}_{\sim r}[I^{=t}]$, however, the local model checking method has to be used.
### Min-max approximation
The key idea of min-max approximation is to approximate the *largest set of states satisfying* $\Phi$, and the *smallest set of states satisfying* $\Phi$ with respect to the space of perturbations $\mathbf{P}$. Let $\mathbf{C}$ be a set of parameterised CTMCs induced by the space of perturbations $\mathbf{P}$ in the system $\mathcal{S}$. We compute the approximation $Sat^{\top}_{\mathbf{C}}(\Phi)$ and $Sat^{\bot}_{\mathbf{C}}(\Phi)$ such that $$Sat^{\top}_{\mathbf{C}}(\Phi) \supseteq \bigcup_{\mathcal{C}_p \in \mathbf{C}} Sat_{\mathcal{C}_p}(\Phi) \ \wedge \ Sat^{\bot}_{\mathbf{C}}(\Phi) \subseteq \bigcap_{\mathcal{C}_p \in \mathbf{C}} Sat_{\mathcal{C}_p}(\Phi)$$ where $s\in Sat_{\mathcal{C}_p}(\Phi)$ iff $s\vDash \Phi$ in CTMC $\mathcal{C}_p$. To obtain such approximations we extended the satisfaction relation $\vDash$ and showed that it is sufficient for an arbitrary path formula $\phi$, and $\mathsf{X} \in
\{\mathsf{X}_ {\mathsf{C}^{\leq t}}, \mathsf{X}_ {\mathsf{I}^{= t}}\}$ to compute the vectors $\overline{Prob}^{\mathbf{C}, \phi}_{\top},\ \overline{Prob}^{\mathbf{C}, \phi}_{\bot},\ \overline{Exp}^{\mathbf{C}, \mathsf{X}}_{\top}$ and $\overline{Exp}^{\mathbf{C}, \mathsf{X}}_{\bot}$ such that for each $s\in \mathbb{S}$ the following holds: $$\label{eq:glob}
\begin{array}{rl}
\overline{Prob}^{\mathbf{C}, \phi}_{\top}(s) & \geq max\{ \overline{Prob}^{\mathcal{C}_p, \phi}(s) \mid \mathcal{C}_p \in \mathbf{C} \} \\
\overline{Prob}^{\mathbf{C}, \phi}_{\bot}(s) & \leq min \{ \overline{Prob}^{\mathcal{C}_p, \phi}(s) \mid \mathcal{C}_p \in \mathbf{C} \} \\
\overline{Exp}^{\mathbf{C}, \mathsf{X}}_{\top}(s) & \geq max\{ \overline{Exp}^{\mathcal{C}_p, \mathsf{X}}(s) \mid \mathcal{C}_p \in \mathbf{C} \} \mbox{~for~} \mathsf{X} \in \{ \mathsf{X}_ {\mathsf{I}^{= t}}, \mathsf{X}_ {\mathsf{C}^{\leq t}} \}\\
\overline{Exp}^{\mathbf{C}, \mathsf{X}}_{\bot}(s) & \leq min\{ \overline{Exp}^{\mathcal{C}_p, \mathsf{X}}(s) \mid \mathcal{C}_p \in \mathbf{C} \} \mbox{~for~}\mathsf{X} \in \{ \mathsf{X}_ {\mathsf{I}^{= t}}, \mathsf{X}_ {\mathsf{C}^{\leq t}} \}.\\
\end{array}$$
The min-max approximation can be easily extended to the operator $\mathsf{E}_{\sim r}[I^{=t}]$. For the given state $s \in \mathbb{S}$ and the time *t* it is sufficient to compute the values $Post^{\mathbf{C}}_{\top}(s,t)$ and $Post^{\mathbf{C}}_{\top}(s,t)$ such that the following holds: $$\label{eq:loc}
\begin{array}{rl}
Post^{\mathbf{C}}_{\top}(s,t) & \geq max\{ Post(\pi^{\mathcal{C}_p,s,t}) \mid \mathcal{C}_p \in \mathbf{C} \} \\
Post^{\mathbf{C}}_{\bot}(s,t) & \leq min\{ Post(\pi^{\mathcal{C}_p,s,t}) \mid \mathcal{C}_p \in \mathbf{C} \}. \\
\end{array}$$
The approximated sets ${Sat}^{\top}_{\mathbf{C}}(\Phi)$ and ${Sat}^{\bot}_{\mathbf{C}}(\Phi)$ are further used in the computation of $D_{\Phi,\mathbf{P},\top}^{\mathcal{C},s}$ and $D_{\Phi,\mathbf{P},\bot}^{\mathcal{C},s}$. If the topmost operator of the formula $\Phi$ is $\mathsf{P}_{= ?}[\phi]$ then $$D_{\Phi,\mathbf{P},\bot}^{\mathcal{C},s} =
\overline{Prob}^{\mathbf{C}, \phi}_{\bot}(s) \wedge D_{\Phi,\mathbf{P},\top}^{\mathcal{C}, s} =
\overline{Prob}^{\mathbf{C}, \phi}_{\top}(s).$$ If the topmost operator of the formula $\Phi$ is $\mathsf{R}_{= ?}[\mathsf{C}^{\leq t}]$ and $\mathsf{R}_{= ?}[\mathsf{I}^{= t}]$ then $$D_{\Phi,\mathbf{P},\bot}^{\mathcal{C},s} =
\overline{Exp}^{\mathbf{C}, \mathsf{X}}_{\bot}(s) \wedge D_{\Phi,\mathbf{P},\top}^{\mathcal{C},s} =
\overline{Exp}^{\mathbf{C}, \mathsf{X}}_{\top}(s) \mbox{~for~} \mathsf{X} =\mathsf{X}_
{\mathsf{C}^{\leq t}} \mbox{~and~} \mathsf{X} = \mathsf{X}_ {\mathsf{I}^{= t}} \mbox{, respectively}.$$ Similarly, if the topmost operator of the formula $\Phi$ is $\mathsf{E}_{= ?}[\mathsf{I}^{= t}]$ then $$D_{\Phi,\mathbf{P},\bot}^{\mathcal{C}, s} = Post^{\mathbf{C}}_{\bot}(s,t)
\wedge D_{\Phi,\mathbf{P},\top}^{\mathcal{C}, s} =
Post^{\mathbf{C}}_{\top}(s,t).$$
### Parameterised uniformisation {#sec:parameteriseduniformisation}
Recall that the most crucial part of the robustness computation is given by the fact that the space of perturbations of stochastic rate constants $\mathbf{P}$ is dense and thus the set $\mathbf{C}$ is infinite. Therefore, it is not possible to employ the standard model checking techniques to compute the result for each CTMC $\mathcal{C}_p \in$ $\mathbf{C}$ individually.
In order to overcome this problem we employ parameterised uniformisation introduced in [@CAV2013]. It is a modification of the standard uniformisation technique that allows us to compute strict approximations of the minimal and maximal transient probability with respect to the set $\mathbf{C}$, moreover, the modification preserves the asymptotic time complexity of standard uniformisation. For the given state $s \in \mathbb{S}$ and time $t \in \mathbb{R}_{\geq 0}$ the parameterised uniformisation returns vectors $\pi^{\mathbf{C}, s, t}_{\top}$ and $\pi^{\mathbf{C}, s,
t}_{\bot}$ such that for each state $s' \in \mathbb{S}$ the following holds: $$\pi^{\mathbf{C}, s, t}_{\top}(s') \geq max\{\pi^{\mathcal{C}_p, s, t}(s') \mid \mathcal{C}_p \in \mathbf{C}\} \ \wedge \ \pi^{\mathbf{C}, s, t}_{\bot}(s') \leq min\{\pi^{\mathcal{C}_p, s, t}(s') \mid \mathcal{C}_p \in \mathbf{C}\}$$
The modification is based on the computation of the local maximum (minimum) of $\pi^{\mathcal{C}_p, s, t}(s')$ over all $\mathcal{C}_p \in \mathbf{C}$ for each state *s’* and in each iteration $i$ of standard uniformisation. It means that in the *i*th iteration of the computation for a state *s’* we consider only the maximal (minimal) values in the relevant states in the iteration *i-1*, i.e., the states that affect $\pi^{\mathcal{C}_p, s, t}(s')$.
In [@CAV2013] we have defined the function $\sigma(s)$ (formally $\sigma(p,s,\pi)$) which for each state $s \in \mathbb{S}$, perturbation point $p\in \mathbf{P}$ and probability distribution $\pi$ (or pseudo-distribution with the sum smaller or larger than $1$) returns the difference of probability mass inflow and outflow to/from state *s*. If all reactions are described by mass action kinetics the resulting $\sigma$ functions are monotonic with respect to any single perturbed stochastic rate constant $k_r$. This allows us to efficiently compute for each state $s'$ the local maximum (minimum) of $\pi^{\mathcal{C}_p, s, t}(s')$ over all $\mathcal{C}_p \in \mathbf{C}$ corresponding to $\mathbf{P}$.
However, in the case of more complex rate functions than those resulting from mass action kinetics, the corresponding $\sigma(s)$ function does not have to be in general monotonic over $k_r \in [k_r^{\bot},k_r^{\top}]$ for all states *s*. This makes the computation of local extremes with respect to $k_r$ more complex however still tractable. In the following let us assume the space of perturbations $\mathbf{P} = [k_r^{\bot},k_r^{\top}] \times \mathbf{P}'$ will be decomposed along the $k_r$ axis.
The key idea is for each state *s* to be able to efficiently decompose $\mathbf{P}$ into subspaces $\mathbf{P} = \mathbf{P}_1 \cup \ldots \cup \mathbf{P}_n$, such that for each $\mathbf{P}_i$ the function $\sigma(s)$ over $\mathbf{P}_i$ is monotonic and then use the original method. The problem is a computation of such a strict decomposition into monotonic subspaces is computationally demanding. Therefore we use a simplification, by off-line functional analysis we identify properties of $\sigma$ functions for a given class of reaction kinetics and then obtain a partial decomposition of $\mathbf{P}$ based on function derivations into subspaces where monotonicity is guaranteed. For the remaining subspaces $\mathbf{P}_j$ where monotonicity of $\sigma$ is not guaranteed we employ a less accurate approximation.
We decompose the function $\sigma(s)$ over ${\mathbf{P}_j}$ into functions $\alpha^{s,\mathbf{P}_j}_k$ and $\beta^{s,\mathbf{P}_j}_l$ such that: $$\sigma(s) = \sum_{k=1}^K \alpha^{s,\mathbf{P}_j}_k - \sum_{l=1}^L \beta^{s,\mathbf{P}_j}_l$$ and each $\alpha^{s,\mathbf{P}_j}_k$ and $\beta^{s,\mathbf{P}_j}_l$ is monotonic. This allows us to use the original method to compute the maximum and minimum of the functions $\alpha^{s,\mathbf{P}_j}_k$ and $\beta^{s,\mathcal{P}_j}_l$ over the interval $\mathbf{P}_j$, denoted as $max(\alpha^{s,\mathbf{P}_j}_k)$, $min(\alpha^{s,\mathbf{P}_j}_k)$, $max(\beta^{s,\mathbf{P}_j}_k)$ and $min(\beta^{s,\mathbf{P}_j}_k)$, respectively. Note that, this decomposition can be easily obtained from the definition of the rate function $f_r$. Now the maximum and minimum of $\sigma(p,s)$ over $\mathbf{P_j}$ can be approximated in the following way: $$max\{\sigma(p,s) \mid p \in \mathbf{P}_j\} \leq \sum_{k=1}^K max(\alpha^{s,\mathbf{P}_j}_k) - \sum_{l=1}^L min(\beta^{s,\mathbf{P}_j}_l)$$ $$min\{\sigma(p,s) \mid p \in \mathbf{P}_i\} \geq \sum_{k=1}^K min(\alpha^{s,\mathbf{P}_i}_k) - \sum_{l=1}^L max(\beta^{s,\mathbf{P}_j}_l).$$ This approximation increases the inaccuracy of parameterised uniformisation, however, the subspaces $\mathbf{P}_j$ where the monotonicity of $\sigma(s)$ is not guaranteed are usually small and together with perturbation space decomposition introduced in the following section keep on getting smaller. Hence, the additional inaccuracy of the presented extension is manageable. Despite the fact that the time demands of this approximation are orders of magnitudes lower than other numerical methods computing maximum/minimum of $\sigma(s)$ over $\mathbf{P}_i$, they still significantly slow down the computation of parameterised uniformisation.
The aforementioned parameterised uniformisation can be straightforwardly employed also for backward transient analysis. It means that we can efficiently compute the vectors $\tau^{\mathbf{C},\mathbb{A},t}_{\top}$ and $\tau^{\mathbf{C},\mathbb{A},t}_{\bot}$ such that for the given set of states $\mathbb{A}$ and each state $s \in \mathbb{S}$ the following holds: $$\tau^{\mathbf{C}, \mathbb{A}, t}_{\top}(s) \geq max\{\tau^{\mathcal{C}_p, \mathbb{A}, t}(s) \mid \mathcal{C}_p \in \mathbf{C}\} \ \wedge \ \tau^{\mathbf{C}, \mathbb{A}, t}_{\bot}(s) \leq min\{\tau^{\mathcal{C}_p, \mathbb{A}, t}(s) \mid \mathcal{C}_p \in \mathbf{C}\}$$
Once we know how to compute the vectors $\tau^{\mathbf{C}, \mathbb{A}, t}_{\top}$ and $\tau^{\mathbf{C}, \mathbb{A}, t}_{\bot}$ the global model checking technique for non-parameterised CTMCs can be directly employed. To obtain the vectors $\overline{Prob}^{\mathbf{C}, \phi}_{\top},\ \overline{Prob}^{\mathbf{C}, \phi}_{\bot},\ \overline{Exp}^{\mathbf{C}, \mathsf{X}}_{\top}$ and $\overline{Exp}^{\mathbf{C}, \mathsf{X}}_{\bot}$ satisfying Equation \[eq:glob\], it is sufficient to replace the backward transient distribution $\tau^{\mathcal{C}, \mathbb{A},t}_{\bot}$ by the vectors $\tau^{\mathbf{C}, \mathbb{A}, t}_{\top}$ and $\tau^{\mathbf{C}, \mathbb{A}, t}_{\top}$. However for a general class of user-defined post-processing functions *Post*, the vectors $\pi^{\mathbf{C}, s, t}_{\top}$ and $\pi^{\mathbf{C}, s,t}_{\bot}$ cannot be directly used to compute values of $Post^{\mathbf{C}}_{\top}(s,t) = Post(\pi^{\mathbf{C}, s, t}_{\top})$ nor $Post^{\mathbf{C}}_{\bot}(s,t) = Post(\pi^{\mathbf{C}, s, t}_{\bot})$ that would satisfy Equation \[eq:loc\] since there is no guarantee about the projective properties of the function *Post*.
Now we show the main idea how to compute $Post^{\mathbf{C}}_{\top}(s,t)$ and $Post^{\mathbf{C}}_{\bot}(s,t)$ for the post-processing function *Post* defined as the mean quadratic deviation of a probability distribution. This function allows us to quantify and analyse a noise in different variants of signalling pathways that are studied in the second case study. The post-processing function is defined as $Post(\pi) = \sum_{s \in \mathbb{S}}{\lvert s(A) - mean(\pi,A)\rvert^2 \cdot \pi(s)}$, where $s(A)$ gives the population of *A* in state *s* and $mean(\pi,A)$ is the mean of the distribution $\pi$ defined as $mean(\pi,A) = \sum_{s \in \mathbb{S}}{s(A) \cdot \pi(s)}$.
Let us suppose we have an upper and lower bound on the probability distribution $\pi_{\top}, \pi_{\bot}$ obtained by the parameterised uniformisation. It means that $\forall \mathcal{C}_p \in \mathbf{C}$ and $\forall s \in \mathbb{S}. \ \pi_{\bot}(s) \leq \pi^{\mathcal{C}_p}(s) \leq \pi_{\top}(s)$. To find the maximal value $max \left\{ Post(\pi^{\mathcal{C}_p}) \mid \mathcal{C}_p \in \mathbb{C} \right\}$ means to find the distribution $\pi^{max}$ such that $\sum_{s \in \mathbb{S}}{\pi^{max}(s)} = 1$, $\forall s\in \mathbb{S}. \ \pi_{\bot}(s) \leq \pi^{max}(s) \leq \pi_{\top}(s)$ and the probability mass in $\pi^{max}$ is distributed with the farest distance from the mean. Clearly, such a distribution has a maximal mean quadratic deviation. Note that the number of distributions satisfying the first two conditions is uncountable. Thus we cannot employ direct searching strategy.
Our searching strategy builds on the observation that only distributions that localise most of the mass as far as possible from the mean (i.e., maximizing the mean quadratic deviation and still meeting the bounds $\pi_{\bot}, \pi_{\top}$), have to be considered. These distributions can be linearly ordered with respect to the sum of mass *x* localised at the low populated part of the state space. It can be shown that the function that evaluates $Post$ on all these distributions is piece-wise quadratic with respect to *x* and has $O(|\mathbb{S}|)$ segments. Therefore, $O(|\mathbb{S}|)$ many steps are sufficient to compute $max \left\{ Post(\pi^{\mathcal{C}_p}) \mid \mathcal{C}_p \in \mathbb{C} \right\}$.
To compute the minimal value $min \left\{ Post(\pi^{\mathcal{C}_p}) \mid \mathcal{C}_p \in \mathbb{C} \right\}$ we proceed analogously, i.e., only the distributions that localise most of the mass as close as possible to the mean are considered. This leads again to a piece-wise quadratic function. It is also important to note that the perturbation space decomposition presented in the next section allows us to obtain the values $Post^{\mathbf{C}}_{\top}(s,t)$ and $Post^{\mathbf{C}}_{\bot}(s,t)$ with the desired precision.
![[**Perturbation space refinement.**]{} Part (A) depicts three resulting probabilities (green dots) of the formula $\Phi$ for three values of the rate $k_1$ corresponding to three perturbation points $p \in \mathbf{P}$ from Figure \[fig:decomposition\] for the initial state $s_0$ denoted as $Prob^{\mathcal{C}_p}(s_0,\Phi)$. The shape of $Prob^{\mathcal{C}_p}(s_0,\Phi)$ for all $p\in \mathbf{P}$ is estimated upon these three points by polynomial interpolation and shown as a black curve. The top four parts (A), (B), (C) and (D) illustrate the min-max approximation of $Prob^{\mathcal{C}_p}(s_0,\Phi)$ for all $p\in \mathbf{P}$ using the decomposition of $\mathbf{P}$ into 2, 4, 8 and 16 subspaces. The exact shape of the probability function for $\Phi$ is visualised as the red thick curve in the (D) and is compared to the initial estimate. Two types of errors are illustrated: the approximation error is depicted as yellow rectangles and the uniformisation error as the pink rectangles. As can be seen a more refined decompositions reduces both types of errors in each further refined subspace. The bottom parts (E) and (F) depict how the errors arise and how they can be reduce using perturbation space decomposition.[]{data-label="fig:refinement"}](minmax_all){width="\textwidth"}
### Perturbation space decomposition
As we already mentioned, a finite decomposition $\mathbf{P} = \mathbf{P}_1 \cup \ldots \cup \mathbf{P}_n$ into perturbation subspaces is used in order to obtain more accurate approximation of the evaluation function $D^{\mathcal{C}}_{\Phi}$ over the perturbation space $\mathbf{P}$. Before we describe perturbation space decomposition we briefly discuss the key characteristics of parameterised uniformisation that helps us to understand the source of the inaccuracy. The most important fact is that parameterised uniformisation for the set $\mathbf{C}$ in general does not correspond to standard uniformisation for any CTMC $\mathcal{C}_p \in \mathbf{C}$. The reason is that we consider a behaviour of a parameterised CTMC that has no equivalent counterpart in any particular $\mathcal{C}_p$. First, the parameter $k_r$ (minimizing/maximizing the inspected value) is determined locally for each state. Therefore, in a single iteration there can exist two different states such that in one state the parameterised uniformisation selects $k_r=k_r^{\top}$ while in another state it selects $k_r=k_r^{\bot}$. Second, the parameter is determined individually for each iteration and thus for a state $s_i$ the parameter $k_r$ can be chosen differently in individual iterations.
Inaccuracy of the proposed min-max approximation related to the computation of parameterised uniformisation, called *unification error*, is given as: $$(D^{\mathcal{C}}_{\Phi,\mathbf{P},\top} - max \{ D^{\mathcal{C}}_{\Phi}(p) \mid p \in \mathbf{P} \}) +
(min \{ D^{\mathcal{C}}_{\Phi}(p) \mid p \in \mathbf{P} \} - D^{\mathcal{C}}_{\Phi,\mathbf{P},\bot}).$$ Apart from the unification error our approach introduces an inaccuracy related to approximation of the evaluation function, called *approximation error*, given as: $$max \{ D^{\mathcal{C}}_{\Phi}(p) \mid p \in \mathbf{P} \} - min \{ D^{\mathcal{C}}_{\Phi}(p) \mid p \in \mathbf{P} \}.$$ Finally, the *overall error* of the min-max approximation, denoted as $\mathsf{Err}^{\mathcal{C}}_{\Phi, \mathbf{P}}$, is defined as a sum of both errors, i.e., $D^{\mathcal{C}}_{\Phi,\mathbf{P},\top} - D^{\mathcal{C}}_{\Phi,\mathbf{P},\bot}$. Figure \[fig:refinement\] illustrates both types of errors. The approximation error is depicted as yellow rectangles and the unification error is depicted as the purple rectangles.
We are not able to effectively distinguish the proportion of the approximation error and the unification error nor to reduce the unification error as such. Therefore, we design a method based on the perturbation space decomposition that allows us to effectively reduce the overall error of the min-max approximation to a user specified *absolute error bound*, denoted as $\textsc{Err}$.
In order to ensure that the min-max approximation meets the given absolute error bound $\textsc{Err}$, we iteratively decompose the perturbation space $\mathbf{P}$ into finitely many subspaces such that $\mathbf{P} = \mathbf{P}_1 \cup \ldots \cup
\mathbf{P}_n$ and each partial result satisfies the overall error bound, i.e., $\forall \ j: 1 \leq j \leq n: \mathsf{Err}^{\mathcal{C}}_{\Phi, \mathbf{P}_j} \leq \textsc{Err}$. Therefore, the overall error equals to $$\mathsf{Err}^{\mathcal{C}}_{\Phi, \mathbf{P}}= \sum_{j =
1}^{n}{\frac{|\mathbf{P}_j|}{|\mathbf{P}|} \left(D^{\mathcal{C}}_{\Phi,\mathbf{P}_j,\top} - D^{\mathcal{C}}_{\Phi,\mathbf{P}_j,\bot} \right)} \leq \sum_{j =
1}^{n}{\frac{|\mathbf{P}_j|}{|\mathbf{P}|} \textsc{Err}} = \textsc{Err}.$$ Figure \[fig:refinement\] illustrates such a decomposition and demonstrates convergence of $\mathsf{Err}^{\mathcal{C}}_{\Phi, \mathbf{P}_j}$ to 0 provided that the evaluation function $D^{\mathcal{C}}_{\Phi}$ over $\mathbf{P}$ is continuous.
For sake of simplicity, we present the parametric decomposition only on the computation of $\pi^{\mathbf{C}, s, t}_{\star}$ and $\tau^{\mathbf{C}, \mathbb{A}, t}_{\star}$ for $\star \in \{\top,\bot\}$ and $\mathbf{P}$ since it can be easily extended to the computation of $D_{\Phi,\mathbf{P},\star}^{\mathcal{C}}$ for any formula $\Phi$. The key part of the parametric decomposition is to decide when the inspected subspace should be further decomposed. The condition for the decomposition is different for $\pi^{\mathbf{C}, s, t}_{\star}$ and $\tau^{\mathbf{C}, \mathbb{A}, t}_{\star}$. Since the vector $\pi^{\mathbf{C}, s, t}_{\star}$ gives us the transient probability distribution from the state $s$ that is further used to compute $D^{\mathcal{C},s}_{\Phi,\mathbf{P},\star}$, we consider the following condition. The space $\mathbf{P}$ (represented by the CTMC $\mathbf{C}$) is decomposed if during the computation of parameterised uniformisation in an iteration $i$ it holds that: $$\sum_{k =1}^{|\mathbb{S}|} {\pi^{\mathbf{C}, s, i}_{\top}(s_k)} - \sum_{k =1}^{|\mathbb{S}|} {\pi^{\mathbf{C}, s, i}_{\bot}(s_k)} > \textsc{Err}$$ where $ \pi^{\mathbf{C}, s, i}_{\star}$ denotes the corresponding approximation of $\pi^{\mathbf{C}, s, 0} \cdot (\mathbf{Q}^{\mathsf{unif}(\mathbf{C})})^i$.
In contrast to $\pi^{\mathbf{C}, s, t}_{\star}$, the value $\tau^{\mathbf{C}, \mathbb{A}, t}_{\top}(s)$ for each state $s\in \mathbb{S}$ is further used to $D^{\mathcal{C},s}_{\Phi,\mathbf{P},\star}$ and thus we consider the different condition. The space $\mathbf{P}$ is decomposed if during the computation of parameterised uniformisation in an iteration $i$ for any state $s$ it holds that: $$\tau^{\mathbf{C}, \mathbb{A}, i}_{\top}(s) - \tau^{\mathbf{C}, \mathbb{A}, i}_{\bot}(s) > \textsc{Err}.$$
If the decomposition takes place we cancel the current computation and decompose the perturbation space $\mathbf{P}$ to $n$ subspaces such that $\mathbf{P} = \mathbf{P}_1 \cup
\ldots \cup \mathbf{P}_n$. Each subspace $\mathbf{P}_j$ defines a new set of CTMCs $\mathbf{C}_j = \{\mathcal{C}_j \mid j \in \mathbf{P}_j \}$ that is independently processed in a new computation branch. Note that we could reuse the previous computation and continue from the iteration $i-1$. However, the most significant part of the error is usually cumulated during the previous iterations and thus the decomposition would have only a negligible impact on error reduction.
*A minimal decomposition with respect to the perturbation space $\mathbf{P}$* defines a minimal number of subspaces *m* such that $\mathbf{P} = \mathbf{P}_1 \cup \ldots
\cup \mathbf{P}_{m}$ and for each subspace $\mathbf{P}_j$ where $1 \leq j \leq m$ holds that $\mathsf{Err}^{\mathcal{C},s}_{\Phi, \mathbf{P}_j} \leq \textsc{Err}$ where $\mathsf{Err}^{\mathcal{C},s}_{\Phi, \mathbf{P}_j} = D^{\mathcal{C}, s}_{\Phi,\mathbf{P}_j,\top} - D^{\mathcal{C},s}_{\Phi,\mathbf{P}_j,\bot}$. Note that the existence of such decomposition is guaranteed only if the evaluation function $D^{\mathcal{C},s}_{\Phi}$ over $\mathbf{P}$ is continuous. If the evaluation function is continuous there can exist more than one minimal decomposition. However, it can not be straightforwardly found. To overcome this problem we have considered and implemented several heuristics allowing to iteratively compute a decomposition satisfying the following: (1) it ensures the required error bound whenever $D^{\mathcal{C},s}_{\Phi}$ over $\mathbf{P}$ is continuous, (2) it guarantees the refinement termination in the situation where$D^{\mathcal{C},s}_{\Phi}$ over $\mathbf{P}$ is not continuous and the discontinuity causes that $\textsc{Err}$ can not be achieved. To ensure the termination an additional parameter has to be introduced as a lower bound on the subspace size. Hence this parameter provides a supplementary termination criterion.
### Implementation
We delivered a prototype implementation of the framework for the robustness analysis on top of the tool PRISM 4.0 [@KNP11]. This tool provides the appropriate modelling and specification language. Our implementation builds on sparse engine that uses data structures based on the sparse matrices. They provides suitable representation of models for the time efficient numerical computation.
In the case that large number of perturbation subspaces is required to obtain the desired accuracy of the approximation the sequential computation can be extremely time consuming. However, our framework allows very efficient parallelization since the the computation of particular subspaces is independent and thus can be executed in parallel. Our implementation enables the parallel computation and thus the robustness analysis can be significantly accelerated using high performance parallel hardware architectures.
Results
=======
Gene Regulation of Mammalian Cell Cycle
---------------------------------------
We have applied the robustness analysis to the gene regulation model published in [@Keletal00], the regulatory network is shown in Fig. \[fig:modelg1s\] (left). The model explains regulation of a transition between early phases of the mammalian cell cycle. In particular, it targets the transition from the control $G_1$-phase to *S*-phase (the synthesis phase). $G_1$-phase makes an important checkpoint controlled by a *bistable regulatory circuit* based on an interplay of the retinoblastoma protein *pRB*, denoted by *A* (the so-called tumour suppressor, HumanCyc:HS06650) and the retinoblastoma-binding transcription factor $E_2F_1$, denoted by *B* (a central regulator of a large set of human genes, HumanCyc:HS02261). In high concentration levels, the $E_2F_1$ protein activates the $G_1$/$S$ transition mechanism. On the other hand, a low concentration of $E_2F_1$ prevents committing to *S*-phase.
Positive autoregulation of *B* causes bi-stability of its concentration depending on the parameters. Especially, of specific interest is the degradation rate of *A*, $\gamma_{A}$. In [@Swatetal04] it is shown that for increasing $\gamma_{A}$ the low stable mode of *B* switches to the high stable mode. When mitogenic stimulation increases under conditions of active growth, rapid phosphorylation of *A* starts and makes the degradation of unphosphorylated *A* stronger (the degradation rate $\gamma_{A}$ increases). This causes *B* to lock in the high stable mode implying the cell cycle commits to *S*-phase. Since mitogenic stimulation influences the degradation rate of *A*, our goal is to study the population distribution around the low and high steady state and to explore the effect of $\gamma_{A}$ by means of the evaluation function.
It is necessary to note that the original ODE model in [@Swatetal04] has been formalised by means of Hill kinetics representing the cooperative action of transcription factor molecules. Since Hill kinetics cannot be directly transferred to stochastic modelling [@Garaietal12; @Sanftetal11], we have reformulated the model in the framework of stochastic mass action kinetics [@Gillespie1977]. The resulting reactions are shown in Fig. \[fig:modelg1s\] (right). Since the detailed knowledge of elementary chemical reactions occurring in the process of transcription and translation is incomplete, we use the simplified form as suggested in [@Gillespieetal05]. In the minimalist setting, the reformulation requires addition of rate parameters describing the transcription factor–gene promoter interaction while neglecting cooperativeness of transcription factors activity. Our parameterisation is based on time-scale orders known for the individual processes [@Yang2003] (parameters considered in $s^{-1}$). Moreover, we assume the numbers of *A* and *B* are bounded by 10 molecules. Correctness of the upper bounds for *A* and *B* was validated by observing thousand independent stochastic simulations. We consider minimal population number distinguishing the two stable modes. All other species are bounded by the initial number of DNA molecules (genes *a* and *b*) which is conserved and set to 1. The corresponding CTMC has 1078 states and 5919 transitions.
![[**Model of regulation of the mammalian cell cycle.**]{} The core gene regulatory module controlling the $G_1/S$-phase transition in the cell cycle of mammalian cells [@Keletal00] is depicted in the upper part. The retinoblastoma protein *pRB* (A) \[HumanCyc:HS06650\] interacts with the retinoblastoma-binding transcription factor $E_2F_1$ (B) \[HumanCyc:HS02261\]. In high concentration levels, the $E_2F_1$ protein activates the $G_1/S$ transition mechanism. On the other hand, a low concentration of $E_2F_1$ prevents committing to $S$-phase. Positive autoregulation of $E_2F_1$ causes bi-stability.\
Stochastic mass action reformulation of the $G_1/S$ regulatory circuit is shown in the table below. The gene regulation is modelled by means of a set of second-order reactions simplifying the elementary processes behind transcription. In particular, the model includes the interactions among transcription factors (*A*, *B* stand for *pRB* and $E_2F_1$, respectively) and respective genes and protein production/degradation reactions. The interactions are represented by reversible TF-gene binding reactions in the second row of the table (genes are denoted by small letters). Individual protein production reactions controlled by these interactions are represented by the irreversible gene expression reactions in the first row of the table. Protein degradation is modelled as spontaneous by means of first-order reactions. Kinetic coefficients are set only approximately provided that they are considered equal for all instances of a particular process (binding, dissociation, promoted protein production). The only exception is the spontaneous (basal) expression of *b* which is set to a low rate. This mimics the fact that $E_2F_1$ is only rapidly produced under the circumstances of self-activation [@Swatetal04]. Degradation parameters are left unspecified.[]{data-label="fig:modelg1s"}](cs1_model){width="50.00000%"}
We consider two hypotheses: (1) stabilisation in the low mode where $B<3$, (2) stabilisation in the high mode where $B>7$. Both hypotheses are expressed within time horizon 1000 seconds reflecting the time scale of gene regulation response. According to [@Swatetal04], we consider the perturbation space $\gamma_{A}\in
[0.005,0.5]$. For both hypothesis we consider three different settings of $\gamma_{B}$: $\gamma_B=0.05$, $\gamma_B=0.10$, and $\gamma_B=0.15$.
We employ two alternative CSL formulations to express the hypothesis (1). First, we express the property of being inside the given bound during the time interval $I=[500,1000]$ using globally operator: $\mathsf{P}_{=?}[\mathsf{G}^I\, (B< 3)]$. The interval starts from 500 seconds in order to bridge the initial fluctuation region and let the system stabilise. The resulting landscape visualisation is depicted in Figure \[fig:cs1\_until\_0-2\] together with the robustness values computed for individual cases. Since the stochastic noise causes molecules to repeatedly escape the requested bound, the resulting probability is significantly lower than expected. Namely, in the case $\gamma_B=0.05$ the resulting probability is close to 0 for almost all considered parameter values implying very small robustness. Increasing of the *B* degradation rate causes an observable increase in robustness.
![[**Results of robustness analysis for hypothesis (1) using a until operator.**]{} Hypothesis (1) requires stabilisation of $E_2F_1$ in the low concentration mode ($B<3$). A CSL formula with the until operator is used in this case. Each of the curves represents the evaluation function over $\gamma_A$ degradation obtained for a particular setting of $\gamma_B$. More precisely, the horizontal axis shows the perturbation of *pRB* degradation rate and the vertical axis shows the probability of the hypothesis to be satisfied. In the upper left corner, robustness values are shown for each of the curves. The values are displayed with the absolute error quantifying the precision of the approximate method. For comparison, the values are computed also on piece-wise affine approximations of the evaluation function. It can be seen that the robustness values are small which is due to the fact that fluctuations of molecular numbers cause frequent exceeding of the required bound in the considered time horizon.[]{data-label="fig:cs1_until_0-2"}](cs1_until_0-2){width="\textwidth"}
In order to avoid fluctuations of affecting the result, we use a cumulative reward property to capture the fraction of the time the system has the required number of molecules within the time interval $[0,1000]$: $\mathsf{R}_{=?}[\mathsf{C}^{\leq t}](B<3)$ where $t = 1000$ and $\mathsf{R}_{=?}[\mathsf{C}^{\leq t}](B \sim X)$ denotes that state reward $\rho$ is defined such that $\forall s\in \mathbb{S}.\rho(s) = 1$ iff $B \sim X$ in $s$. The resulting landscape visualisation is shown in Figure \[fig:cs1\_reward\_0-2\]. Here the effect of increase of robustness value with respect to increasing $\gamma_B$ is significantly stronger.
After normalising the robustness values, we can observe that the model is significantly more robust with respect to the cumulative reward-based formulation of the hypothesis. This goes with the fact that the reward property neglects the frequent fluctuations in the given time horizon.
When focusing on the phenomenon of bistability, we can conclude that the most significant variance in the molecule population with respect to the two stable modes is observed in the range $\gamma_A=[0.15,0.3]$ with $\gamma_B=0.10$. Here the distribution of the behaviour targeting the low and high mode is diversified nearly uniformly (especially for $\gamma_A=0.2$). Note that in this case there is a significant amount of behaviour (around $40\%$) not converging to either of the two modes.
![[**Results of robustness analysis for hypothesis (1) using a reward operator.**]{} Hypothesis (1) requires stabilisation of $E_2F_1$ in the low concentration mode ($B<3$). A CSL formula with cumulative reward operator is used in this case. Each of the curves represents the evaluation function over $\gamma_A$ degradation obtained for a particular setting of $\gamma_B$. More precisely, the horizontal axis shows the perturbation of *pRB* degradation rate and the vertical axis shows the probability of the hypothesis to be satisfied. In the upper left corner, robustness values are shown for each of the curves. The values are displayed with the absolute error quantifying the precision of the approximate method. For comparison, the values are computed also on piece-wise affine approximations of the evaluation function. It can be seen that the robustness values change rapidly with different settings of $\gamma_B$. This observation goes with the fact that with faster degradation of $E_2F_1$ there is a higher probability that the positively self-regulated protein is locked in the stable mode of no production. The decrease of the value with increasing $\gamma_A$ is due to the weakening effect of inhibition by *pRB*.[]{data-label="fig:cs1_reward_0-2"}](cs1_reward_0-2){width="\textwidth"}
To encode the hypothesis $2$ we employ the reward-based formulation: $\mathsf{R}_{=?}[\mathsf{C}^{\leq t}](B>7)$. The time interval is set to be the same as in the previous case ($t=1000$). The resulting landscape visualisations for individual settings of $\gamma_B$ are depicted in Figure \[fig:cs1\_reward\_8-10\]. It can be observed that the effect of $\gamma_B$ is now inverse which goes with the fact that higher rate of $E_2F_1$ degradation causes the rapid dynamics of the protein and decreases the amenability of the cell to commit to *S*-phase (by making the hypothesis $1$ more robust than hypothesis $2$).
![[**Results of robustness analysis for hypothesis $2$.**]{} Hypothesis (2) requires stabilisation of $E_2F_1$ in the high concentration mode ($B>7$). A CSL formula with cumulative reward operator is employed. Each of the curves represents the evaluation function over $\gamma_A$ degradation obtained for a particular setting of $\gamma_B$. The horizontal axis shows the perturbation of *pRB* degradation rate and the vertical axis shows the probability of the hypothesis to be satisfied. In the upper left corner, robustness values are shown for each of the curves. The values are displayed with the absolute error quantifying the precision of the approximate method. For comparison, the values are computed also on piece-wise affine approximations of the evaluation function. It can be seen that the robustness values change rapidly with different settings of $\gamma_B$. This observation goes with the fact that with faster degradation of $E_2F_1$ there is a lower probability that the positively self-regulated protein is locked in the stable mode of no production. In particular, the high stable mode is preferred for lower values of $\gamma_B$. The increase of the value with increasing $\gamma_A$ is due to the weakening effect of inhibition by *pRB*.[]{data-label="fig:cs1_reward_8-10"}](cs1_reward_8-10){width="\textwidth"}
An interesting observation coming out of the analysis is that the selection of an initial state has only a negligible impact on the result. This is exploited in Figure \[fig:cs1\_init\_states\_combine\] where we have selected 11 states uniformly distributed throughout the state space. Although low initial numbers of *B* slightly decrease robustness of hypothesis (2), the difference is not very big.
More detailed insight can be inferred from Figure \[fig:cs1\_selected\_interval\_statemap\] where hypothesis (2) evaluation is exploited for a small perturbation of $\gamma_A$ with respect to the entire initial state space. The considered perturbation is highlighted in Figure \[fig:cs1\_init\_states\_combine\] by the grey vertical line. The colour intensity of the grid shows the upper bound of the cumulative reward evaluated for the respective initial state. It can be seen that the hypothesis is really insensitive to selection of initial states. Only the initial zero level of *B* causes a decrease of the resulting value. Moreover, this happens (naturally) just in two kinds of states: (*i*) no molecule of *B* is bound to any of the genes, i.e., self-activation of *b* is inactive and the expression of *b* occurs in the spontaneous mode having a low rate 0.05; (*ii*) a molecule of *A* is bound to *b* thus imposing the inhibition of *b* and causing the same scenario.
![[**Landscape visualisation for hypothesis (2) and several selected initial states.**]{} The landscape visualisation of hypothesis (2) (stabilisation of $E_2F_1$ in the high concentration mode $B>7$) is shown for several selected initial states of the whole state space. A CSL formula with cumulative reward operator is employed. Each of the curves represents the evaluation function over $\gamma_A$ degradation obtained for a particular initial state and $\gamma_B$ set to 0.05. The legend shows the amount of individual species in particular initial states and the robustness of the hypothesis is given together with the absolute error. The results obtained by piece-wise affine approximation are also shown. It can be seen that the hypothesis is only negligibly sensitive to initial conditions. Especially, only states with zero initial concentration of $E_2F_1$ cause $E_2F_1$ to attain low molecular numbers thus lowering the robustness of the hypothesis. The grey vertical line shows the small perturbation in $\gamma_A$ which is further explored in detail in Figure \[fig:cs1\_selected\_interval\_statemap\].[]{data-label="fig:cs1_init_states_combine"}](cs1_init_states_combine){width="\textwidth"}
![[**Analysis of hypothesis (2) for all initial states.**]{} Hypothesis (2) (stabilisation of $E_2F_1$ in the high concentration mode $B>7$) is computed and visualised for all initial states in the considered perturbation space $(\gamma_A,\gamma_B) \in [0.10168, 0.10555] \times [0.05]$. Because we assume at most a single molecule of DNA in the system, state variables denoting genes and gene-protein complexes have a binary domain. There are only two variables having a larger domain (0-10), in particular, these are the proteins *pRB* and $E_2F_1$. Therefore each of the (binary) combinations is visualised for the entire domain of *A* and *B* in a separate box. The colour intensity of each box in the grid shows the upper bound of the cumulative reward evaluated for the respective initial state. It can be seen that the hypothesis is mostly insensitive to selection of initial states. Only the initial zero level of $E_2F_1$ (*B, bB, aB*) causes a decrease of the resulting value. States selected in Figure \[fig:cs1\_init\_states\_combine\] are highlighted in red.[]{data-label="fig:cs1_selected_interval_statemap"}](cs1_selected_interval_statemap){width=".8\textwidth"}
Robustness of two-component signalling systems response
-------------------------------------------------------
Signalling pathways make the main interface between cells and their environment. Their main role is to sense biochemical conditions outside the cell and to transfer this information into the internal logical circuits (gene regulation) of the cell. Since signal processing is realised by several dedicated protein complexes (signalling components), it is naturally amenable to intrinsic noise in these protein populations caused by stochasticity of transcription/translation processes. Robust input-output signal mapping is crucial for cell functionality. Many models and experimental studies have been conducted attempting to explain mechanisms of robust signal processing in procaryotic cells, e.g., [@Batchelor21012003; @Shinar11122007].
In order to construct robust signalling circuits in synthetically modified procaryotic cells, Steuer et al. [@steuer2011robust] has suggested and analysed a modification of a well-studied two-component signalling pathway that is insensitive to signalling component concentration fluctuations. The study has been performed by using a simplified model consisting of the two signalling components each considered in both phosphorylated and unphosporylated forms. The first component, the histidine kinase *H*, is a membrane-bound receptor phosphorylated by an external signalling ligand *S*. In its phosphorylated form *Hp*, the histidine kinase transfers the phospho-group onto the second component – the response regulator *R*. That way it activates the response regulator by transforming it into the phosphorylated form *Rp* which is diffusible and functions as the internal signal for the cell. The basic topology of the pathway is depicted in Figure \[fig:sigtopol\]A. The modification suggested by Steuer et al. is depicted in Figure \[fig:sigtopol\]B. The difference is in the addition of catalytic activation of *Rp* dephosporylation by the unphosphoshorylated histidine kinase *H*. In [@steuer2011robust] it has been rigorously proven that under the deterministic setting this modification leads to globally robust steady-state response of the signalling pathway that is not achievable with the basic topology.
We reformulate the model in the stochastic setting and employ our method to provide detailed analysis of the input-output signal response under fluctuations in population of both signalling components. In contrast to [@steuer2011robust] where average steady-state population is analysed with respect to fluctuations in signalling components, our analysis refines the steady population in terms of distributions. That way we obtain for a stable input signal a detailed view of distribution of the output response. In particular, instead of studying the effect of perturbations on the average population, we see how perturbations affect the distribution, i.e., the variance (fluctuation) in the output response. That way the stochastic framework gives a more detailed insight into the input-output signal response mechanism.
The biochemical model of both topology variants is given in Figure \[fig:sigtopol\]C. The input signal *S* is considered to be fixed and therefore it makes a constant parameter of the model. The signalling components in both phosporylated and unphosporylated forms make the model variables *H, Hp, R*, and *Rp*.
![[**Model of a two-component signalling pathway.**]{} (A) Basic topology of the two-component signalling pathway. (B) Modified topology of the two-component signalling pathway, additionally, histidine kinase *H* catalyses dephosporylation of the response regulator *R*. (C) Reactions specifying the biochemical model of the two considered topologies of the two-component signalling pathway. Phosphorylation of the first component *H* catalysed by the input signal *S* and phosporylation of the second component *R* are shared by both topologies, the only difference is in the second component dephophorylation. Additionally, we consider unregulated proteosynthesis/degradation reactions for both topology variants. Reaction topology in (A) and (B) was created using CellDesigner [@Celldesigner]. []{data-label="fig:sigtopol"}](cs2_model){width="50.00000%"}
Depending on which topology is chosen the original deterministic model [@steuer2011robust] exhibits different relationships between the steady-state concentrations of the input signal *S* and the output signal *Rp*:
$$\begin{aligned}
Rp \text{ steady-state in model 1} \hspace{2cm}& Rp \text{ steady-state in model 2}\\
[Rp]=\frac{k_1}{k_{31}}[S][H] \hspace{3cm} &\hspace{1.5cm} [Rp]=\frac{k_1}{k_{32}}[S]\end{aligned}$$
In particular, it can be seen that the steady-state concentration of the output signal \[*Rp*\] in model 1 is affected not only by the input signal *S* but also by the number of unphosphorylated receptors *R*, this can be interpreted in such a way that the concentration of the signalling components should be kept stable in order to obtain the robust output. This is, however, not an issue in model 2 where *Rp* depends only on *S*. Since the steady-state analysis has been carried out under the deterministic setting additionally imposing assumptions of conserved total amounts of *H + Hp* and *R + Rp*, it is appropriate only for high molecular populations.
The question we want to answer is “Is there a difference in the way the two models handle noise (fluctuations) for low molecular numbers of signalling components?” In such conditions, populations of *H + Hp* and *R + Rp* can not be considered conserved since the proteins are subject to degradation and production. Production of proteins from genes as well as degradation is inherently noisy as it has been demonstrated in the previous case study. Different levels of noise can be affected by, e.g., regulatory feedback loops or varying numbers of gene copies. Even for a noiseless output signal *S* these internal fluctuations of protein concentrations transfer noise to *Rp*. We formalise our question in terms of the CSL property $\mathsf{E}_{=?}[\mathsf{I}^{=t}]$ which asks for the value of a post-processing function in a future time *t*, where the post-processing function is defined as the *mean quadratic deviation* of the distribution of *Rp*.
For the model to have low numbers of molecules to exhibit stochastic fluctuations and enable responses to varying levels of *S* we have chosen $k_{p} = 0.3$ *molecules*$\cdot s^{-1}$ and $k_{d} = 0.01 \mbox{~} s^{-1}$ which leads to an average total population of 30 molecules for both $H + Hp$ and $R + Rp$. To make the analysis straightforward we assume same speed of degradation of phosphorylated and unphosphorylated variants of each protein.
To reduce the size of the state space we have truncated total populations to $25 \leq H + Hp \leq 35$ and $25 \leq R + Rp \leq 35$ which leads to $116281$ states in total. The initial state is considered with populations $s_0 = (H = 30, Hp = 0, R = 30, Rp = 0)$. The state space reduction has a significant impact on the measured absolute values of noise but conserves general trends as is shown in Figure \[fig:cs2\_truncation\_noise\].
In order to control fluctuations in protein production we extend our model with two populations of genes, one for *H* and one for *R*, respectively, and for each of the genes we introduce an autoregulatory negative feedback loop via binding of the proteins to their corresponding genes. That way we restrict the protein production. By modifying the number of gene copies in the cell and the rate of protein-gene binding we are able to regulate the overall noise in the transcription. This approach however leads to rapid increase in state space size because of the necessary introduction of new variables representing genes and protein-gene complexes thus making the analysis inefficient. To this end, we decided to abstract from details of the underlying autoregulatory mechanism and to model it using a sigmoid production function which mimics the desired behaviour accordingly. By numerical analysis, we have verified that such an approximation can be employed in the stochastic framework. The function is defined in the following way: $$\emptyset \stackrel{sig(k_p,n)}{\longrightarrow} X \hspace{4em} sig(k_p,n) = \frac{2}{1 + \left( \frac{X}{30} \right)^n} \cdot k_p$$ where $n$ is the so-called Hill coefficient controlling the steepness of the sigmoid (caused by cooperativity of transcription factors in protein-gene interactions) and $k_P$ is the maximal production rate. We use this approach for modelling the production of both species *H* and *R* by sigmoid coefficients denoted $n_H$ and $n_R$, respectively. The sigmoid function regulates the population by enabling production when it is below average and represses it when the population is above the average. The larger *n* is the more steep the sigmoid function is leading to stronger regulation and lower noise. The case *n=0* corresponds to an unregulated model and when increased to *n=20* it corresponds to over 10 copies of each gene in the fully modelled feedback loop mechanism. The effect of different levels of sigmoid regulation to noise can be seen in a simplified birth death model in Figure \[fig:cs2\_truncation\_noise\].
![[**Influence of state space truncation to mean quadratic deviation of a distribution.**]{} A simple birth death model is considered to show the influence of different settings of the state space truncation on the measured noise evaluated in the form of a *mean quadratic deviation* (*mqd*) of the state space distribution. The model has a single species *X* and two reactions $\emptyset \stackrel{sig(0.3,n)}{\longrightarrow} X, X \stackrel{0.01}{\longrightarrow} \emptyset$ which stabilise the population around an average of 30. For different values of the sigmoid coefficient *n* we can see different *mqd* values, the larger the *n* the smaller the noise. If *X* is restricted to $25\leq X\leq 35$ the overall noise is smaller since the probability mass can not spread to states placed further from the mean. In a less restricted version with populations between 20 and 40 the noise is about $2.5\times$ larger. If sigmoid regulation is weak and the regulation is strong then the difference in the amount of noise is less then 20%.[]{data-label="fig:cs2_truncation_noise"}](cs2_truncation_noise){width="\textwidth"}
To see long term effects of intrinsic noise we decided to examine the system in the situation when the output response is stabilised. Since the min-max approximation method cannot be employed with steady-state computation, transient analysis in a suitable time horizon has been performed instead. To estimate the closest time *t* when the system behaviour can be observed stable, we have computed values of output response noise for the unregulated variant of the model (*n = 0*) using standard numerical steady state numerical analysis (we employed the tool PRISM [@KNP11]) and compare it to probability distributions obtained by transient analysis in $t=20$, $t=50$ and $t=100$ seconds. Consequently, we have compared the probability distribution in the steady state with the probability distribution in $t=100$ seconds. The results clearly show that that the difference in distributions is negligible and the transient distribution can be considered stable after $t=100$.
To further speed up the computation, we have precomputed the distribution of *H* and *R* in the time horizon $t=100$ without enabling phosphorylation reactions. This has lead to a significant reduction to $121$ states. Starting with the achieved probability distribution, we have subsequently computed the transient analysis with enabled phosporylation reactions in next $5$ seconds. The rationale behind is that the protein production and degradation are two orders of magnitude slower than phosphorylation. Therefore total populations of *H* and *R* dictate the time at which the system is nearly stable and thus the next $5$ seconds are sufficient for the fast-scale phosporylation to stabilise the fractions $\frac{H}{Hp}$ and $\frac{R}{Rp}$.
To compute the noise (variance) in *Rp* we employ the *mean quadratic deviation* post-processing function for state space distributions. Our goal is to compare the levels of *Rp* noise in both models for different levels of the output signal *S* and for different values of intrinsic noise appearing in protein production (controlled by sigmoid coefficients $n_H$ and $n_R$). After computing lower and upper bounds of the state space distributions, we have computed the lower and upper bounds of the post-processing function using the algorithm informally introduced in Section \[sec:parameteriseduniformisation\]. Consequently, we obtain robustness values for the output response $R_p$ over the respective perturbation subspaces in the form *average* $\pm$ *error*. Finally, we define the perturbation space of the interest. In particular, for the signal we choose the value interval $S \in [2.0, 20.0]$ and for sigmoid coefficients $n_H, n_R \in [0.1, 10.0]$.
Since the full computation over the 3-dimensional perturbation space has turned out to be intractable, we have to find a way how to reduce its dimension. To this end, we focus on a subspace $S = 15.0, (n_H,n_R) \in [3.0,4.0] \times [3.0,4.0]$ where both models have symmetric sensitivity to both sigmoid production coefficients $n_H, n_R$. This symmetry allows us to merge $n_H, n_R$ into a single coefficient *n*. Results are visualised in Figure \[fig:cs2\_models\_nhnr\] where it can be seen that in Model 1 the influence of $n_H$ and $n_R$ is almost perfectly symmetrical with $n_H$ being slightly more influential. In Model 2 the influence is evidently stronger in $n_R$ but the response seems to be symmetrical enough to justify the sigmoid coefficients merging. An interesting property of parameterised uniformization and the perturbation space decomposition algorithm can be seen in Figure \[fig:cs2\_models\_nhnr\] where the decomposition of the perturbation spaces around both sigmoid coefficients set to $3.1$ is very dense. This is due to the non-linearity of the sigmoid production functions which leads to non-monotonicity of probability inflow/outflow differences in states during parameterised uniformization (see Section \[sec:methods\]). In order to preserve conservativeness of estimates we have to locally over/under approximate these inflow/outflow rates thus leading to increase of error. To obtain the desired level of accuracy, we dynamically refine all those subspaces where this has occurred.
Finally, we inspect selected subintervals of the perturbation space given by five exclusive intervals of the input signal value domain, $S \in [2,3] \cup [6,7] \cup [10,11] \cup [14,15] \cup [19,20]$, and three distinct levels of production noise represented by sigmoid coefficient $n \in \left\{0.1, 4.0, 10.0\right\}$. The results of this main experiment can be seen in Figure \[fig:cs2\_model12\_rp\_noise\] and Figure \[fig:cs2\_model12\_hr\_noise\]. The trends that can be seen in Figure \[fig:cs2\_model12\_rp\_noise\] are that for lower signals up to *S=10*. Model 2 has encountered lower noise in *Rp* than Model 1 but in the higher signal region it is outperformed by Model 1 which quickly converges to values between $8$ and $10$. However, *Rp* noise produced in Model 2 linearly increases with increasing value of the input signal *S*. For most of the inspected subspaces a stronger regulation of *H* and *R* production by the sigmoid coefficient *n* leads to a reduction of *Rp* noise. An exception to this observation can be seen in Model 2 at the signal interval $[19.0, 20.0]$ where this trend is inverted. To show that this is an emergent behaviour arising from the nontrivial interaction of phosphorylation and dephosphorylation reactions not present in the basic production and degradation of components *H* and *R*, their respective influences are displayed in Figure \[fig:cs2\_model12\_hr\_noise\]. There we can see that in Model 1 both *H* and *R* follow an initial increase of noise with increasing *S* but then the noise stabilises. This leads us to a hypothesis that the regulation of noise in signalling components dynamics looses its influence as signal *S* increases. This is however due to the fact that more *S* leads to faster phosphorylation of *H* which effectively reduces the population of *H* thus also reducing its absolute noise. In the case of Model 2 the situation is different since we can observe a permanent increase of noise in both *H* and *R* populations. The inversion of noise with increased regulation seen in \[fig:cs2\_model12\_rp\_noise\] and closely shown in Figure \[fig:cs2\_model2highsignal\] has not yet been explained satisfactorily.
![[**Influence of genetic regulation on noise in model 1 and 2.**]{} In the upper part two schemes noise of *Rp* in model 1 is computed over perturbations of both sigmoid production constants $n_H$ and $n_R$ in $[3.0,4.0] \times [3.0,4.0]$. The upper and lower bounds on noise (mean quadratic deviation of the resulting probability distribution projected onto populations of *Rp*) are recomputed into the form *average* $\pm$ *error*, the average values are shown on the left and errors are shown on the right. The densely subdivided subspaces around the value $3.1$ are due to conservative over/under approximations in the computation of the probability distribution in states where inflow and outflow of the probability mass is not strictly a monotonous function over the given perturbation interval, thus the error is locally increased and the subspaces must be further divided to obtain the required precision. The lower two schemes show the same results for model 2. By comparing both results we can see that model 1 has an overall lower noise and also computation error given the same level of refinement then model 2, in model 1 the results are symmetrical with respect to perturbations in $n_H$ and $n_R$ with $n_H$ having a slightly larger influence. In model 2 $n_R$ has a larger influence, however we considered the difference negligible and combined both parameters into a single sigmoid production constant *n*.[]{data-label="fig:cs2_models_nhnr"}](cs2_model12_nRnH_3-4){width="\textwidth"}
![[**Comparison of models by *Rp* noise robustness.**]{} Robustness *Rp* noise in both models has been computed with respect to perturbations of signal *S* over five selected intervals of the input signal $S \in [2,3] \cup [6,7] \cup [10,11] \cup [14,15] \cup [19,20]$ and for three distinct levels of the intrinsic noise in signalling component dynamics represented by sigmoid coefficient $n \in \left\{0.1, 4.0, 10.0\right\}$. Perturbations were not computed over the whole interval $(S,n) \in [2, 20] \times [0.1, 10.0]$ due to very high computational demands. From the computed values of individual refined subspaces as well as the aggregated robustness values for each input signal interval we can see that for lower values of signal *S* (up-to 10) Model 2 embodies lower output response noise then Model 1 (spontaneous dephosphorylation). While output response noise in Model 1 tends to converge to values between 8 and 10, Model 2 exhibits a permanent (almost linear) increase in the output response noise over most of the studied portion of the perturbation space. A super-linear increase of the noise is observed for strong input signals. Another interesting aspect is that while with increasing levels of gene regulation given by sigmoid coefficient *n* the overall noise in *Rp* decreases over the whole interval of signal values for Model 1 and most of it for Model 2, there is an anomaly in Model 2 in the high signal region \[19.0, 20.0\] where with decreasing noise in *R* and *H* (see Figure \[fig:cs2\_model12\_hr\_noise\]) the noise in *Rp* increases. We have not yet explained this phenomenon satisfactorily.[]{data-label="fig:cs2_model12_rp_noise"}](cs2_models_combine_all_kn){width="\textwidth"}
![[**Noise in populations or *H* and *R* in both models.**]{} Noise in *H* (A) and *R* (B) in both models has been computed with respect to perturbations of signal *S* over five selected intervals $S \in [2,3] \cup [6,7] \cup [10,11] \cup [14,15] \cup [19,20]$ and for three distinct levels of inherent production noise represented by sigmoid coefficient $n \in \left\{0.1, 4.0, 10.0\right\}$. We can see that in all cases with increasing regulation by *n* the intrinsic noise in the dynamics of each of the signalling components decreases.[]{data-label="fig:cs2_model12_hr_noise"}](cs2_hr_noise_combine){width="\textwidth"}
![[**High signal region in model 2.**]{} A closeup of the high signal region in model 2, where increasing levels of regulation by the sigmoid coefficient *n* leads to a paradoxical increase of output response noise instead of decrease. Even though the inaccuracy is large we consider the trend to be strong and thus real.[]{data-label="fig:cs2_model2highsignal"}](cs2_model_2_all_kN_19-20_combine){width="\textwidth"}
Discussion
==========
In this paper we proposed a novel framework for robustness analysis of stochastic biochemical systems. It allows us to quantify and analyse how the validity of a hypothesis formulated as a temporal property depends on the perturbations of stochastic kinetic parameters and initial concentrations. The framework extends the quantitative model checking techniques and numerical methods for CTMCs and adapts them to the needs of stochastic modelling in biology. Therefore, in contrast to statistical methods such as Monte Carlo simulation and parameter sampling our framework is customizable with respect to the required precision of computation. This is obtained by providing the lower and upper bounds of the results.
Case studies have demonstrated that the framework can be successfully applied to the robustness analysis of nontrivial biochemical systems. They have shown how to use CSL to specify properties targeting transient behaviour under fluctuations. From the first case study we can conclude that the reward-based formulation of stability properties is more appropriate to distinguish the individual parameter settings under the requested range of uncertainty. The inspected biological hypothesis in the second case study can not be directly formulated using CSL with rewards. Therefore, we have employed post-processing functions to express and study the mean quadratic deviation of the molecule population distribution of the signal response regulator protein.
The time complexity of our framework in practice depends mainly on the size of the state space, the number of reaction steps that have to be considered, and the number of perturbation sets that have to be analysed to provide the desired precision. The size of the state space is given by the number of species and their populations. The framework is suitable for low populations and is relevant especially in the case of gene regulation. In the first case study we have considered only a single molecule of DNA and thus the state space of resulting CTMC was manageable. In the second case study we had to abstract from the feedback loop mechanism using a sigmoid production function to reduce the state space and to make the analysis feasible. If such an abstraction can not be used, our framework can be effectively combined with general state space reduction methods for CTMCs, e.g., finite projection techniques [@munsky2006finite; @Henzinger2009] and dynamic state space truncation [@Didier]. The number of reaction steps can be reduced using separation of fast and slow reactions as demonstrated in the second case study or using adaptive uniformisation [@Moorsel94; @Didier].
In the first case study several hundreds of perturbation subsets had to be analysed and the overall robustness analysis took a few hours. However, in the second case study several thousands of perturbation subsets were required to achieve reasonable precision. In order to speedup the computation we analysed the subsets in parallel using a high performance multi-core workstation were the analysis took several hours. To further improve the accuracy of the robustness analysis without decreasing the performance, we have employed a piecewise linear approximation. It allows us to obtain more precise result without increasing the number of perturbation sets, however, it does not guarantee the conservative error bounds.
The presented method as employed in the first case study gives us a tool for exact analysis of bistability from the global point of view (with respect to all initial conditions, the considered time bound, and the given range of parameters). It can be considered as an analogy to bifurcation analysis known from the ODE world. When comparing our approach with the bifurcation analysis performed in [@Swatetal04], our approach provides a detailed mesoscopic insight into the analysed phenomenon. Instead of identifying just the points where the population diverges, we obtain the precise knowledge of how the population is distributed around the two stable states. Especially, the method shows that reachability of the cancer-inducing high stable mode of the retinoblastoma-binding transcription factor is almost always possible despite the initial state of the regulatory system. The exhaustive analysis is performed with uncertainty in the degradation parameters of the two most important cell-cycle regulating proteins. However, if the degradation of the tumour suppressor protein is sufficiently high, there is always possibility allowing the population to switch into the safe low stable mode. Moreover, robustness of having the possibility to avoid the cell malfunction is positively affected by increasing the retinoblastoma-binding transcription factor degradation. In contrast to [@Swatetal04], the switching mechanism is described at the single cell level which allows to quantify the portion of population amenable to mall-function and thus can provide a preliminary guide to further analysis targeting elimination of the undesired behaviour.
The second case study has shown new insights into the phenomenon of noise in two-component signalling pathways appearing in procaryotic organisms. The previous study [@steuer2011robust] conducted in the framework of deterministic models targeted global robustness of steady concentrations of output signalling components by means of analytically finding invariant perturbation space. The result has shown that a synthetic pathway topology including additional catalysis of signal response regulator by histidine kinase leads to globally robust input-output signal mapping with respect to fluctuations in signalling components concentration. On the contrary, the basic topology without histidine-modulated dephosphorylation does not fulfil global robustness. Since signalling pathways are understood to be amenable to intrinsic noise due to relatively low molecule populations of signalling proteins (typically hundreds of molecules), the respective stochasticity might affect the input-output signal response. To this end, we have reformulated the model in the stochastic framework and instead of studying the effect of perturbations on the average population, we study in detail how perturbations affect the distribution, i.e., the variance (fluctuation) in the output response. Our study has shown that both pathway topologies result with fluctuations in output response, but robustness of input-output mapping varies in both models with increasing the level of the (constant) input signal. For low input signals the synthetic topology gives response with smaller variance in the output whereas for high input signals the output variance rapidly increases. Therefore the basic topology seems to be more suitable for processing of strong signals while the synthetic topology is more appropriate for low level signals. Our study has also shown that both topologies are quite robust with respect to scaling the noise in signalling components dynamics.
Acknowledgments {#acknowledgments .unnumbered}
===============
This work has been supported by the Czech Science Foundation grant No. GAP202/11/0312. M. Češka has been supported by Ministry of Education, Youth, and Sport project No. CZ.1.07/2.3.00/30.0009 - Employment of Newly Graduated Doctors of Science for Scientific Excellence. D. Šafránek has been supported by EC OP project No. CZ.1.07/2.3.00/20.0256.
|
Just another WordPress.com site
Posts tagged ‘therapy’
Today I started my first ever “detox diet.” I always thought detox diets were for people trying to flush drugs out of their system before a drug test. Don’t psychoanalyze where I got that idea. With all of my stomach pain, someone suggested I try to cleanse my system by doing one of these detoxes to start over the probiotics in my stomach. I asked my doctor about it, he seemed cynical. He also said if this will make me more mindful of eating healthy and drinking lots of fluids for 7 days, then he supports me trying it. Today I had a hard time passing up the desserts. In fact, I feel like I’ve been hungry all day. Fruits, veggies, and occasional whole grains or protein just don’t fill me up like the fattening food I love. I’m hoping it will get easier as the week goes on, but I’m still wondering how I’m going to make it a full week without chocolate.
Tomorrow I’m going to a new therapist. My brother, who is getting his PhD in Clinical Psych, was kind enough to research the resources in my area and even called the woman to make sure she takes my insurance. Going to new therapists is always intimidating, but I’ve gotten to be a pro at it over the last few years. Since I started seeing a therapist, five or six years ago, I’ve only had one long-term therapist. Having a long-term therapist makes a huge difference. I was lucky enough to have one see me directly following my hospitalizations all the way through to graduating college. Therapy is such a long, draining, and sometimes downright upsetting process, but all the progress I made in those couple years blows my mind. My quality of life is much better because of my work in therapy, so I just have to go into my appointment tomorrow with an open mind and the desire to improve myself. It’s interesting that admitting to going to therapy is so taboo. I try to be open about it, but I sometimes forget it’s still not talked about in some places. For instance, in a group of small town folks, without thinking I said, “My therapist says…” and then I realized I’d just made everyone uncomfortable by admitting I have a therapist. I, personally, think everyone should have a therapist. What better way is there to self-reflect and develop self-awareness? You get to talk for one hour straight while someone listens attentively, provides alternative ways to view situations, and always supports you and reminds you to be nice to yourself.
My dad seemed jealous of my self-assurance the other day. I was packing a bunch of my painting supplies, and it was a lot of crap to hall back with me to the convent. I didn’t think it was a big deal, because one nun had suggested I bring my painting stuff, not to mention, I have my own room where I can store it, so I don’t know why anyone would care.
My dad said, “Are you really taking all that shit?”
“Yeah, why?” I said
“Is it your meds that gives you the confidence to do that, or therapy, or what?” he asked.
This response shocked me for several reasons. The first and foremost: I’m far from being a confident person, and did he really think self-assurance can come in a pill? My parents also ask me if I’ve taken my meds if I’m upset about something. People struggle to understand that anti-depressants aren’t really happy pills. They don’t protect you from experiencing all of the emotions that life brings. There’s a quote I like, although I can’t remember the author. It’s something along the line of, “The opposite of depression isn’t happiness. It’s human vitality.” The medication just balances out the chemicals to take care of the physical symptoms keeping you from having the energy to do things that help you feel better. I don’t think people understand that mental illness can be physically paralyzing. During my second hospitalization, I could not get out of bed, and I honestly didn’t care if I got better. I didn’t have the energy to fight anymore.
Thinking about how far I’ve come, I try not to get discouraged that I just had to get my meds increased. I’m trying to remind myself that brain chemistry changes and that going on more medication is not the same as getting worse. It’s actually me fighting and my refusal to go back to where I was. I really don’t feel bad mentally yet, but it’s interesting how our bodies can sometimes tell us things we’re not yet conscious of. Maybe this physical detox will help detox my mind.
|
A powered toothbrush is designed to assist a user by mechanically moving the head of the toothbrush. One approach in the art is to provide a vibratory element in the body of the toothbrush. U.S. Published Application No. 2002/0124333 relates to a mechanical vibratory device which causes the head part to vibrate. The vibratory device is accommodated in a front head part of the toothbrush, or in a neck-part region adjacent to the head part, said neck part connecting the head part to the handle, and is operatively connected to a power source, accommodated in the handle, via electrical connections running in the neck part, vibration-damping means preferably being provided in order to prevent vibration transmission to the handle, this achieves the situation where the vibrations which effect the improved cleaning action are produced predominantly in the head part and can only be felt to a slight extent in the handle, as a result of which comfortable handling of the toothbrush is achieved.
A number of approaches have been taken in the prior art to provide flexibility to the bristles during use of a toothbrush. U.S. Pat. No. 5,970,564, for example, discloses a toothbrush having an elastomeric ridge wherein there is a center array of bristles and there is a side array of bristles mounted in elastomeric boots. A number of patents disclose a toothbrush head having sets of bristles, each of which is mounted to a non-rigid or elastic support element. Examples of these approaches are found in U.S. Pat. Nos. 1,770,195, 2,244,098, 6,161,245 and 6,311,360 and in French Patent No. 38440.
The head of a conventional toothbrush usually has a flat or slightly altered surface to which cleaning elements are attached. Usually the cleaning elements are strands of plastic material(s) formed into tufts, bundles or other groupings. A goal of many toothbrushes is to accommodate the cleaning element profile to that of the teeth being cleaned. Achieving that goal is complicated by the difficulty in matching a toothbrush profile to the complex surface of a typical set of human teeth. The latter generally lie in a “C” shaped curve which presents the need for a brush to address a convex outer curve and a concave inner curve. In addition, the toothbrush should be capable of cleaning irregularities on the tooth surface as well as the interproximal area between teeth.
Blakeman U.S. Pat. No. 2,706,825 issued Apr. 26, 1955 discloses a replaceable bristle head for a toothbrush. The flexible head undulates in a manner so that rows of bristles move in a direction aligned with the axis of the toothbrush handle. U.S. Pat. Nos. 5,355,546, 5,483,722 and 5,839,149, all issued to Scheirer et al disclose cleaning elements mounted on a flexible membrane supported between a horseshoe shaped handle extension.
U.S. Pat. No. 5,651,158 issued Jul. 29, 1997 to Hans Halm discloses a toothbrush handle with a segmented head wherein adjoining segments are linked by an elastomeric material. The segments are primarily oriented transverse to the longitudinal axis of the toothbrush but may also be oriented parallel to that axis.
U.S. Pat. No. 6,088,870 discloses a latticework arrangement across the face of the toothbrush head. Tufts of cleaning elements are situated between the boundaries of the lattice and are mounted in the head so that each tuft flexes such that during brushing the bristle tufts will deflect in a manner that increases cleaning of the tooth surface.
U.S. Pat. No. 6,219,874 B1 issued Apr. 24, 2001 to Gelder, et al. discloses flexible mounting of toothbrush cleaning elements accomplished by segmenting portions of the toothbrush head, which segments are connected by flexible hinges.
U.S. Pat. No. 6,408,476 discloses another form of segmented toothbrush head with transverse grooves and an elastomeric portion joining the segments. A method of manufacturing this head is also disclosed.
It is well known that the ideal brushing technique from a dental hygiene perspective is an up and down stroke along the vertical surface of teeth which massages the gums while cleaning the teeth. However, due to a number of factors, including ergonomic difficulties, haste, lack of education or the like, few consumers use the recommended brushing technique. Rather, the typical consumer brushes across their teeth in a horizontal motion rather than a vertical movement. Various approaches have been taken by others to translate horizontal brush movement into partial vertical movement of the bristles or cleaning elements.
Translation of horizontal to vertical movement of cleaning elements is accomplished in U.S. Pat. No. 4,783,869 through use of a helix groove in a movable shaft within a toothbrush handle. The groove receives a pin which rides in the groove. This mechanism causes the toothbrush head to partially rotate or oscillate as the handle moves left-to-right or vice versa in the user's mouth. That rotation or oscillation causes the cleaning elements to move in a vertical plane perpendicular to movement of the toothbrush handle.
U.S. Pat. No. 5,481,775 discloses an arcuate shaped base for a toothbrush head aligned with the longitudinal axis of the head. A movable arcuate block containing cleaning elements is flexibly mounted on the toothbrush head. The block is free to slide on the head in a manner whereby the cleaning elements may travel in a vertical direction generally transverse to the typical side-to-side motion of the toothbrush.
U.S. Pat. No. 5,528,786 discloses pivotal mounting of cleaning elements that allows those elements to move up and down in concert with a side-to-side stroke along the teeth.
A general disclosure of flexible mounting for cleaning elements on a toothbrush head is contained in U.S. Pat. No. 5,839,149. In this patent the cleaning elements are mounted on a flexible membrane supported between a horseshoe shaped handle extension.
U.S. Pat. No. 6,141,817 discloses cleaning elements mounted on a flexible membrane that splay outward when the toothbrush is pressed against the user's teeth.
U.S. Pat. No. 6,338,176 B1 issued Jan. 15, 2002 to Smith, et al. discloses round sections of cleaning bristles mounted on individual pads that rotate within a toothbrush body. This converts backward and forward motion of the toothbrush into circular motion of the cleaning elements. The bristles associated with each pad are of varying height to accommodate irregularities, gaps, pockets and contours in natural tooth formation. The rotating cleaning elements can be supplemented with fixed cleaning elements adjacent thereto.
A toothbrush head should provide both proper support for the bristles, and be flexible enough during use to allow the bristles to conform to the shape of a user's mouth or teeth. Additionally, construction techniques should be inexpensive, versatile and consistent.
In an attempt to meet these criteria, a process known as “Anchor Free Tufting” (“AFT”) has been used in the formation of toothbrush heads. In such an AFT process, a head plate for holding toothbrush bristles, and for eventual insertion into a toothbrush body, is typically formed of a rigid plastic that is conducive to sonic welding. The head plate is formed with a solid perimeter and defines a field of variously shaped and sized holes within this perimeter. Fibers that are to form the tufts are then placed in the holes in the field of the head plate, and the backs of the tufts are melted together to fix their position relative to one another.
The tufted head plate is then inserted into a predefined receiving portion of the head portion of a toothbrush handle and is sonically welded into place. The brush is then end-rounded and packaged for sale as a traditional toothbrush.
However, this manufacturing process results in a toothbrush with a very rigid head that does not easily conform to the physical characteristics of a user when brushing. Therefore, it would be desirable to provide a toothbrush that can be conveniently manufactured by the above process, but provides the desired flexibility of the head of the toothbrush during use.
It would also be desirable if a toothbrush could be provided having various cleaning elements on a head, such as bristles with different degrees of flexible mounting, to have a enhanced cleaning effect when moved by a mechanical vibratory element.
|
More Headlines
Posted: 10/14/2010
- The Justice Department asked a federal judge Thursday to allow the military's "don't ask, don't tell' policy to continue while it appeals her order to end the ban on gays serving openly.
Updated: 10/14/2010
- A school bus and vehicle have collided in Pendleton County. While no students on the bus were injured, there are reports a teenager in the vehicle had to be air-lifted from the scene.
Posted: 10/14/2010
- A federal court has refused to reconsider the case of a Kentucky death row inmate. But the one judge who didn't agree with the decision said he thinks the state's legal system is unreliable when it comes to determining if someone should die for a crime.
Updated: 10/13/2010
- The last of the Chilean miners, the foreman who held them together when they were feared lost, was raised from the depths of the earth Wednesday night - a joyous ending to a 69-day ordeal that riveted the world.
Updated: 10/14/2010
- Lenders seized more U.S. homes this summer than in any three-month stretch since the housing market began to collapse in 2006, but many of the foreclosures may soon be challenged in court, because of allegations that banks evicted people without reading the documents.
Updated: 10/13/2010
- A Central Kentucky town is dealing with a foul problem after an odor has started coming from trees.
This is happening to some trees in the downtown area, and the smell is becoming a nuisance to some area shop owners.
Updated: 10/13/2010
- First it was bed bugs. Now stink bugs! Experts say the hot, dry weather has proven to be ideal conditions for the smelly pests to thrive on crops. But problems with them may become even bigger than those with bed bugs.
Online Public Information File
Viewers with disabilities can get assistance accessing this station's FCC Public Inspection File by contacting the station with the information listed below. Questions or concerns relating to the accessibility of the FCC's online public file system should be directed to the FCC at 888-225-5322, 888-835-5322 (TTY), or [email protected].
|
Compile and install
===================
Dependencies
------------
`sslh` uses:
* [libconfig](http://www.hyperrealm.com/libconfig/). For
Debian this is contained in package `libconfig8-dev`. You
can compile with or without it using USELIBCONFIG in the
Makefile.
* [libwrap](http://packages.debian.org/source/unstable/tcp-wrappers).
For Debian, this is contained in packages
`libwrap0-dev`. You
can compile with or without it using USELIBWRAP in the
Makefile.
* [libsystemd](http://packages.debian.org/source/unstable/libsystemd-dev), in package `libsystemd-dev`. You
can compile with or without it using USESYSTEMD in the
Makefile.
* [libcap](http://packages.debian.org/source/unstable/libcap-dev), in package `libcap-dev`. You can compile with or without it using USELIBCAP in the Makefile
* libbsd, to enable to change the process name (as shown in
`ps`, so each forked process shows what protocol and what
connection it is serving),
which requires `libbsd` at runtime, and `libbsd-dev` at
compile-time.
For OpenSUSE, these are contained in packages libconfig9 and
libconfig-dev in repository
<http://download.opensuse.org/repositories/multimedia:/libs/openSUSE_12.1/>
For Fedora, you'll need packages `libconfig` and
`libconfig-devel`:
yum install libconfig libconfig-devel
If you want to rebuild `sslh-conf.c` (after a `make
distclean` for example), you will also need to add
[conf2struct](https://www.rutschle.net/tech/conf2struct/README.html)
(v1.4) to your path.
Compilation
-----------
After this, the Makefile should work:
make install
There are a couple of configuration options at the beginning
of the Makefile:
* `USELIBWRAP` compiles support for host access control (see
`hosts_access(3)`), you will need `libwrap` headers and
library to compile (`libwrap0-dev` in Debian).
* `USELIBCONFIG` compiles support for the configuration
file. You will need `libconfig` headers to compile
(`libconfig8-dev` in Debian).
* `USESYSTEMD` compiles support for using systemd socket activation.
You will need `systemd` headers to compile (`systemd-devel` in Fedora).
* `USELIBBSD` compiles support for updating the process name (as shown
by `ps`).
Binaries
--------
The Makefile produces two different executables: `sslh-fork`
and `sslh-select`:
* `sslh-fork` forks a new process for each incoming connection.
It is well-tested and very reliable, but incurs the overhead
of many processes.
If you are going to use `sslh` for a "small" setup (less than
a dozen ssh connections and a low-traffic https server) then
`sslh-fork` is probably more suited for you.
* `sslh-select` uses only one thread, which monitors all connections
at once. It is more recent and less tested, but only incurs a 16
byte overhead per connection. Also, if it stops, you'll lose all
connections, which means you can't upgrade it remotely.
If you are going to use `sslh` on a "medium" setup (a few thousand ssh
connections, and another few thousand ssl connections),
`sslh-select` will be better.
If you have a very large site (tens of thousands of connections),
you'll need a vapourware version that would use libevent or
something like that.
Installation
------------
* In general:
make
cp sslh-fork /usr/local/sbin/sslh
cp basic.cfg /etc/sslh.cfg
vi /etc/sslh.cfg
* For Debian:
cp scripts/etc.init.d.sslh /etc/init.d/sslh
* For CentOS:
cp scripts/etc.rc.d.init.d.sslh.centos /etc/rc.d/init.d/sslh
You might need to create links in /etc/rc<x>.d so that the server
start automatically at boot-up, e.g. under Debian:
update-rc.d sslh defaults
|
526 U.S. 603 (1999)
WILSON et al.
v.
LAYNE, DEPUTY UNITED STATES MARSHAL, et al.
No. 98-83.
United States Supreme Court.
Argued March 24, 1999.
Decided May 24, 1999.
CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT
*604 *605 Rehnquist, C. J., delivered the opinion for a unanimous Court with respect to Parts I and II, and the opinion of the Court with respect to Part III, in which O'Connor, Scalia, Kennedy, Souter, Thomas, Ginsburg, and Breyer, JJ., joined. Stevens, J., filed an opinion concurring in part and dissenting in part, post, p. 618.
Richard K. Willard argued the cause for petitioners. With him on the briefs were David H. Coburn, James S. Felt, Richard Seligman, Steven R. Shapiro, Arthur B. Spitzer, and Dwight H. Sullivan.
Lawrence P. Fletcher-Hill, Assistant Attorney General of Maryland, argued the cause for the state respondents. With him on the brief were J. Joseph Curran, Jr., Attorney General, Carmen M. Shepard, Deputy Attorney General, and Andrew H. Baida and John B. Howard, Jr., Assistant Attorneys General. Richard A. Cordray filed a brief for the federal respondents.[*]
Chief Justice Rehnquist delivered the opinion of the Court.
While executing an arrest warrant in a private home, police officers invited representatives of the media to accompany them. We hold that such a "media ride-along" does violate the Fourth Amendment, but that because the state *606 of the law was not clearly established at the time the search in this case took place, the officers are entitled to the defense of qualified immunity.
I
In early 1992, the Attorney General of the United States approved "Operation Gunsmoke," a special national fugitive apprehension program in which United States Marshals worked with state and local police to apprehend dangerous criminals. The "Operation Gunsmoke" policy statement explained that the operation was to concentrate on "armed individuals wanted on federal and/or state and local warrants for serious drug and other violent felonies." App. 15. This effective program ultimately resulted in over 3,000 arrests in 40 metropolitan areas. Brief for Federal Respondents Layne et al. 2.
One of the dangerous fugitives identified as a target of "Operation Gunsmoke" was Dominic Wilson, the son of petitioners Charles and Geraldine Wilson. Dominic Wilson had violated his probation on previous felony charges of robbery, theft, and assault with intent to rob, and the police computer listed "caution indicators" that he was likely to be armed, to resist arrest, and to "assaul[t] police." App. 40. The computer also listed his address as 909 North StoneStreet Avenue in Rockville, Maryland. Unknown to the police, this was actually the home of petitioners, Dominic Wilson's parents. Thus, in April 1992, the Circuit Court for Montgomery County issued three arrest warrants for Dominic Wilson, one for each of his probation violations. The warrants were each addressed to "any duly authorized peace officer," and commanded such officers to arrest him and bring him "immediately" before the Circuit Court to answer an indictment as to his probation violation. The warrants made no mention of media presence or assistance.[1]
*607 In the early morning hours of April 16, 1992, a Gunsmoke team of Deputy United States Marshals and Montgomery County Police officers assembled to execute the Dominic Wilson warrants. The team was accompanied by a reporter and a photographer from the Washington Post, who had been invited by the Marshals to accompany them on their mission as part of a Marshals Service ride-along policy.
At around 6:45 a.m., the officers, with media representatives in tow, entered the dwelling at 909 North StoneStreet Avenue in the Lincoln Park neighborhood of Rockville. Petitioners Charles and Geraldine Wilson were still in bed when they heard the officers enter the home. Petitioner Charles Wilson, dressed only in a pair of briefs, ran into the living room to investigate. Discovering at least five men in street clothes with guns in his living room, he angrily demanded that they state their business, and repeatedly cursed the officers. Believing him to be an angry Dominic Wilson, the officers quickly subdued him on the floor. Geraldine Wilson next entered the living room to investigate, wearing only a nightgown. She observed her husband being restrained by the armed officers.
When their protective sweep was completed, the officers learned that Dominic Wilson was not in the house, and they departed. During the time that the officers were in the home, the Washington Post photographer took numerous pictures. The print reporter was also apparently in the living room observing the confrontation between the police and *608 Charles Wilson. At no time, however, were the reporters involved in the execution of the arrest warrant. Brief for Federal Respondents Layne et al. 4. The Washington Post never published its photographs of the incident.
Petitioners sued the law enforcement officials in their personal capacities for money damages under Bivens v. Six Unknown Fed. Narcotics Agents, 403 U. S. 388 (1971) (the U. S. Marshals Service respondents), and Rev. Stat. § 1979, 42 U. S. C. § 1983 (the Montgomery County Sheriff's Department respondents). They contended that the officers' actions in bringing members of the media to observe and record the attempted execution of the arrest warrant violated their Fourth Amendment rights. The District Court denied respondents' motion for summary judgment on the basis of qualified immunity.
On interlocutory appeal to the Court of Appeals, a divided panel reversed and held that respondents were entitled to qualified immunity. The case was twice reheard en banc, where a divided Court of Appeals again upheld the defense of qualified immunity. The Court of Appeals declined to decide whether the actions of the police violated the Fourth Amendment. It concluded instead that because no court had held (at the time of the search) that media presence during a police entry into a residence violated the Fourth Amendment, the right allegedly violated by respondents was not "clearly established" and thus qualified immunity was proper. 141 F. 3d 111 (CA4 1998). Five judges dissented, arguing that the officers' actions did violate the Fourth Amendment, and that the clearly established protections of the Fourth Amendment were violated in this case. Id., at 119 (opinion of Murnaghan, J.)
Recognizing a split among the Circuits on this issue, we granted certiorari in this case and another raising the same question, Hanlon v. Berger, 525 U. S. 981 (1998), and now affirm the Court of Appeals, although by different reasoning.
*609 II
Petitioners sued the federal officials under Bivens and the state officials under § 1983. Both Bivens and § 1983 allow a plaintiff to seek money damages from government officials who have violated his Fourth Amendment rights. See § 1983; Bivens, supra, at 397. But government officials performing discretionary functions generally are granted a qualified immunity and are "shielded from liability for civil damages insofar as their conduct does not violate clearly established statutory or constitutional rights of which a reasonable person would have known." Harlow v. Fitzgerald, 457 U. S. 800, 818 (1982).
Although this case involves suits under both § 1983 and Bivens, the qualified immunity analysis is identical under either cause of action. See, e. g., Graham v. Connor, 490 U. S. 386, 394, n. 9 (1989); Malley v. Briggs, 475 U. S. 335, 340, n. 2 (1986). A court evaluating a claim of qualified immunity "must first determine whether the plaintiff has alleged the deprivation of an actual constitutional right at all, and if so, proceed to determine whether that right was clearly established at the time of the alleged violation." Conn v. Gabbert, ante, at 290. This order of procedure is designed to "spare a defendant not only unwarranted liability, but unwarranted demands customarily imposed upon those defending a long drawn out lawsuit." Siegert v. Gilley, 500 U. S. 226, 232 (1991). Deciding the constitutional question before addressing the qualified immunity question also promotes clarity in the legal standards for official conduct, to the benefit of both the officers and the general public. See County of Sacramento v. Lewis, 523 U. S. 833, 840-842, n. 5 (1998). We now turn to the Fourth Amendment question.
In 1604, an English court made the now-famous observation that "the house of every one is to him as his castle and fortress, as well for his defence against injury and violence, as for his repose." Semayne's Case, 5 Co. Rep. 91a, 91b, 77 *610 Eng. Rep. 194, 195 (K. B.). In his Commentaries on the Laws of England, William Blackstone noted that
"the law of England has so particular and tender a regard to the immunity of a man's house, that it stiles it his castle, and will never suffer it to be violated with impunity: agreeing herein with the sentiments of antient Rome . . . . For this reason no doors can in general be broken open to execute any civil process; though, in criminal causes, the public safety supersedes the private." 4 Commentaries 223 (1765-1769).
The Fourth Amendment embodies this centuries-old principle of respect for the privacy of the home: "The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized." U. S. Const., Amdt. 4 (emphasis added). See also United States v. United States Dist. Court for Eastern Dist. of Mich., 407 U. S. 297, 313 (1972) ("[P]hysical entry of the home is the chief evil against which the wording of the Fourth Amendment is directed").
Our decisions have applied these basic principles of the Fourth Amendment to situations, like the one in this case, in which police enter a home under the authority of an arrest warrant in order to take into custody the suspect named in the warrant. In Payton v. New York, 445 U. S. 573, 602 (1980), we noted that although clear in its protection of the home, the common-law tradition at the time of the drafting of the Fourth Amendment was ambivalent on the question whether police could enter a home without a warrant. We were ultimately persuaded that the "overriding respect for the sanctity of the home that has been embedded in our traditions since the origins of the Republic" meant that absent a warrant or exigent circumstances, police could not *611 enter a home to make an arrest. Id., at 601, 603-604. We decided that "an arrest warrant founded on probable cause implicitly carries with it the limited authority to enter a dwelling in which the suspect lives when there is reason to believe the suspect is within." Id., at 603.
Here, of course, the officers had such a warrant, and they were undoubtedly entitled to enter the Wilson home in order to execute the arrest warrant for Dominic Wilson. But it does not necessarily follow that they were entitled to bring a newspaper reporter and a photographer with them. In Horton v. California, 496 U. S. 128, 140 (1990), we held "[i]f the scope of the search exceeds that permitted by the terms of a validly issued warrant or the character of the relevant exception from the warrant requirement, the subsequent seizure is unconstitutional without more." While this does not mean that every police action while inside a home must be explicitly authorized by the text of the warrant, see Michigan v. Summers, 452 U. S. 692, 705 (1981) (Fourth Amendment allows temporary detainer of homeowner while police search the home pursuant to warrant), the Fourth Amendment does require that police actions in execution of a warrant be related to the objectives of the authorized intrusion, see Arizona v. Hicks, 480 U. S. 321, 325 (1987). See also Maryland v. Garrison, 480 U. S. 79, 87 (1987) ("[T]he purposes justifying a police search strictly limit the permissible extent of the search").
Certainly the presence of reporters inside the home was not related to the objectives of the authorized intrusion. Respondents concede that the reporters did not engage in the execution of the warrant, and did not assist the police in their task. The reporters therefore were not present for any reason related to the justification for police entry into the homethe apprehension of Dominic Wilson.
This is not a case in which the presence of the third parties directly aided in the execution of the warrant. Where the police enter a home under the authority of a warrant to *612 search for stolen property, the presence of third parties for the purpose of identifying the stolen property has long been approved by this Court and our common-law tradition. See, e. g., Entick v. Carrington, 19 How. St. Tr. 1029, 1067 (K. B. 1765) (in search for stolen goods case, "`[t]he owner must swear that the goods are lodged in such a place. He must attend at the execution of the warrant to shew them to the officer, who must see that they answer the description") (quoted with approval in Boyd v. United States, 116 U. S. 616, 628 (1886)).
Respondents argue that the presence of the Washington Post reporters in the Wilsons' home nonetheless served a number of legitimate law enforcement purposes. They first assert that officers should be able to exercise reasonable discretion about when it would "further their law enforcement mission to permit members of the news media to accompany them in executing a warrant." Brief for Federal Respondents Layne et al. 15. But this claim ignores the importance of the right of residential privacy at the core of the Fourth Amendment. It may well be that media ride-alongs further the law enforcement objectives of the police in a general sense, but that is not the same as furthering the purposes of the search. Were such generalized "law enforcement objectives" themselves sufficient to trump the Fourth Amendment, the protections guaranteed by that Amendment's text would be significantly watered down.
Respondents next argue that the presence of third parties could serve the law enforcement purpose of publicizing the government's efforts to combat crime, and facilitate accurate reporting on law enforcement activities. There is certainly language in our opinions interpreting the First Amendment which points to the importance of "the press" in informing the general public about the administration of criminal justice. In Cox Broadcasting Corp. v. Cohn, 420 U. S. 469, 491 492 (1975), for example, we said "in a society in which each individual has but limited time and resources with which to *613 observe at first hand the operations of his government, he relies necessarily upon the press to bring to him in convenient form the facts of those operations." See also Richmond Newspapers, Inc. v. Virginia, 448 U. S. 555, 572-573 (1980). No one could gainsay the truth of these observations, or the importance of the First Amendment in protecting press freedom from abridgment by the government. But the Fourth Amendment also protects a very important right, and in the present case it is in terms of that right that the media ridealongs must be judged.
Surely the possibility of good public relations for the police is simply not enough, standing alone, to justify the ride-along intrusion into a private home. And even the need for accurate reporting on police issues in general bears no direct relation to the constitutional justification for the police intrusion into a home in order to execute a felony arrest warrant.
Finally, respondents argue that the presence of third parties could serve in some situations to minimize police abuses and protect suspects, and also to protect the safety of the officers. While it might be reasonable for police officers to themselves videotape home entries as part of a "quality control" effort to ensure that the rights of homeowners are being respected, or even to preserve evidence, cf. Ohio v. Robinette, 519 U. S. 33, 35 (1996) (noting the use of a "mounted video camera" to record the details of a routine traffic stop), such a situation is significantly different from the media presence in this case. The Washington Post reporters in the Wilsons' home were working on a story for their own purposes. They were not present for the purpose of protecting the officers, much less the Wilsons. A private photographer was acting for private purposes, as evidenced in part by the fact that the newspaper and not the police retained the photographs. Thus, although the presence of third parties during the execution of a warrant may in some circumstances be constitutionally permissible, see supra, at 611-612, the presence of these third parties was not.
*614 The reasons advanced by respondents, taken in their entirety, fall short of justifying the presence of media inside a home. We hold that it is a violation of the Fourth Amendment for police to bring members of the media or other third parties into a home during the execution of a warrant when the presence of the third parties in the home was not in aid of the execution of the warrant.[2]
III
Since the police action in this case violated petitioners' Fourth Amendment right, we now must decide whether this right was clearly established at the time of the search. See Siegert, 500 U. S., at 232-233. As noted above, Part II, supra, government officials performing discretionary functions generally are granted a qualified immunity and are "shielded from liability for civil damages insofar as their conduct does not violate clearly established statutory or constitutional rights of which a reasonable person would have known." Harlow v. Fitzgerald, 457 U. S., at 818. What this means in practice is that "whether an official protected by qualified immunity may be held personally liable for an allegedly unlawful official action generally turns on the `objective legal reasonableness' of the action, assessed in light of the legal rules that were `clearly established' at the time it was taken." Anderson v. Creighton, 483 U. S. 635, 639 (1987) (citing Harlow, supra, at 819); see also Graham v. Connor, 490 U. S., at 397.
In Anderson, we explained that what "clearly established" means in this context depends largely "upon the level of generality at which the relevant `legal rule' is to be identified." 483 U. S., at 639. "[C]learly established" for purposes of *615 qualified immunity means that "[t]he contours of the right must be sufficiently clear that a reasonable official would understand that what he is doing violates that right. This is not to say that an official action is protected by qualified immunity unless the very action in question has previously been held unlawful, but it is to say that in the light of preexisting law the unlawfulness must be apparent." Id., at 640 (citations omitted); see also United States v. Lanier, 520 U. S. 259, 270 (1997).
It could plausibly be asserted that any violation of the Fourth Amendment is "clearly established," since it is clearly established that the protections of the Fourth Amendment apply to the actions of police. Some variation of this theory of qualified immunity is urged upon us by petitioners, Brief for Petitioners 37, and seems to have been at the core of the dissenting opinion in the Court of Appeals, see 141 F. 3d, at 123. However, as we explained in Anderson, the right allegedly violated must be defined at the appropriate level of specificity before a court can determine if it was clearly established. 483 U. S., at 641. In this case, the appropriate question is the objective inquiry whether a reasonable officer could have believed that bringing members of the media into a home during the execution of an arrest warrant was lawful, in light of clearly established law and the information the officers possessed. Cf. ibid.
We hold that it was not unreasonable for a police officer in April 1992 to have believed that bringing media observers along during the execution of an arrest warrant (even in a home) was lawful. First, the constitutional question presented by this case is by no means open and shut. The Fourth Amendment protects the rights of homeowners from entry without a warrant, but there was a warrant here. The question is whether the invitation to the media exceeded the scope of the search authorized by the warrant. Accurate media coverage of police activities serves an important public purpose, and it is not obvious from the general principles *616 of the Fourth Amendment that the conduct of the officers in this case violated the Amendment.
Second, although media ride-alongs of one sort or another had apparently become a common police practice,[3] in 1992 there were no judicial opinions holding that this practice became unlawful when it entered a home. The only published decision directly on point was a state intermediate court decision which, though it did not engage in an extensive Fourth Amendment analysis, nonetheless held that such conduct was not unreasonable. Prahl v. Brosamle, 98 Wis. 2d 130, 154 155, 295 N. W. 2d 768, 782 (App. 1980). From the federal courts, the parties have only identified two unpublished District Court decisions dealing with media entry into homes, each of which upheld the search on unorthodox non-Fourth Amendment right to privacy theories. Moncrief v. Hanton, 10 Media L. Rptr. 1620 (ND Ohio 1984); Higbee v. TimesAdvocate, 5 Media L. Rptr. 2372 (SD Cal. 1980). These cases, of course, cannot "clearly establish" that media entry into homes during a police ride-along violates the Fourth Amendment.
At a slightly higher level of generality, petitioners point to Bills v. Aseltine, 958 F. 2d 697 (CA6 1992), in which the Court of Appeals for the Sixth Circuit held that there were material issues of fact precluding summary judgment on the question whether police exceeded the scope of a search warrant by allowing a private security guard to participate in the search to identify stolen property other than that described in the warrant. Id., at 709. Bills, which was decided a mere five weeks before the events of this case, did anticipate today's holding that police may not bring along third parties during an entry into a private home pursuant *617 to a warrant for purposes unrelated to those justifying the warrant. Id., at 706. However, we cannot say that even in light of Bills, the law on third-party entry into homes was clearly established in April 1992. Petitioners have not brought to our attention any cases of controlling authority in their jurisdiction at the time of the incident that clearly established the rule on which they seek to rely, nor have they identified a consensus of cases of persuasive authority such that a reasonable officer could not have believed that his actions were lawful.
Finally, important to our conclusion was the reliance by the United States marshals in this case on a Marshals Service ride-along policy that explicitly contemplated that media who engaged in ride-alongs might enter private homes with their cameras as part of fugitive apprehension arrests.[4] The Montgomery County Sheriff's Department also at this time had a ride-along program that did not expressly prohibit media entry into private homes. Deposition of Sheriff Raymond M. Kight, in No. PJM-94-1718, p. 8. Such a policy, of course, could not make reasonable a belief that was contrary to a decided body of case law. But here the state of the law as to third parties accompanying police on home entries was at best undeveloped, and it was not unreasonable for law enforcement officers to look and rely on their formal ridealong policies.
Given such an undeveloped state of the law, the officers in this case cannot have been "expected to predict the future course of constitutional law." Procunier v. Navarette, 434 *618 U. S. 555, 562 (1978). See also Wood v. Strickland, 420 U. S. 308, 321 (1975); Pierson v. Ray, 386 U. S. 547, 557 (1967). Between the time of the events of this case and today's decision, a split among the Federal Circuits in fact developed on the question whether media ride-alongs that enter homes subject the police to money damages. See 141 F. 3d, at 118 119; Ayeni v. Mottola, 35 F. 3d 680 (CA2 1994), cert. denied, 514 U. S. 1062 (1995); Parker v. Boyer, 93 F. 3d 445 (CA8 1996), cert. denied, 519 U. S. 1148 (1997); Berger v. Hanlon, 129 F. 3d 505 (CA9 1997), cert. granted, 525 U. S. 981 (1998). If judges thus disagree on a constitutional question, it is unfair to subject police to money damages for picking the losing side of the controversy.
For the foregoing reasons, the judgment of the Court of Appeals is affirmed.
It is so ordered.
Justice Stevens, concurring in part and dissenting in part.
Like every other federal appellate judge who has addressed the question, I share the Court's opinion that it violates the Fourth Amendment for police to bring members of the media or other third parties into a private dwelling during the execution of a warrant unless the homeowner has consented or the presence of the third parties is in aid of the execution of the warrant. I therefore join Parts I and II of the Court's opinion.
In my view, however, the homeowner's right to protection against this type of trespass was clearly established long before April 16, 1992. My sincere respect for the competence of the typical member of the law enforcement profession precludes my assent to the suggestion that "a reasonable officer could have believed that bringing members of the media into a home during the execution of an arrest warrant was lawful." Ante, at 615. I therefore disagree with the Court's *619 resolution of the conflict in the Circuits on the qualified immunity issue.[1] The clarity of the constitutional rule, a federal statute (18 U. S. C. § 3105), common-law decisions, and the testimony of the senior law enforcement officer all support my position that it has long been clearly established that officers may not bring third parties into private homes to witness the execution of a warrant. By contrast, the Court's opposing view finds support in the following sources: its bare assertion that the constitutional question "is by no means open and shut," ante, at 615; three judicial opinions that did not directly address the constitutional question, ante, at 616; and a public relations booklet prepared by someone in the United States Marshals Service that never mentions allowing representatives of the media to enter private property without the owner's consent, ante, at 617.
I
In its decision today the Court has not announced a new rule of constitutional law. Rather, it has refused to recognize an entirely unprecedented request for an exception to a well-established principle. Police action in the execution of a warrant must be strictly limited to the objectives of the authorized intrusion. That principle, like the broader protection provided by the Fourth Amendment itself, represents the confluence of two important sources: our English forefathers' traditional respect for the sanctity of the private *620 home and the American colonists' hatred of the general warrant.
The contours of the rule are fairly described by the Court, ante, at 609-611 of its opinion, and in the cases that it cites on those pages. All of those cases were decided before 1992. None of those casesnor, indeed, any other of which I am awareidentified any exception to the rule of law that the Court repeats today. In fact, the Court's opinion fails to identify a colorable rationale for any such exception. Respondents' position on the merits consisted entirely of their unpersuasive factual submission that the presence of representatives of the news media served various legitimate albeit nebulouslaw enforcement purposes. The Court's cogent rejection of those post hoc rationalizations cannot be characterized as the announcement of a new rule of law.
During my service on the Court, I have heard lawyers argue scores of cases raising Fourth Amendment issues. Generally speaking, the Members of the Court have been sensitive to the needs of the law enforcement community. In virtually all of them at least one Justice thought that the police conduct was reasonable. In fact, in only a handful did the Court unanimously find a Fourth Amendment violation. That the Court today speaks with a single voice on the merits of the constitutional question is unusual and certainly lends support to the notion that the question is indeed "open and shut." Ante, at 615.
But the more important basis for my opinion is that it should have been perfectly obvious to the officers that their "invitation to the media exceeded the scope of the search authorized by the warrant." Ibid. Despite reaffirming that clear rule, the Court nonetheless finds that the mere presence of a warrant rendered the officers' conduct reasonable. The Court fails to cite a single case that even arguably supports the proposition that using official power to enable news photographers and reporters to enter a private home for purposes unrelated to the execution of a warrant could *621 be regarded as a "reasonable" invasion of either property or privacy.
II
The absence of judicial opinions expressly holding that police violate the Fourth Amendment if they bring media representatives into private homes provides scant support for the conclusion that in 1992 a competent officer could reasonably believe that it would be lawful to do so. Prior to our decision in United States v. Lanier, 520 U. S. 259 (1997), no judicial opinion specifically held that it was unconstitutional for a state judge to use his official power to extort sexual favors from a potential litigant. Yet, we unanimously concluded that the defendant had fair warning that he was violating his victim's constitutional rights. Id., at 271 ("The easiest cases don't even arise" (citations and internal quotation marks omitted)).
Nor am I persuaded that the absence of rulings on the precise Fourth Amendment issue presented in this case can plausibly be explained by the assumption that the police practice was common. I assume that the practice of allowing media personnel to "ride along" with police officers was common, but that does not mean that the officers routinely allowed the media to enter homes without the consent of the owners. As the Florida Supreme Court noted in Florida Publishing Co. v. Fletcher, 340 So. 2d 914, 918 (1976), there has long been a widespread practice for firefighters to allow photographers to enter disaster areas to take pictures, for example, of the interior of buildings severely damaged by fire. But its conclusion that such media personnel were not trespassers rested on a doctrine of implied consent[2]a theory *622 wholly inapplicable to forcible entries in connection with the execution of a warrant.[3]
In addition to this case, the Court points to three lower court opinionsnone of which addresses the Fourth Amendmentas the ostensible basis for a reasonable officer's belief that the rule in Semayne's Case[4] was ripe for reevaluation.[5] See ante, at 616. Two of the cases were decided in 1980 and the third in 1984. In view of the clear restatement of the rule in the later opinions of this Court, cited ante, at 611, those three earlier decisions could not possibly provide a *623 basis for a claim by the police that they reasonably relied on judicial recognition of an exception to the basic rule that the purposes of the police intrusion strictly limit its scope.
That the two federal decisions were not officially reported makes such theoretical reliance especially anomalous.[6] Moreover, as the Court acknowledges, the claim rejected in each of those cases was predicated on the media's alleged violation of the plaintiffs' "unorthodox non-Fourth Amendment right to privacy theories," ante, at 616, rather than a claim that the officers violated the Fourth Amendment by allowing the press to observe the execution of the warrant. Moncrief v. Hanton, 10 Media L. Rptr. 1620 (ND Ohio 1984); Higbee v. Times-Advocate, 5 Media L. Rptr. 2372 (SD Cal. 1980). As for the other case, Prahl v. Brosamle, 98 Wis. 2d 130, 295 N. W. 2d 768 (App. 1980)cited by the Court, ante, at 616, for the proposition that the officer's conduct was "not unreasonable"it actually held that the defendants' motion to dismiss should have been denied because the allegations supported the conclusion that the officer committed a trespass when he allowed a third party to enter the plaintiff's property.[7] Since that conclusion was fully consistent with a *624 number of common-law cases holding that similar conduct constituted a trespass,[8] it surely does not provide any support for an officer's assumption that a similar trespass would be lawful.
Far better evidence of an officer's reasonable understanding of the relevant law is provided by the testimony of the Sheriff of Montgomery County, the commanding officer of three of the respondents: "`We would never let a civilian into a home. . . . That's just not allowed.' " Brief for Petitioners 41.
III
The most disturbing aspect of the Court's ruling on the qualified immunity issue is its reliance on a document discussing "ride-alongs" apparently prepared by an employee in the public relations office of the United States Marshals Service. The text of the document, portions of which are set out in an appendix, makes it quite clear that its author was not a lawyer, but rather a person concerned with developing the proper public image of the Service, with a special interest in creating a favorable impression with the Congress. Although the document occupies 14 pages in the joint *625 appendix and suggests handing out free Marshals Service T-shirts and caps to "grease the skids," it contains no discussion of the conditions which must be satisfied before a newsperson may be authorized to enter private property during the execution of a warrant. App. 12. There are guidelines about how officers should act and speak in front of the camera, and the document does indicate that "the camera" should not enter a private home until a "signal" is given. Id., at 7. It does not, however, purport to give any guidance to the marshals regarding when such a signal should be given, whether it should ever be given without the consent of the homeowner, or indeed on how to carry out any part of their law enforcement mission. The notion that any member of that well-trained cadre of professionals would rely on such a document for guidance in the performance of dangerous law enforcement assignments is too farfetched to merit serious consideration.
* * *
The defense of qualified immunity exists to protect reasonable officers from personal liability for official actions later found to be in violation of constitutional rights that were not clearly established. The conduct in this case, as the Court itself reminds us, contravened the Fourth Amendment's core protection of the home. In shielding this conduct as if it implicated only the unsettled margins of our jurisprudence, the Court today authorizes one free violation of the wellestablished rule it reaffirms.
I respectfully dissent.
APPENDIX TO OPINION OF STEVENS, J. "MEDIA RIDE-ALONGS
"The U. S. Marshals Service, like all federal agencies, ultimately serves the needs and interests of the American public *626 when it accomplishes its designated duties. Keeping the public adequately informed of what the Service does can be viewed as a duty in its own right, and we depend on the news media to accomplish that. "Media `ride-alongs' are one effective method to promote an accurate picture of Deputy Marshals at work. Ride-alongs, as the name implies, are simply opportunities for reporters and camera crews to go along with Deputies on operational missions so they can see, and record, what actually happens. The result is usually a very graphic and dynamic look at the operational activities of the Marshals Service, which is subsequently aired on TV or printed in a newspaper, magazine, or book.
"However, successful ride-alongs don't just `happen' in a spontaneous fashion. They require careful planning and attention to detail to ensure that all goes smoothly and that the media receive an accurate picture of how the Marshals Service operates. This booklet describes considerations that are important in nearly every ride-along." App. 4. "Establish Ground Rules "Another good ideaactually, it's an essential oneis to establish ground rules at the start and convey them to the reporter and camera person. Address such things as what can be covered with cameras and when, any privacy restrictions that may be encountered, and interview guidelines. "Emphasize the need for safety considerations and explain any dangers that might be involved. Make the ground rules realistic but balancedremember, the media will want good action footage, not just a mop-up scene. If the arrest is planned to take place inside a house or building, agree ahead of time on when the camera can enter and who will give the signal." Id., at 7.
"The very best planning won't result in a good ride-along if the Marshals Service personnel involved do not do their part. It's a case of actions speaking as loudly as words, and both *627 are important in getting the best media exposure possible." Id., at 9.
"`Waving the Flag'
"One action of special consequence is `waving the flag' of the Marshals Service. This is accomplished when Deputies can easily be recognized as USMS Deputies because they are wearing raid jackets, prominently displaying their badges, or exhibiting other easily identifiable marks of the Service. We want the public to know who you are and what kind of job you do. That is one of the goals of the ride-along. So having Deputy Marshals easily identified as such on camera is not just a whimit's important to the overall success of the ride-along.
"Of course, how the Deputies act and what they say is also crucial. During the ride-along virtually any statement made by Deputies just might end up as a quote, attributed to the person who made it. Sometimes that could prove embarrassing. A Deputy must try to visualize what his or her words will look like in a newspaper or sound like on TV. Being pleasant and professional at all times is key, and that includes not being drawn into statements of personal opinion or inappropriate comments. Using common sense is the rule." Id., at 9-10.
"You also need to find out when the coverage will air or end up in print. Ask the reporter if he or she can keep you informed on that matter. You might `grease the skids' for this by offering the reporter, camera person, or other media representatives involved a memento of the Marshals Service. Marshals Service caps, mugs, T-shirts, and the like can help establish a rapport with a reporter that can benefit you in the future." Id., at 12.
"Getting to the Final Product
"Naturally, it's important to see the final product of the ridealong when it airs on TV or appears in the newspaper. You *628 should arrange to videotape any TV news coverage or clip the resulting newspaper stories and send a copy of the videotape or news clipping to the Office of Congressional and Public Affairs." Id., at 13.
NOTES
[*] A brief of amici curiae urging affirmance was filed for ABC, Inc., et al. by Lee Levine, James E. Grossberg, Jay Ward Brown, Henry S. Hoberman, Richard M. Schmidt, Jr., Susanna M. Lowy, Harold W. Fuson, Jr., Barbara Wartelle Wall, Ralph E. Goldberg, Karlene W. Goller, Jerry S. Birenz, Slade R. Metcalf, Jack N. Goodman, David S. J. Brown, René P. Milam, George Freeman, and Jane E. Kirtley.
[1] The warrants were identical in all relevant respects.By way of example, one of them read as follows:
"The State of Maryland, to any duly authorized peace officer, greeting: you are hereby commanded to take Dominic Jerome Wilson if he/she shall be found in your bailiwick, and have him immediately before the Circuit Court for Montgomery County, now in session, at the Judicial Center, in Rockville, to answer an indictment, or information, or criminal appeals unto the State of Maryland, of and concerning a certain charge of Robbery [Violation of Probation] by him committed, as hath been presented, and so forth. Hereof fail not at your peril, and have you then and there this writ. Witness." App. 36-37.
[2] Even though such actions might violate the Fourth Amendment, if the police are lawfully present, the violation of the Fourth Amendment is the presence of the media and not the presence of the police in the home. We have no occasion here to decide whether the exclusionary rule would apply to any evidence discovered or developed by the media representatives.
[3] See, e. g., Florida Publishing Co. v. Fletcher, 340 So. 2d 914, 919 (1976) (it "`is a widespread practice of long-standing' " for media to accompany officers into homes), cert. denied, 431 U. S. 930 (1977); Zoglin, Live on the Vice Beat, Time, Dec. 22, 1986, p. 60 (noting "the increasingly common practice of letting TV crews tag along on drug raids").
[4] A booklet distributed to marshals recommended that "fugitive apprehension cases . .. normally offer the best possibilities for ride-alongs." App. 4-5. In its discussion of the best way to make ride-alongs useful to the media and portray the Marshals Service in a favorable light, the booklet noted that reporters were likely to want to be able to shoot "good action footage, not just a mop-up scene." It advised agents that "[i]f the arrest is planned to take place inside a house or building, agree ahead of time on when the camera can enter and who will give the signal." Id., at 7.
[1] It is important to emphasize that there is no split in Circuit authority on the merits of the constitutional issue. Nor, as I explain infra, at 622-624, do I believe that any District Court had reached a conclusion at odds with the Court's Fourth Amendment holding. Any conflict was limited to the qualified immunity issue. Three Circuits rejected the defense whereas the Fourth and the Eighth accepted it. See Ayeni v. Mottola, 35 F. 3d 680, 686 (CA2 1994); Bills v. Aseltine, 958 F. 2d 697 (CA6 1992); Berger v.Hanlon, 129 F. 3d 505 (CA9 1997); 141 F. 3d 111 (CA4 1998) (en banc); Parker v. Boyer, 93 F. 3d 445 (CA8 1996).
[2] The Florida Supreme Court held: "The trial court properly determined from the record before it that there was no genuine issue of material fact insofar as the entry into respondent's home by petitioner's employees became lawful and non-actionable pursuant to the doctrine of common custom, usage, and practice and since it had been shown that it was common usage, custom and practice for news media to enter private premises and homes under the circumstances present here.
. . . . .
"`The fire was a disaster of great public interest . . . . [I]t has been a longstanding custom and practice throughout the country for representatives of the news media to enter upon private property where disaster of great public interest has occurred.' " 340 So. 2d, at 917-918.
The Court's reference to this case, ante, at 616, n. 3, misleadingly suggests that the "widespread practice" referred to in the Florida court's opinion was police practice; it was not.
[3] Indeed, the Wisconsin state-court decision, cited by the Court as contrary authority, took pains to distinguish this case: "We will not imply a consent as a matter of law. It is of course well known that news representatives want to enter a private building after or even during a newsworthy event within the building. That knowledge is no basis for an implied consent by the possessor of the building to the entry. . . . We conclude that custom and usage have not been shown in fact or law to confer an implied consent upon news representatives to enter a building under the circumstances presented by this case." Prahl v. Brosamle, 98 Wis. 2d 130, 149-150, 295 N. W. 2d 768, 780 (App. 1980).
[4] 5 Co. Rep. 91a, 77 Eng. Rep. 194 (K. B. 1604).
[5] As the Court notes, the only Federal Court of Appeals authority on the subject, Bills v. Aseltine, 958 F. 2d 697 (CA6 1992), "anticipate[d] today's holding that police may not bring along third parties during an entry into a private home pursuant to a warrant for purposes unrelated to those justifying the warrant." Ante, at 616-617.
[6] In the Fourth Circuit, unreported opinions may not be considered in the course of determining qualified immunity. Hogan v. Carter, 85 F. 3d 1113, 1118 (1996).
[7] Prahl v. Brosamle, 98 Wis. 2d, at 154-155, 295 N. W. 2d, at 782 ("A new trial must be had with respect to the plaintiffs' claims for trespass against Lieutenant Kuenning and Dane Country . . . . Lieutenant Kuenning had no authority to extend a consent to [the press] to enter the land of another. Although entry by Lieutenant Kuenning was privileged, he committed a trespass by participating in the trespass by [the press]").
The Court is correct that the Wisconsin Court of Appeals upheld dismissal of the plaintiff's 42 U. S. C. § 1983 claim against the newscaster because he was not acting under color of state law. As the basis for rejecting the § 1983 action "for invasion of privacy based on disclosure of the incident," the court further held that "[w]e are unwilling to accept the proposition that the filming and television broadcast of a reasonable search and seizure, without more, result in unreasonableness." 98 Wis. 2d, at 138, 295 N. W. 2d, at 774. Important to its conclusion was its observation that, unlike the unnecessary male participation in body searches of schoolgirls in Doe v. Duter, 407 F. Supp. 922 (WD Wis. 1976), "[n]either the search of Dr. Prahl and his premises nor the film or its broadcast has been shown to include intimate, offensive or vulgar aspects." 98 Wis. 2d, at 138, 295 N. W. 2d, at 774. The reporter in question was stationed in the entryway of the building and was able to film into the plaintiff's office during the police interview.
[8] See, e. g., Daingerfield v. Thompson, 74 Va. 136, 151 (1880) ("There seems, indeed, to be no principle of law better settled, and for which numerous authorities may be cited if necessary, than this: that all persons who wrongfully contribute in any manner to the commission of a trespass, are responsible as principals, and each one is liable to the extent of the injury done"); see also W. Keeton, D. Dobbs, R. Keeton, & D. Owen, Prosser and Keeton on Law of Torts § 13, p. 72 (5th ed. 1984).
|
Four construction sites are currently in operation, two in Austria and two in Italy. You can see the details of the construction lots by clicking on our interactive route plan. The excavation progress reports of the individual tunnel structures in the various construction lots are updated weekly. Work progress is shown on a graph.
Tulfes-Pfons
GENERAL INFORMATION ON THE CONSTRUCTION LOT
This construction lot, worth 380 million Euro, was awarded to the Strabag/Salini-Impregilo bidding consortium in Summer of 2014. Excavation works began in September 2014 and will last until Spring of 2019. The construction lot includes 38 km of tunnel excavation work.
It consists of several structures:
Tulfes emergency tunnel
Innsbruck emergency stop
Main tunnel tubes
Connecting tunnels
Ahrental-Pfons exploratory tunnel
CURRENTLY ONGOING CONSTRUCTION WORK
Emergency tunnel
The emergency tunnel is being driven parallel to the existing Innsbruck railway bypass; it will be 9 km long and the excavation cross-section is 35 m². The drill-and-blast excavation work on this tunnel starts from three points at the same time: from Tulfes westwards (already completed), from the Ampass access tunnel eastwards and again westwards. The emergency tunnel will be completed in Summer of 2017.
Main tunnel tubes
The main tubes have been excavated by blasting according to the so-called New Austrian Tunnel Construction Method. The ceremonial first volley was set off on 19/03/2015 in the presence of EU Commissar Violeta Bulc and the Ministers of Transportation of the 7 Alpine countries (Austria, Italy, Germany, Switzerland, France, Liechtenstein and Slovenia. This stretch includes 3 km of main tunnels with an excavated cross-section of about 70 m2 .
Connecting tunnel between the BBT and the Innsbruck bypass
Construction has been under way since Summer of 2015 on the connecting tunnels between the Brenner Base Tunnel and the existing Innsbruck bypass. The connecting tunnels will also be excavated by blasting, with a cross-section of about 115 m2 .
Exploratory Tunnel
The construction lot includes the construction of the 15-km stretch of exploratory tunnel between the Ahrental junction point and the town of Pfons. The open gripper TBM with a length of 200 m has started on September 26th, 2015. Until spring 2015 the machine will excavate 15 km of the exploratory tunnel in south direction.
Wolf
GENERAL INFORMATION ON THE CONSTRUCTION LOT
This construction lot, worth 104 million Euro, was awarded to the Swietelsky/Swietelsky Tunnelbau bidding consortium in the fall of 2013. Excavation work on the Wolf access tunnel began on December 4th 2013.
The construction lot includes not only the work on the Wolf access tunnel and part of the exploratory tunnel, but also a series of logistics and safety works for the disposal site in the Padaster valley.
The construcion lot includes:
The Wolf access tunnel
The Wolf exploratory tunnel
The Padaster brook re-routing tunnel
The spoil removal tunnel
The debris barrier in the Padaster valley
The intake structure for the Padaster brook
The inner lining of the Padaster tunnel
The inner lining for the Saxen tunnel
Wolf access tunnel and exploratory tunnel
The 3.257 metre Wolf access tunnel, with an excavation cross-section of 120 m², was excavated by blasting from December 2013 to November 2015. It covers a 400-meter incline and ends in a junction chamber south of St. Jodok that has a cross-section of 280 m². Subsequently, the base slab of the access tunnel was further re-worked with the installation of the concrete track bed.
The Wolf 2 construction lot also includes a 1.2 km stretch of the exploratory tunnel to be excavated towards Brenner (to the south), a junction chamber and connecting tunnels.
All the works in this construction lot, which were carried out in preparation for the main Pfons-Brenner construction lot starting in Fall of 2018, were completed in September of 2017.
Exploratory tunnel Wolf (completed)
Excavation1,200 m
Length1,200 m
Cross section49 m²
Access tunnel Wolf (completed)
Excavation4,000 m
Length4,000 m
Cross section120 m²
Overview map
CONSTRUCTION PROGRESS: WOLF
Gallery
Construction site Wolf Construction site Wolf Construction site Wolf
Mules 2-3
GENERAL INFORMATION ON THE CONSTRUCTION LOT
The construction lot “Periadriatic seam”, carried out between October 2011 and the summer of 2015, comprised the excavation of 3.7 km of the main tubes and 1.5 km of the exploratory tunnel. The crossing of the Periadriatic fault line, one of the biggest geological fault lines in the Alps, was accomplished without a hitch.
In May 2016 the biggest construction lot of the Brenner Base Tunnel project, known as “Mules 2-3”, with a contract volume of 993 mln euro, was awarded to the bidding consortium Astaldi S.p.A., Ghella S.p.A., Oberosler Cav Pietro S.r.l., Cogeis S.p.A. and PAC S.p.A. Works started in September 2016.
This lot reaches from the construction lot “Isarco river underpass” up to the border with Austria. In the course of a 7-year construction period, 39,8 km of the main tubes and 14.8 km of the exploratory tunnel will be excavated, including the emergency stop in Trens and its access tunnel as well as the bypasses which connect the main tubes every 333 m. A total of approximately 65 kilometers of tunnels will be excavated.
Once this construction lot is finished, all excavation activities in the Italian project area will be complete.
CURRENT CONSTRUCTION WORK
Access to emergency stop in TrensThe blasting work to connect the Mules access tunnel with the future emergency stop in Trens started in December 2016. The 3,8 km long tunnel with a cross-section of approximately 80 m2 will presumably be excavated by 2021.
Exploratory TunnelThe drill and blast excavation works on the exploratory tunnel northwards resumed on 13.02.2017. After approximately 600 m, the rest of the stretch, all the way to the border, will be excavated with a tunnel boring machine (TBM). The driving operations are expected to be completed by the end of 2021.
Main tunnelsThe drift of the main tubes northwards resumed on the 13th of march 2017. The first section of the stretch up to the future location of the Trens emergency stop, where new assembly chambers will be built, is going to be excavated by drill and blast in nearly two years. From here TBMs will drive the remaining section of the tunnels northwards up to the border of the construction lot, which should be reached during the first months of 2022.The first blast of the southwards drift took place on the 3rd of april 2017. The southern end of the construction lot is expected to be reached in mid-2020.
EXCAVATION DATA (Jenuary 14th, 2019)
Access tunnel to the emergency stop
Excavation2.001 m
Length3,797 m
Cross-section80 m²
Main tunnels
Excavation6,468 m
Length39,856 m
Cross-section85 m²
Exploratory tunnel
Excavation3,487 m
Length14,758 m
Cross-section35 m²
Safety and logistics tunnels
Excavation754 m
Length6,930 m
Cross-section26-56 m²
Overview map
CONSTRUCTION PROGRESS: MULES 2-3
Gallery
TBM for the exploratory tunnel Mules-Brennero (excavation since May 2018) Arrival of a TBM component for the exploratory tunnel Mules-Brennero (excavation since May 2018) Tübbing production plant at Hinterrigger Tübbing production plant at Hinterrigger Tübbing production plant at Hinterrigger Northwards construction of the main tubes TBM "Virginia" for the excavation of one of two main tubes in northward direction TBM "Flavia" for one of the two main tunnels towards Brennero
Isarco river underpass
GENERAL INFORMATION ON THE CONSTRUCTION LOT
The southernmost construction lot of the Brenner Base Tunnel, worth 303 million Euro, was awarded to the RTI Salini-Impregilo S.p.A., Strabag AG, Strabag S.p.A., Consorzio Integra and Collini Lavori S.p.A. bidding consortium in October of 2014. This lot links the Brenner Base Tunnel with the existing Brenner line and the railway station in Fortezza. The work is scheduled to be completed in 2022.
Since in this section the tunnel tubes are just a few meters below the surface, a portion of the activities pertaining to this construction lot will be carried out building artificial tunnels. Furthermore, as loose fluvioglacial materials and the groundwater layer will be crossed, it will be necessary to adopt specific ground consolidation procedures including ground freezing and the so-called jet grouting.
CURRENT CONSTRUCTION WORK
As part of the preparatory work, by the end of 2016 the state road SS12 was re-routed and two bridges, an underpass under the Brenner railway line and a direct highway access were built. This means that both supply and removal transports can take place while avoiding traffic through the surrounding towns.
The current phase (main construction phase I) provides for the construction of four shafts, 20-25 metres deep, to access the rock faces of the Isarco River Underpass. Furthermore, it includes the excavation - already completed - of the access tunnel on the right side of the valley using drill and blast, and of part of the main tubes and the connecting tunnels, to be built with the drill and blast method.
As a preparatory measure for the relocation of the existing line which will be carried out during the following construction phase, slope stabilization and earthmoving measures will be implemented.
EXCAVATION DATA (December 14th, 2019)
Main tunnels
Excavation2,151 m
Length4,478 m
Cross-section60-150 m²
Connecting tunnels
Excavation576 m
Length1,186 m
Cross-section60-80 m²
Connecting side tunnels
Excavation96 m
Length170 m
Cross-section35-45 m²
Access tunnel (completed)
Excavation192 m
Length192 m
Cross-section65 m²
Overview map
CONSTRUCTION PROGRESS AT THE ISARCO RIVER UNDERPASS
Gallery
Excavation of the Western main tube Isarco river underpass Construction of the shafts Meeting point between access tunnel and connecting tunnel Consolidation with jet-grouting Construction of the tunnel floor in the connecting tunnel Isarco river underpass Isarco river underpass Construction of the shafts
|
Cephalosporin-induced leukopenia following rechallenge with cefoxitin.
To describe a case of cefazolin-induced leukopenia in a critically ill patient who developed this adverse reaction upon rechallenge with cefoxitin. A 22-year-old male was admitted after a motor vehicle crash. beta-Lactam therapy was initiated with vancomycin, cefepime, and metronidazole and, upon identification of methicillin-sensitive Staphylococcus aureus bacteremia 4 days later, therapy was narrowed to cefazolin 1 g every 12 hours. The dose was adjusted to 1 g every 12 hours during continuous venovenous hemodialysis. Imipenem was given for 2 days, resulting in a total of 18 days of beta-lactam treatment, at which time he developed significant leukopenia (white blood cell [WBC] count 0.9 x 10(3)/microL). Antimicrobial treatment was changed to tigecycline and continued for suspected pleural space infection. The patient's WBC count recovered within 4 days after the change in therapy. He was taken to surgery 8 days after cefazolin was discontinued and received perioperative prophylaxis with cefoxitin (total dose 3 g). Subsequently, the patient again became severely leukopenic (WBC count 2.4 x 10(3)/microL). Within a week after surgery, the patient developed septic shock secondary to multidrug-resistant Escherichia coli bacteremia and died. beta-Lactam-induced leukopenia is a rare but well-described adverse drug reaction. It is a cumulative dose-dependent phenomenon reported to occur most often after 2 weeks of therapy. The mechanism of leukopenia is thought to be secondary to either an immune-mediated response or direct bone marrow toxicity. Rechallenge with a different beta-lactam antibiotic has not been shown to consistently cause recurrent leukopenia. The case described here suggests an immune-related mechanism for the development of leukopenia. Use of the Naranjo probability scale determined the association between cephalosporin use and leukopenia to be probable. Cefazolin was a probable cause of this patient's leukopenia. It is important for clinicians to recognize beta-lactam-induced leukopenia and maybe recommend use of a drug from a different antibiotic class if continued treatment is indicated.
|
WVI Blog
Our blog is a combination of inspiration and achievements. A blend of the right and left sides of our brains.
Albert Einstein said, “We cannot solve our problems with the same level of thinking we used when we created them.” The question is, how do we go about changing our thinking? Worldview Intelligence has been proven to provide a mechanism to do just that. It is an approach that can be applied to a variety of
In September 2011, Jerry Nagel was invited to do a TEDx talk on worldview – When Worldviews Collide. His understanding of worldview has evolved from his research for his PhD on Worldview, Art of Hosting and Social Constructionist theory and the practical experience in AoH trainings, other events and now specifically the rapidly evolving curriculum for Worldview
The Worldview Intelligence Framework is rooted in solid academic and philosophical research and rigour thanks to the work and insight of Belgian philosopher and logician Leo Apostel (1925-1995). In his day, Apostel felt the world was becoming increasingly fragmented, particularly in the sciences, and he wanted to create a way to encourage communication across disciplines. He invited many
What is Worldview Intelligence? Where did it come from? Where is it being used? We are asked these questions regularly and here is a bit of what we have been telling people about our work. Theory, Philosophy and Practice Behind Worldview Intelligence Leo Apostel (1925-95) was a Belgium philosopher at the Vrije Universiteit Brussel (VUB) who was concerned
|
Conformations and electronic structures of axially coordinated fullerene-porphyrin-fullerene triad (C60CHCOO)2-Sn(IV) porphyrin.
The conformational (cis and trans) stability and electronic structures of (C(60)CHCOO)(2)-Sn(IV) porphyrin, recently synthesized as a novel fullerene-porphyrin-fullerene triad linked by metal axial coordination, have been studied by ab initio calculations. The cis conformer was found to be slightly more stable than the trans by 1.38 kcal/mol in the neutral compound. Upon the addition of an electron to the triad, the relative stability of the cis conformer was found to be higher (3.29 kcal/mol) than that in the neutral one. From the investigation of frontier molecular orbitals, for the cis conformer, it was found that the electrons are localized in HOMO of the porphyrin, while the electrons are localized in LUMO of the syn-fullerene. For the trans conformer, it was found that the electrons are localized in HOMO of the porphyrin, while the electrons are localized in LUMO of one of the two fullerene moieties, and the electrons are localized in LUMO2 of the other fullerene moiety, but the LUMO and LUMO2 have the same orbital energy. Thus, the PET may take place unidirectionally in the cis conformer from the porphyrin to the syn-fullerene, while it is bidirectional from the porphyrin to both of the fullerene moieties.
|
Theories of Criminal Behavior Documents
Showing 1 to 30 of 36
1
Running Head: FINAL CONFESSIONS
Final Confessions
James Knowlton
Southern Oregon University
2[Type text]
[Type text]
[Type text]
Introduction
In Brian Wallace and Bill Crowleys true story Final Confession, they tell the
story of one of Bostons most noto
THEORIES OF CRIMINAL BEHAVIOR
FROM THEORY TO PRACTICE
Crime and the American Dream
THEORY
Associated Theorist(s): Richard Rosenfeld and Steven F. Messner
Orientation/Classification: Strain
Basic Premise, Description, Elements of the Theory:
Why does the
Running head: Seven Crimes Essay
1
Seven Crimes Essay
Alexandria Trujillo
Southern Oregon University
Seven Crimes Essay
2
Introduction
Theories are often derived from a serious of events that require an explanation and are
based on the principles that hel
Social Structure and Anomie - Robert K. Merton
This is one of the most famous pieces of early 20th century sociological writings and
crime, and, relatively speaking, fairly easy to read (trust me: try reading some Talcott
Parsons!). Robert K Merton (1910
GENE-BASED EVOLUTIONARY THEORIES
IN CRIMINOLOGY*
LEE ELLIS
Minot State University
ANTHONY WALSH
Boise State University
In the past 20 years, several theories of criminal (and antisocial)
behavior have been proposed from an evolutionary perspective, some
o
1. Routine Activities Theory Routine Activity Theory is a subfield of rational choice and criminology, developed by Marcus
Felson and Lawrence E Cohen.
2. Routine Activity Theory says that crime is normal and
depending on the circumstances crime will be c
THEORIES OF CRIMINAL BEHAVIOR
FROM THEORY TO PRACTCE
Social Structure and Anomie
THEORY
Associated Theorist: Robert K. Merton
Orientation/Classification: Sociology/Learning
Basic Premise, Description, Elements of the Theory:
The social order is solely a
|
Q:
Displaying a PDF in a WPF Application Not Working - WebBrowser or Adobe Control
I need to display a PDF in a WPF application. From all the reading I've been doing online, it seems the [only?] ways of displaying a PDF within a WPF application is via Adobe's controls or a WebBrowser control. I've tried to use Adobe's controls, however, I have been unable to add the Reader control because for some reason I can't find the it as something I can add into my Toolbox (even once adding the reference needed). I'm running Windows 7 (64-bit), VS2010, .NET 4.0, and have Adobe Acrobat 7.0 Professional and Adobe Acrobat 9 Pro Extended installed, if that has anything to do with it. So anyway, I decided to try it in a WebBrowser control hosted in a WindowsFormsHost. The XAML I have is this:
<WindowsFormsHost x:Name="FormsHost" Visibility="Visible" Grid.Column="1" Margin="7,0,0,0">
<WF:WebBrowser x:Name="WebBrowser" Dock="Fill" IsWebBrowserContextMenuEnabled="False" ScriptErrorsSuppressed="True" WebBrowserShortcutsEnabled="False" Margin="7,0,0,0" />
</WindowsFormsHost>
and then this in the C# code behind:
WebBrowser.Navigate(new System.Uri(FileName));
where FileName == the exact location of the .pdf file I need to display. However, when I run this all I see is a completely blank, white area where the WebBrowser is. I've also tried setting the .pdf file like this:
WebBrowser.Url = new System.Uri(FileName);
and I get the exact same thing. I know the PDF is being created in the correct location, as I can manually browse to it and open it fine.
Does anyone have any ideas as to why this isn't working? Or possibly why I don't seem to have the Reader control as an option? At this point, either solution would be fine, they just neither one seem to be working!
Thanks!
A:
Here is what I did...
MAIN WINDOW XAML
<Window x:Class="PDFView.MainWindow"
xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
Title="PDFView">
<Grid>
<WebBrowser x:Name="Browser" />
</Grid>
</Window>
MAIN WINDOW CODE BEHIND
using System.IO;
using System.Net;
using System.Windows;
using System.Windows.Navigation;
namespace PDFView {
/// <summary>
/// Interaction logic for MainWindow.xaml
/// </summary>
public partial class MainWindow : Window
{
private const string hostUri = "http://localhost:8088/PsuedoWebHost/";
private HttpListener _httpListener;
public MainWindow() {
InitializeComponent();
this.Loaded += OnLoaded;
}
/// <summary>
/// Loads the specified PDF in the WebBrowser control
/// </summary>
/// <param name="sender"></param>
/// <param name="routedEventArgs"></param>
private void OnLoaded(object sender, RoutedEventArgs routedEventArgs)
{
// Get the PDF from the 'database'
byte[] pdfBytes = GetPdfData();
// Create a PDF server that serves the PDF to a browser
CreatePdfServer(pdfBytes);
// Cleanup after the browser finishes navigating
Browser.Navigated += BrowserOnNavigated;
Browser.Navigate(hostUri);
}
/// <summary>
/// Retrieve a byte array from a 'database record'
/// </summary>
/// <returns></returns>
private byte[] GetPdfData() {
// TODO: Replace this code with data access code
// TODO: Pick a file from the file system for this demo.
string path = @"c:\Users\Me\Documents\MyPDFDocument.pdf";
byte[] pdfBytes = File.ReadAllBytes(path);
// Return the raw data
return pdfBytes;
}
/// <summary>
/// Creates an HTTP server that will return the PDF
/// </summary>
/// <param name="pdfBytes"></param>
private void CreatePdfServer(byte[] pdfBytes) {
_httpListener = new HttpListener();
_httpListener.Prefixes.Add(hostUri);
_httpListener.Start();
_httpListener.BeginGetContext((ar) => {
HttpListenerContext context = _httpListener.EndGetContext(ar);
// Obtain a response object.
HttpListenerResponse response = context.Response;
response.StatusCode = (int)HttpStatusCode.OK;
response.ContentType = "application/pdf";
// Construct a response.
if (pdfBytes != null) {
response.ContentLength64 = pdfBytes.Length;
// Get a response stream and write the PDF to it.
Stream oStream = response.OutputStream;
oStream.Write(pdfBytes, 0, pdfBytes.Length);
oStream.Flush();
}
response.Close();
}, null);
}
/// <summary>
/// Stops the http listener after the browser has finished loading the document
/// </summary>
/// <param name="sender"></param>
/// <param name="navigationEventArgs"></param>
private void BrowserOnNavigated(object sender, NavigationEventArgs navigationEventArgs)
{
_httpListener.Stop();
Browser.Navigated -= BrowserOnNavigated;
}
}
}
A:
I'm afraid you are stuck using the WinForms PDF control and hosting it in your WPF application. Its a little annoying, but its not difficult to do. Here is an article and some sample source code:
http://hugeonion.com/2009/04/06/displaying-a-pdf-file-within-a-wpf-application/
|
I ported BOOM over to our fav platform to play around with Graffiti support (IndivisionAGA Mrk2). I thought I may as well release it now while I'm waiting for the new IndivisionAGA.
BOOM is an enhanced DOOM port that supports:
1) Removal of engine limits and bugs. This includes the visplane limit, the tutti-frutti and medusa effects, the save game size limit, the venetian blind crash, and many others.
2) SEO's to the engine.
3) extra editing features. These include: Configurable animated and switch textures. Deep effects water. scrolling walls, floors, and ceilings (including support for conveyor belts). Translucent walls. friction effects, examined as mud and ice. custom color maps (which can provide, for example to underwater blue "tint"). silent teleporters, Which can be useful for fake "room over room" effects. linedef Generic types - A Particular linedef behavior can be "calculated" using a separate program called TRIGCALC.EXE linedef calculator. A DeHackEd extension standard, BEX.
Hi,
checked Boom on my A4000D/060 with PicassoIV. Runs great with AGA but with Tooltype RTG set i only can choose a Screenmode, then nothing happens. SnoopDOS says nothing, Amiga acts after that like i not startet Boom.
Must i set up a special Screenmode?
Are you further enchancing this port Nova? Optimizing speed etc?
So I did it and began playing it all over again with ultra-violence!
It's about as fast as I can get it, when Jens releases the IndivisionAGA Mrk2 I'll add Graffiti support which should make it quicker for slower machines (060's should already be able to run it full speed).
It's about as fast as I can get it, when Jens releases the IndivisionAGA Mrk2 I'll add Graffiti support which should make it quicker for slower machines (060's should already be able to run it full speed).
Sorry, surely this has been asked, but will ScummVm take benefit of that Graffiti support too?
Sorry, surely this has been asked, but will ScummVm take benefit of that Graffiti support too?
That was my original plan but I don't think it will really help much to be honest. I'm already happy with the performance of the 040/060 C2P versions already and of course the RTG versions. The ECS version is very slow but obviously Graffiti won't help there
The 030 version is also a little slow but the problem is not with screen drawing. To get it running at a decent speed on an 030 I'd need to use an older version of ScummVM.
I ported BOOM over to our fav platform to play around with Graffiti support (IndivisionAGA Mrk2). I thought I may as well release it now while I'm waiting for the new IndivisionAGA.
BOOM is an enhanced DOOM port that supports:
1) Removal of engine limits and bugs. This includes the visplane limit, the tutti-frutti and medusa effects, the save game size limit, the venetian blind crash, and many others.
2) SEO's to the engine.
3) extra editing features. These include: Configurable animated and switch textures. Deep effects water. scrolling walls, floors, and ceilings (including support for conveyor belts). Translucent walls. friction effects, examined as mud and ice. custom color maps (which can provide, for example to underwater blue "tint"). silent teleporters, Which can be useful for fake "room over room" effects. linedef Generic types - A Particular linedef behavior can be "calculated" using a separate program called TRIGCALC.EXE linedef calculator. A DeHackEd extension standard, BEX.
|
Q:
Only Open Folds With z Commands
If I navigate the first line of a folded function with h,l, etc it will open the fold. Navigating search matches with n will also open a fold if there is a match inside the fold.
Can you configure folds to only open when you explicitly open them with z commands?
A:
You can set the 'foldopen' option to control which commands or movements will open folds.
The specific cases you mentioned:
h or l: these are horizontal movements, controlled by the hor setting.
search or n: these are controlled by the search setting.
So you might want to remove these two specific cases, which you can do with:
set foldopen-=hor foldopen-=search
Or maybe you want to disable all of them and just set 'foldopen' to be empty:
set foldopen=
|
Jarosz Folder, Tanto Md: 7506
The KA-BAR/Jarosz Tanto Folder, a liner lock folding knife, features a hollow ground AUS 8A stainless steel blade and glass-filled nylon handle. Designed with a moveable clip for tip up or tip down carry, this modestly sized folder is perfect for pocket or pack. All at a fraction of the cost of Jesse's custom folders.
|
Introduction: Eclipsed by the disastrous financial news of recent months, as well as relatively superficial considerations-lipstick, hockey mom, moose huntress-Sarah Palin’s political resume and heritage is genuinely frightening. Although downplayed or ignored by mainstream media, Sarah Palin’s profound relationship with the Alaskan Independence Party (AIP) augurs very poorly for American politics. Anti-American, politically retrograde, deeply racist and dedicated to fracturing the republic, the AIP is the Alaskan element of the Constitution Party. This fascist political party has run candidates endorsed by the Aryan Nations and dedicated to the successful re-establishment of the Confederacy. In addition to echoing the neo-Confederate endorsement of Southern Secession, the AIP has actively sought the support of the Republic of Iran in attempts to separate Alaska from the United States. Earlier in 2008, Palin sent a greeting to the AIP’s convention, endorsing their efforts. Those knowledgeable about Palin’s political heritage assert that this greeting was predictable. In fact, the AIP had much to do with shaping Palin’s political agenda when she was mayor of Wasilla and she continued to permit access to her AIP mentors when she became governor. (Her husband Todd Palin was a member of the AIP until recently and is described by acquaintances as a “shadow governor.”) Former Nixon cabinet official Walter Hickel became governor of Alaska running as the AIP candidate! Although he later switched back to the GOP, Hickel takes credit for Palin’s election. Of great significance is the AIP’s ideological affiliation with other secessionist elements around the world, including elements associated with the UNPO, an organization that champions the independence of “unrepresented peoples.” Although it represents itself as egalitarian and progressive, the UNPO’s agenda actually works to undermine the political and geographical integrity of larger countries that might present a rival to the Underground Reich. Of more than passing significance is the fact that Palin quoted fascist columnist Westbrook Pegler in her speech before the Republican Convention. It is also worth contemplating what the Nazi elements in the Bush administration will do with a seated African-American President.
Program Highlights Include: The AIP’s support for Tibetan secession, the AIP’s support for the Lakota secessionist movement; the AIP’s support for Hawaiian secession; the AIP’s support for Chechen secession; the endorsement of Constitution Party Presidential candidate Michael Peroutka; by Aryan Nations head Richard Butler; Sarah Palin’s origins in Sandpoint Idaho-an epicenter of the White Supremacist movement.
1. The secessionist party with which Sarah Palin is closely affiliated has a violently anti-American agenda.
“‘My government is my worst enemy. I’m going to fight them with any means at hand.’
This was former revolutionary terrorist Bill Ayers back in his old Weather Underground days, right? Imagine what Sarah Palin is going to do with this incendiary quote as she tears into Barack Obama this week.
Only one problem. The quote is from Joe Vogler, the raging anti-American who founded the Alaska Independence Party. Inconveniently for Palin, that’s the very same secessionist party that her husband, Todd, belonged to for seven years and that she sent a shout-out to as Alaska governor earlier this year. (‘Keep up the good work,’ Palin told AIP members. ‘And God bless you.’)
AIP chairwoman Lynette Clark told me recently that Sarah Palin is her kind of gal. ‘She’s Alaskan to the bone … she sounds just like Joe Vogler.’
So who are these America-haters that the Palins are pallin’ around with? . . .”
” . . .Vogler wasn’t just a blowhard either. He put his secessionist ideas into action, working to build AIP membership to 20,000 — an impressive figure by Alaska standards — and to elect party member Walter Hickel as governor in 1990.
Vogler’s greatest moment of glory was to be his 1993 appearance before the United Nations to denounce United States ‘tyranny’ before the entire world and to demand Alaska’s freedom. The Alaska secessionist had persuaded the government of Iran to sponsor his anti-American harangue.
That’s right … Iran. The Islamic dictatorship. The taker of American hostages. The rogue nation that McCain and Palin have excoriated Obama for suggesting we diplomatically engage. That Iran. . . .”
3. Although the mainstream media have ignored or downplayed the connection, the AIP is very close to Sarah Palin and has helped to shape her political agenda. AIP luminary Mark Chryson has helped to forge the link between Palin and his party.
” . . . Though Chryson belongs to a fringe political party, one that advocates the secession of Alaska from the Union, and that organizes with other like-minded secessionist movements from Canada to the Deep South, he is not without peculiar influence in state politics, especially the rise of Sarah Palin. An obscure figure outside of Alaska, Chryson has been a political fixture in the hometown of the Republican vice-presidential nominee for over a decade. During the 1990s, when Chryson directed the AIP, he and another radical right-winger, Steve Stoll, played a quiet but pivotal role in electing Palin as mayor of Wasilla and shaping her political agenda afterward. Both Stoll and Chryson not only contributed to Palin’s campaign financially, they played major behind-the-scenes roles in the Palin camp before, during and after her victory. . . .”
4. Chryson and the AIP network with White Supremacist and neo-Confederate organizations. The AIP’s web site links to many of the site maintained by groups supported by the UNPO (as well as the UNPO’s website), such as the Tibetan, Hawaiian, Lakota and Mongolian independence movements. Note that the UNPO–represented as a progressive, egalitarian organization–supports the rights of minority peoples in order to weaken and breakup larger states that might prove a rival to the Underground Reich. The possibility that an economically and politically devastated U.S. might disintegrate is one to seriously consider. Certainly, having a major-party Vice-Presidential candidate affiliated with an organization such as the AIP is very disturbing. Note that the AIP’s website has a quote from Confederate President Jefferson Davis at the top of its front page. It is also significant that the Constitution Party, which counts the AIP as its Alaskan affiliate, ran Michael Peroutka for President. Peroutka is a member of the League of the South and his candidacy was endorsed by Richard Butler, the now-deceased head of the Aryan Nations.
” . . . Yet Chryson maintains that his party remains committed to full independence. “The Alaskan Independence Party has got links to almost every independence-minded movement in the world,” Chryson exclaimed. “And Alaska is not the only place that’s about separation. There’s at least 30 different states that are talking about some type of separation from the United States.”
This has meant rubbing shoulders and forging alliances with outright white supremacists and far-right theocrats, particularly those who dominate the proceedings at such gatherings as the North American Secessionist conventions, which AIP delegates have attended in recent years. The AIP’s affiliation with neo-Confederate organizations is motivated as much by ideological affinity as by organizational convenience. Indeed, Chryson makes no secret of his sympathy for the Lost Cause. ‘Should the Confederate states have been allowed to separate and go their peaceful ways?’ Chryson asked rhetorically. “Yes. The War of Northern Aggression, or the Civil War, or the War Between the States — however you want to refer to it — was not about slavery, it was about states’ rights.”
Another far-right organization with whom the AIP has long been aligned is Howard Phillips’ militia-minded Constitution Party. The AIP has been listed as the Constitution Party’s state affiliate since the late 1990s, and it has endorsed the Constitution Party’s presidential candidates (Michael Peroutka and Chuck Baldwin) in the past two elections.
The Constitution Party boasts an openly theocratic platform that reads, “It is our goal to limit the federal government to its delegated, enumerated, Constitutional functions and to restore American jurisprudence to its original Biblical common-law foundations.” In its 1990s incarnation as the U.S. Taxpayers Party, it was on the front lines in promoting the “militia” movement, and a significant portion of its membership comprises former and current militia members.
At its 1992 convention, the AIP hosted both Phillips — the USTP’s presidential candidate — and militia-movement leader Col. James ‘Bo’ Gritz, who was campaigning for president under the banner of the far-right Populist Party. According to Chryson, AIP regulars heavily supported Gritz, but the party deferred to Phillips’ presence and issued no official endorsements. . . .”
5. In her speech at the Republican convention, Palin quoted Westbrook Pegler, a reactionary columnist viewed as a fascist by many, including George Seldes. (See below)
” . . . ‘We grow good people in our small towns,’ Palin said, quoting someone identified only as a writer, ‘with honesty and sincerity and dignity.’ That ‘writer,’ Wall Street Journal columnist Thomas Frank notes, is a man named Westbrook Pegler. You have probably never heard of him, but he was a very popular and very right-wing columnist from the first half of the 20th century. How right-wing? He openly wished for the assassination of Franklin Roosevelt, for one. . . .”
5. Author George Seldes devoted a chapter of his 1943 opus Facts and Fascism to Pegler.
“It is the opinion of many persons and organizations that one of the most widely known and read newspaper columnists, Westbrook Pegler, is aiding the Axis rather than the United States in this war; it is a fact that the New York Newspaper Guild, the organization of thousands of Pegler’s colleagues, so stated when it sent President Roosevelt a documentation of twelve instances out of Pegler’s writings.
It is a fact that newspapers, columnists, radio orators and others who form public opinion have served the Axis propaganda. It is also true that too frequently those who know and who make these charges do not name names. They, therefore, emasculate their own words.
For example, here is the director of the U.S. Conciliation Service, Dr. John S. Steelman, who states:
‘Careless recital of the dramatic side of strikes in the press and on the screen and over the radio has given too many people the impression that our war efforts are being held up in a serious way because of willful strife in a major part of American industry. This is a dangerous lie that serves the purpose of the Axis, but serves no good end among us.’ Dr. Steelman knows the culprits, but he is not in a position to name them. He is aware that Kaltenborn on the radio, Pegler in the newspapers, the entire Hearst and Howard press are those guilty of careless recitals about strikes.
On the other hand here is the statement issued by the conference directors of C.I.O. editors (Washington, April 11, 1942):
‘Labor has been subjected to an infamous campaign of misrepresentation, for the purpose of cutting wages, destroying union organization, and in advancing the profits of special interest groups. Most of the daily press has joined in this campaign, together with such radio commentators as H.V. Kaltenborn and others of his type. This anti-labor propaganda campaign, if not directly inspired by Axis agents and American appeasers, at any rate plays their game by striking at national unity and undermining labor morale.’
Again, the official organ of the National Maritime Union, The Pilot, states that ‘Pegler’s sales talk has the made-in-Berlin label.’ . . .”
7. Whether relevant or not, it is interesting (and possibly significant) that Palin is a native of Sandpoint, Idaho, an epicenter of the white supremacist community.
8. It is also worth contemplating what the Nazi elements in the Bush administration will do with an African-American President.
Discussion
4 comments for “FTR #652 Palin by Comparison”
Just FYI, if you’ve ever been exposed to the herbicide paraquat (I’m looking at you, hippies) and you have a family history of Parkinson’s disease, new research suggests that you may want to take extra precautions to avoid any sort of head trauma. It appears that head trauma and paraquat exposure triple the likelihood of developing Parkinson’s.
These findings, of course, also means you may need to avoid any US political news for the next, oh, four years or so. It’s going to bea bumpy ride.
Political Research AssociatesIf Democracy is a Crime Under Religious Right’s Biblical Law, What is the Punishment?
By Frederick Clarkson, on August 21, 2014
The League of the South, best known for its advocacy of white supremacy and the secession of Southern states in the name of Southern Independence, has another less well known dimension: The advocacy of Old Testament notions of the “law of the Bible” as the standard for contemporary civil law— including the prescribed criminal punishments for non-believers.
The League and its leaders are not exactly household names, although New York Times best-selling author Thomas Woods is an unapologetic founding member and close confederate of Ron Paul. But they overlap with the more extreme elements of anti-LGBTQ and anti-choice movements, and led by elements of the virulently theocratic theological strain, Christian Reconstructionism.
Contemporary Christian Right leaders have focused on matters of sexuality, notably homosexuality and women’s sexual and reproductive health and rights (SRHR). But it wasn’t until the notorious “kill the gays bill” in Uganda, where the notion of the death penalty was formally raised. But if the Old Testament Biblical Laws that are foundational to the contemporary Christian Right are to be the standard, then it stands to reason that the biblically prescribed punishments would, in their view, fit the crimes. If that is so, then many of us who think that democracy and religious pluralism are good things may, come the theocracy, find ourselves guilty of capital crimes.
League of the South president Michael Hill laid it out in stark terms in a recent column.
In a piece titled The League and Theocracy, Hill claims that theocracy is what God wants for us, and that anyone who says otherwise is committing treason against God’s Laws. While this is serious enough coming from an organizational leader, in the wake of the Hobby Lobby decision of the U.S. Supreme Court, it is worth considering that such views maybe more common than usually meet the eye.
Hill argues that religious pluralism is an affront to “the law of the Bible,” and anything other than a theocratic approach to Christianity and government is, he declares, “watered-down, emasculated, wimpy, liberal-sotted ‘Christianity-lite’.” A theocracy, he insists, “is a government ruled by the triune God of the Bible: Father, Son, and Holy Spirit. More precisely, it is a government whose code of law is firmly grounded in the law of the Bible.”
…
In light of this, Hill’s argument may sound arcane and his reasoning circular—but it is nevertheless worth understanding because Hill is far from alone in such views, and arguably versions of his thesis are part of the driving ideology of the wider leadership of the Christian Right and their agenda.
Hill continues:
“Simply put, locate the source of law for a society and you have found its god. In a democracy, for instance, the people serve as god – the ultimate source of sovereignty…. All societies are theocracies, whether they realize it or not. But there is a major difference in what traditional Christians (regardless of denomination), on the one hand, and pagans, on the other, have believed and that is this: a society that is not explicitly Christian is a theocracy under the sway of a false god. The false god of the modern American Empire is the god who demands tolerance and pluralism…”
In this way, Hill identifies democracy as idolatry (worship of a false god) or apostasy (abandonment of the faith.) Either way, as Hill understands it, these will not be matters of religious but civil law. And they may well be capital offenses. These, along with more than 30 others, were enumerated by the leading theocratic theologian of the 20thcentury, R. J. Rushdoony—an American whose work has been profoundly influential in catalyzing the contemporary Christian Right. Beyond such crimes as murder and kidnapping, death penalties apply, according to Rushdoony and the influential Christian Reconstructionist movement he launched, are mostly related to religion and sexuality. In addition to idolatry and apostasy, there is blasphemy, and the propagation of false religions, and specifically, astrology and witchcraft. Crimes related to sexuality include adultery, homosexuality, and premarital sex (for women only).
In his foundational tome, Institutes of Biblical Law, Rushdoony called idolatry, “treason” against “the one true God.”
These things said, there are many contemporary theocrats who do not follow Rushdoony down every detail of what ought to be the basis of a criminal code in the U.S. and everywhere else in the world. But since Rushdoony was the first to systematize what he calls “Biblical Law,” everyone interested in the topic measures where they are stand in relation to him. (Biblically approved methods of execution, according to Reconstructionists, include stoning, hanging and “the sword.”) How the law would be implemented, would depend entirely on the interpretation of the Bible, and the political predilections of the leading faction at the time.
Hill writes: “There is really only the choice between pagan law and Christian law and nothing else. There is no neutral position where one can comfortably sit. The God of the Bible specifically forbids pluralism (“Thou shalt have no other gods before me,” Ex 20:3).” Hill couldn’t be clearer that in his view that religious pluralism under the law, is idolatry. “The simple choice that lies before nations,” he concludes, “is either pluralism or faithful obedience to God’s word, and the two are mutually exclusive.”
Like his fellow members of the League of the South, Maryland Republican candidates Michael Peroutka and David Whitney, Hill argues that if governmental leaders fall out of synch with Biblical Law—“If they rule unfaithfully, and thus tyrannically, they are illegitimate and their decrees have no authority.” Peroutka and Whitney have previously made this point regarding the Maryland legislature’s endorsement of marriage equality.
So this is the problem for contemporary Americans wrangling with the definition of religious liberty. Is religious liberty to be reserved only for Christians of the right sort, as defined by the likes of Hill, Peroutka and Whitney (or other leaders of the Religious Right and the Catholic Bishops)? Or is religious liberty something that belongs to all citizens without regard to their stated religious or non-religious identity at any particular time?
Such questions have, in centuries past, been the stuff of religious wars. And some Christian Right leaders are coming to see violence as an inevitable result of contemporary religious and political tensions on these matters. The question of whether, or to what extent, opponents of contraception and abortion get to define the standard for religious liberty in these matters was not settled by the Hobby Lobby decision. And who gets to be the arbiter of religious liberty on marriage equality is also deeply contentious, in light of the spread of the Religious Freedom Restoration Act, which has been proposed in several states and signed into law in Mississippi and allows businesses to discriminate against LGTBQ employees and customers if their personal faith does not approve of homosexuality. The legislation is modeled on a bill which was originally crafted in large part by the Alliance Defending Freedom (ADF, formerly known as the Alliance Defense Fund), and the Arizona political affiliate of Focus on the Family. But in North Carolina, the mainline protestant United Church of Christ (UCC) has filed a federal lawsuit against the state’s constitutional amendment banning the performance of same-sex marriage ceremonies. The UCC argues that this law violates the religious liberty of their church, its clergy and its members. The million member denomination has officially recognized same-sex marriages since 2005. As it happens, a recent federal court ruling in Virginia overturning a similar amendment, may also apply to North Carolina which is located in the same federal court jurisdiction. At this writing,, it appears that will be the case.
But even as the courts are recognizing marriage equality and the debate over religious liberty continues, the view of doctrinaire theocrats that governments and government officials and others who do not comport with God’s Law as they understand it, remains unchanged. How they reconcile their views with the contemporary struggles over the law is also a struggle whose outcome remains to be seen. However, it is worth not losing sight of the fact that for some, there is no answer but armed resistance. Hill has called not only for the South to rise again, but to lead the resurrection with vigilante death squads targeting government officials, journalists and others who do not comport with their particular religious and political views.
We have become almost accustomed to hearing declarations from the likes of these men justifying vigilante violence against everyone from abortion providers to government officials, said to be tyrants. But it is helpful to remind ourselves that the society that they envision is not only based on Biblical Laws, but on Biblical punishments. And those who advocate for religious liberty, and if the theocrats ever achieve their desired governmental control, are likely to find themselves charged with a variety of crimes, and punished accordingly.
House Majority Whip Steve Scalise, a vehement opponent of gay marriage, issued a statement of support for Alabama Chief Justice Roy Moore in his battle with the Federal government over gay marriage, calling for like-minded citizens to get ready to stand with Moore. Oh wait, that wasn’t Scalise’s press office, but a different like-minded organization making that ‘call to arms’. Easy mistake:
“The Mississippi Klan salutes Alabama’s chief justice Roy Moore, for refusing to bow to the yoke of Federal tyranny. The Feds have no authority over individual States marriage laws,” the United Dixie White Knights’ Imperial Wizard, Brent Waller wrote, wrote in a post on the group’s website and on the white supremacist forum Stormfront.
“We call upon all Klansman and White Southern Nationalist to help in the massive protest’s coming, Not by wearing your colors, but by joining in with the Christian community’s protests that are surly coming against tyrannical Federal judges,” Waller said.
However, the group realized its support may have made an implication about its relationship with Moore and clarified that he is not a member of the Klu Klux Klan.
“The UDWK has read many posts in the media put out mostly by Homosexual reporters and Atheists that claim Alabama Chief Justice Roy Moore is a member of the KKK, and activly sought and gained our support. To answer in short, these are lies and typical of that crowd. We simply agree with anyone, who stands against Tyranny, stands for the word of God, and believes in the American Constitution as it was originally written,” Waller wrote on the group’s website.
“We regret any problems our post may have caused the Chief Justice or his Family. Roy Moore is not a Klansman, but he is a man, who believes in God,” Waller continued. “This Country is sick of the spineless people who put out such lies. We need more Justices who do not see the Constitution as a living breathing document to be shreded (sic) and Amended at will.”
…
“We call upon all Klansman and White Southern Nationalist to help in the massive protest’s coming, Not by wearing your colors, but by joining in with the Christian community’s protests that are surly coming against tyrannical Federal judges,” Waller said.
Right Wing WatchRoy Moore’s Showdown With The Courts Over Gay Marriage Tied To The Work Of Neo-Confederate Leader
Submitted by Brian Tashman on Thursday, 2/12/2015 1:30 pm
Alabama Supreme Court Chief Justice Roy Moore threw his state into turmoil this week when he ordered probate judges to defy a federal judge’s ruling striking down the state’s ban on same-sex marriage and refuse to issue marriage licenses to gay and lesbian couples. Moore, who has a history of making extreme anti-gay statements, insists that the federal judge is the one who is really breaking the law since she violateddivinelaw by ruling for marriage equality.
Moore’s call for statewide defiance of the federal judiciary’s “tyranny” stems from a belief that the Constitution was made to protect biblical commandments, so that anything that goes against his personal interpretation of the Bible is therefore in violation of the Constitution.
Moore shares that belief with a powerful ally: Michael Peroutka, a neo-Confederate activist who is also one of the most influential behind-the-scenes figures in the Religious Right’s reimagining of American law.
Peroutka, who once held a leadership position in the neo-Confederate League of the South and remained a member of the group until it hampered his run for a local office in Maryland last year, promotes this theocratic view of the law through his group the Institute on the Constitution. Speaking at an event at the Institute in 2011, Moore gushed that Peroutka would help lead America to a “glorious triumph” over the federal government’s “tyranny.”
After Moore was removed from his original position on Alabama’s high court in 2003 for defying a federal court order to remove a monument of the Ten Commandments from the state judicial building, Peroutka paid for the ousted judge to go on a national speaking tour to build support for his cause. He also funded a group that held rallies in support of Moore.
Over the nine years, Peroutka contributed over a quarter of a million dollars to two groups founded by Moore, the Foundation for Moral Law (which is now run by Moore’s wife Kayla) and the now-defunct Coalition to Restore America.
In 2004, the far-right Constitution Party tried to recruit Moore to run for president on its ticket. When he declined, Peroutka stepped in to run in his place.
This neo-Confederate leader helped to lay the ideological groundwork for Moore’s current standoff with the federal courts, a standoff which many commentators have compared to Alabama Gov. George Wallace’s decision to defy federal law on desegregation.
Peroutka said last year that such rulings would “coerce” state officials to “declare that which is sinful and immoral” to be “valid and right,” even forcing them to “participate in it.” Such “evil” decisions, according to Peroutka, must be “resisted at every level of government, even the lower levels of government, most especially the lower levels of government,” since local governments are the true “protectors against those who would force these things on us tyrannically from above.”
For example, after a federal judge struck down Kentucky’s ban on same-sex marriage last year, Peroutka insisted that Sen. Rand Paul move to impeach the judge who made the decision, defund the court, and press for his state to defy the ruling: “He should use every influence he has in Kentucky to have people not obey this; the Kentucky legislature, the Kentucky courts, should not obey this, this is not lawful.”
Peroutka also believes that local officials should defy their state legislatures on issues like marriage equality. After Maryland’s general assembly voted in 2012 to legalize same-sex marriage in the state, Peroutka declared that the assembly’s decision to “violate God’s laws” effectively invalidated its legal authority, since any law that contradicts divine law does “not constitute a law – even if it were enacted and signed.”
Using an argument similar to the one Moore is now making in Alabama, Peroutka said that lower-level officials could ignore not only the marriage equality law but any law passed by the state’s general assembly, since it had invalidated itself by breaking biblical decrees: “Is it possible that those who are sworn to uphold the law, such as police and sheriffs and judges and prosecutors, may soon come to the conclusion that the enactments of this body should be ignored because they are based not in law, but in lawlessness?”
“I always go back to these two standards: What does God say and what does the Constitution say?” Peroutka explained in 2013.
He added that the United States will have a small, limited government as long as it adheres to biblical standards. But he believes that the Union’s victory in the Civil War — or as he calls it, “The War Between the States” — enabled the federal government to greatly expand its powers, thus undermining the authority of biblical law and leading to such evils as same-sex marriage.
“Ever since then, there’s been this huge black hole of centralized power that’s formed in Washington, D.C.” he said. “People sometimes talk about ‘The War Between the States’ as being about the issue of slavery. I believe that history is written by the winners, it wasn’t about that at all. What it was about was consolidating power into the hands of a few people.”
“[T]he real effect of the war and the Reconstruction after the war was to take the very foundation of our understanding of our rights away from us, that is to say that they come from God, and put them in the hands of men and say that they come from the Supreme Court or they come from the legislature or they come from the executive,” he added.
The end of the Civil War, Peroutka claims, produced an “evil anti-God, anti-Christian revolution” that led to a “tyrannical consolidation of power” in Washington, D.C., undermining the “biblical worldview that acknowledges Christ’s authority over all things.”
Peroutka also contends that the gay rights movement isn’t just “federalizing homosexuality” but “federalizing perversion,” even claiming that the federal government violated the Constitution by imposing civil rights laws on the states.
“[T]he so-called civil rights laws are not law,” he said in 2013. “They never should’ve been passed. They’re not law now, they weren’t law then. They aren’t law now because there is no such thing as a civil right.”
Since Peroutka believes “rights come from God” and not civil government, he argues that all civil rights laws are illegitimate since “the term ‘civil rights’ is kind of an oxymoron. There’s no ‘right’ in the sense of a permanent, fixed, thing that you have, that can be defended, if in fact it comes from the civil government.”
Now Moore is once again putting Peroutka’s words into action, threatening state judges who lawfully issue a marriage licenses to a same-sex couples. Because in the eyes of Moore and Peroutka, their personal reading of the Bible takes precedence over the law of the land.
The end of the Civil War, Peroutka claims, produced an “evil anti-God, anti-Christian revolution” that led to a “tyrannical consolidation of power” in Washington, D.C., undermining the “biblical worldview that acknowledges Christ’s authority over all things.”
…
“[T]he so-called civil rights laws are not law,” he said in 2013. “They never should’ve been passed. They’re not law now, they weren’t law then. They aren’t law now because there is no such thing as a civil right.”
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.