| { | |
| "paper_id": "1997", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T10:30:14.000249Z" | |
| }, | |
| "title": "Localizing Canon's User Documentation in Europe", | |
| "authors": [ | |
| { | |
| "first": "Tim", | |
| "middle": [], | |
| "last": "O'donoghue", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Canon Research Centre Europe Ltd", | |
| "location": { | |
| "addrLine": "Bovenkerkerweg 59-61 20 Alan Turing Road", | |
| "postCode": "1185 XB, GU2 5YF", | |
| "settlement": "Amstelveen Guildford", | |
| "country": "The Netherlands UK" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Europa", | |
| "middle": [], | |
| "last": "Nv", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Canon Research Centre Europe Ltd", | |
| "location": { | |
| "addrLine": "Bovenkerkerweg 59-61 20 Alan Turing Road", | |
| "postCode": "1185 XB, GU2 5YF", | |
| "settlement": "Amstelveen Guildford", | |
| "country": "The Netherlands UK" | |
| } | |
| }, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "Canon is a company that produces a wide range of products in the areas of imaging, information and industry. Canon's innovative products are marketed the world over and user documentation is a key component in each Canon product. Canon Europa NV is one of Canon's regional headquarters. Even though the name implies responsibility for Europe alone, Canon Europa is also responsible for Eastern Europe, Russia, Africa and the Middle East. One of Canon Europa's responsibilities is the localization of user documentation. This is quite a responsibility when you consider the number of languages spoken in Canon Europa's territories (over 25 major languages) and the number of products which Canon produces. Business machines-copiers, faxes, printers, filing systems, etc.-account for a large proportion of Canon's sales, and for these products it is often necessary to produce user documentation in a number of different formats, both print and electronic, depending upon the product in question and the users needs. Canon Research Centre Europe, one of Canon's global R&D centres, has been working with Canon Europa over the past 5 years to develop an efficient localization and production process, in which: \u2022 SGML is used as a publishing base to enable efficient and flexible publication; \u2022 machine assisted translation, in the form of a translation memory, and translators' tools are used to enable efficient localization. Dr. Tim O'Donoghue Dr. Tim O'Donoghue leads the Solutions Applications Services group at Canon Research Centre Europe. He is a consultant to Canon Europa in the area of document localization and production.", | |
| "pdf_parse": { | |
| "paper_id": "1997", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "Canon is a company that produces a wide range of products in the areas of imaging, information and industry. Canon's innovative products are marketed the world over and user documentation is a key component in each Canon product. Canon Europa NV is one of Canon's regional headquarters. Even though the name implies responsibility for Europe alone, Canon Europa is also responsible for Eastern Europe, Russia, Africa and the Middle East. One of Canon Europa's responsibilities is the localization of user documentation. This is quite a responsibility when you consider the number of languages spoken in Canon Europa's territories (over 25 major languages) and the number of products which Canon produces. Business machines-copiers, faxes, printers, filing systems, etc.-account for a large proportion of Canon's sales, and for these products it is often necessary to produce user documentation in a number of different formats, both print and electronic, depending upon the product in question and the users needs. Canon Research Centre Europe, one of Canon's global R&D centres, has been working with Canon Europa over the past 5 years to develop an efficient localization and production process, in which: \u2022 SGML is used as a publishing base to enable efficient and flexible publication; \u2022 machine assisted translation, in the form of a translation memory, and translators' tools are used to enable efficient localization. Dr. Tim O'Donoghue Dr. Tim O'Donoghue leads the Solutions Applications Services group at Canon Research Centre Europe. He is a consultant to Canon Europa in the area of document localization and production.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "Canon is a company that produces a wide range of products in the areas of imaging, information and industry. Canon's innovative products are marketed the world over and user documentation is a key component in each Canon product.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Localizing Canon's User Documentation in Europe", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Canon Europa NV -one of Canon's regional headquarters -is responsible for Europe (Western, Central and Eastern), Russia, Africa and the Middle East. One responsibility is product localisation, which includes the translation of user documentation. This is quite a task when you consider the number of languages spoken in Canon Europa's territories -over 25 major languages -and the number of products which Canon produces.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Localizing Canon's User Documentation in Europe", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Business machines -copiers, faxes, printers, filing systems, etc. -account for a large proportion of Canon's sales, and for these products it is often necessary to produce user documentation in a number of different formats, both print and electronic, depending upon the product in question and the users needs.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Localizing Canon's User Documentation in Europe", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "To enable this to be done effectively, Canon Europa has:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Localizing Canon's User Documentation in Europe", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 implemented a SGML [1] -based process for flexibility and manageability", | |
| "cite_spans": [ | |
| { | |
| "start": 21, | |
| "end": 24, | |
| "text": "[1]", | |
| "ref_id": "BIBREF0" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Localizing Canon's User Documentation in Europe", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 developed and deployed a translation memory system -Adroit -to assist the localisation process EAMT Workshop, Copenhagen, May 1997 Figure 1 gives an overview of the process by which Canon Europa (abbreviated CENV in the diagram) produces localised user documentation for business machines:", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 133, | |
| "end": 141, | |
| "text": "Figure 1", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Localizing Canon's User Documentation in Europe", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "1) The original user documentation is authored by Canon Inc., Canon Europa's parent company based in Japan. The Japanese original is then translated into English by Canon Inc. and it is this English version which is delivered to Canon Europa.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Figure 1: Overview of Production Process", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "2) The delivered material is usually a DTP document, typically PageMaker or Quark XPress, depending upon the product group.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Figure 1: Overview of Production Process", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "3) The first step of Canon Europa's process is to de-purpose this DTP document. De-purposing is the process by which the DTP document is transformed into a publication-independent form.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Figure 1: Overview of Production Process", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The publication-independent form chosen by Canon Europa is SGML. The specific SGML application used is TranScribe-III, a DTD developed specifically for Canon's business machine user documentation.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "4)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The English SGML document is localised into the necessary language versions for the areas in which the product is to be marketed. This typically ranges from 10 to 20 different language versions.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "5)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The result of localisation is a set of SGML documents, one for each of the target languages.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "6)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "For each of the SGML documents, a publication process is applied to convert the SGML document into its final published form. ", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "7)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Upon reception of the DTP document from Canon Inc., Canon Europa's first step is to de-purpose this material to ensure a solid foundation for the subsequent stages of production. De-purposing can be a very expensive task, but Canon Europa's experience has shown that effort invested early in the production process is recovered with \"interest added\" in the later localisation and publishing stages. This solid foundation simplifies the management of the localisation process and allows increased automation of the publishing processes.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "One example of where SGML has been used to simplify the localisation process is in managing the consistency between hardware localisation and the translation of user documentation. User documentation cannot be translated in isolation, that is, without reference to the complete product kit, of which it is a part.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The complete localisation of a product requires the localisation of its various components, including:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 The product hardware. E.g. the fax machine, including the various user interface components (buttons, display messages on LCD panels, etc.)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 The product software. E.g. the printer driver for a Bubble Jet printer.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 The user documentation, which documents both the hardware and the software.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Importantly, the localisation of these various components must be consistent.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "In the SGML document, (hardware) display messages are identified by the appropriate SGML tags. These tagged messages are then used to query a separate database in which the previously translated display messages are stored (as used to localise the hardware), and the correct translated message is interpolated into the translated user document. Using the same display message database for both the production of localised hardware and documentation ensures consistency between these components in the overall localised product.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "A second example of where SGML has simplified the localisation process is in the handling of non-ASCII character data. Canon's TranScribe-III SGML application uses the SPREAD [2] Entities -which themselves are derived from Unicode [3] character set -to represent all non-ASCII characters in the SGML documents. And Unicode mapping tablesavailable from the Unicode Consortium [4] -are used to enable import and export to data which is encoded with respect to a given character set or code page.", | |
| "cite_spans": [ | |
| { | |
| "start": 175, | |
| "end": 178, | |
| "text": "[2]", | |
| "ref_id": "BIBREF1" | |
| }, | |
| { | |
| "start": 231, | |
| "end": 234, | |
| "text": "[3]", | |
| "ref_id": "BIBREF2" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Why SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Canon uses a machine-assisted translation system as part of its localisation process for user documentation. This system, named Adroit, was developed by Canon Research Centre Europe for Canon Europa. Adroit is based on the translation memory concept whereby a database of translations is used to assist in the translation of new documentation. New documents are compared against older documents, stored in the database, and only those areas identified as \"new\" need be manually translated. The areas identified as \"old\" can be translated using older previouslytranslated documents.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Adroit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The Adroit database is built from previously translated user documentation, extracting translations from pairs of documents which are mutual translations of one another. The extraction of translations relies upon the structure of the documents being invariant between the pairs of documents. This necessary structure is provided by the logical structure imposed by the TranScribe-III SGML application and protected during translation (by only sending textual data to translators, not SGML structural data). \u2022 A hierarchically structured translation database, rather than a single large database into which all translations are deposited. The database is structured to reflect the different types of documents from which the translations were obtained. For example:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Adroit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 copier/black+white/NP6112", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Development of Adroit started at Canon", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 fax/multi-function/MPC30/hardware", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Development of Adroit started at Canon", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 fax/multi-function/MPC30/software", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Development of Adroit started at Canon", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 The ability to quickly compare a new document against the hierarchical database to determine which segments of the database should be used for memory translation.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Development of Adroit started at Canon", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The output of Adroit is a partially translated document, that is, a document which contains sentences both in the source language (English) and the target language (as obtained from the translation memory). This document is sent to a translator.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Adroit only sends sentential data (#PCDATA -in SGML parlance -and those minor elements which can appear in sentences, such as font changes etc.) and a majority of the SGML structure is retained by Adroit. When returned by the translator, the fully translated text is merged with the SGML structure of the original English document to yield a fully translated TranScribe-III document instance.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "This process of separating SGML structure from text content in the translation phase is convenient both for Canon and its translators:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 It ensures that translators cannot disturb the SGML structure, which is essential for correct updating of the Adroit translation database.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 It allows translators to concentrate on translating the text without having to locate it in the often complex SGML structure.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The translator performs three types of tasks when working on a document sent from Adroit: 1) Translating, from scratch, any text in the source language.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "2) Proof reading the exact translations supplied by Adroit.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "EAMT Workshop, Copenhagen, May 1997", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "3) Checking, and possibly modifying, the approximate translations supplied by Adroit.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "It is possible, in the case of 2), for the translator to modify an exact translation supplied by Adroit. However, this seldom occurs and when it does, it is usually justifiable.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Translators use the Star's Transit system which provides them with a suitable interface to perform these and other localisation related tasks. A screen shot of the Transit system is shown in Figure 2 : \u2022 The background colour is used to indicate the different types of text: grey for the current sentence, white for a sentence exactly translated by Adroit, blue for an approximate translation and yellow for untranslated text.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 191, | |
| "end": 199, | |
| "text": "Figure 2", | |
| "ref_id": "FIGREF0" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "\u2022 Elements, and other protected areas are coloured blue. These cannot be edited by the translator.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Transit does contain its own translation memory facility, but this is not directly used since this function is provided by Adroit. Within Canon's localisation workflow, Transit's role is primarily as an editing tool, allowing translators to work with the output of Adroit. However, Transit's translation memory facility is used by the translators who build their own memories. These local memories allow the translator to handle repetition in the current translation job as well as providing a memory for whatever other translation work they undertake.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Future versions of Adroit will support other translators environments -Trados' Translators Workbench for example -so that translators can use their chosen tool rather than one specified by Canon. immediately compatible; they have to be glued together in some way. No two programs speak the same language; each has its own API or data format.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The ideal situation for integrators is where the various componentsauthoring tools, translation memories, translation environments,... -can be treated as modular objects, each sharing a standard API and a standard format for data interchange. How long until this pipe dream is reality? Who knows? The development of localisation-aware authoring environments will be crucial. There are already quite promising industry initiatives such as OpenTag[5], in which companies from different areas of the localisation and documentation industry are developing an XML [6] based standard that will support open data encoding methods during the localisation process. If such initiatives are successful then designing filters and 'system integration glue' will be a thing of the past, and designing localisation workflows will be more like playing with Lego. This will result in systems which are easier to implement and more reliable in operation.", | |
| "cite_spans": [ | |
| { | |
| "start": 559, | |
| "end": 562, | |
| "text": "[6]", | |
| "ref_id": "BIBREF3" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Transit", | |
| "sec_num": null | |
| } | |
| ], | |
| "back_matter": [ | |
| { | |
| "text": "While there are clear benefits -both in localisation and publication -from using SGML, there is also a cost: conversion. The current de-purposing processes are highly dependent upon the DTP material received from Canon Inc. and often a great deal of manual effort is required in this phase. However, this effort can be reduced: If the original DTP material is authored in a structured manner, e.g. by using a standard set of styles, then it can be -partially at least -mapped via a filter onto an equivalent SGML structure.Canon Europa is aiming to increase the amount of automation in this depurposing phase by influencing upstream document production at Canon Inc. Our short term aim is for Canon Inc. to implement a set of DTP authoring guidelines that will assist the automatic conversion from DTP formats to SGML or other document formats. A longer term goal is of course to push SGML further upstream, eventually to the original Japanese authoring environment, but there are many issues which need to be addressed before this can happen.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "At what cost SGML?", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Within the workflow of Canon's document localisation and production process there are a number of discrete components which must be glued together so that they can work in harmony:\u2022 Adroit, the translation memory This integration is typically enabled by creating various conversion programs -filters -which covert one component's data format into a format suitable for another component. Within Canon's workflow there are many such filters, a prime example being the two filters that enable Adroit to 'export data to' and 'import data from' Transit. If Adroit were to support a second system for translators -Translators Workbench for examplethen this would require the creation of a further two filters to handle Trados' data format.One of they problems facing system designers and integrators is that the many components which constitute a complete localisation system are not", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Woes of System Integration", | |
| "sec_num": null | |
| } | |
| ], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "The Standard Generalised Markup Language", | |
| "authors": [], | |
| "year": 1986, | |
| "venue": "", | |
| "volume": "8879", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "The Standard Generalised Markup Language. ISO 8879:1986.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "for details of the SPREAD (Standardisation Project Regarding East Asian Documents) Public Entity Set", | |
| "authors": [], | |
| "year": null, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "See http://www.allette.corn.au/sgml/ercs/allent.htm for details of the SPREAD (Standardisation Project Regarding East Asian Documents) Public Entity Set.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "The Unicode character encoding standard, an international character code for information processing", | |
| "authors": [], | |
| "year": 1993, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "10646--10647", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "The Unicode character encoding standard, an international character code for information processing. ISO/IEC 10646-1:1993.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Extensible Markup Language is a data format for structured document interchange on the Web", | |
| "authors": [], | |
| "year": null, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Extensible Markup Language is a data format for structured document interchange on the Web. See http://www.w3.org/XML/ for details.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "num": null, | |
| "uris": null, | |
| "type_str": "figure", | |
| "text": "Screen shot of the Translator's Interface \u2022 The original English is shown in the top pane with the partial translation produced by Adroit in the bottom." | |
| }, | |
| "TABREF0": { | |
| "html": null, | |
| "content": "<table><tr><td>currently produced by this process.</td></tr><tr><td>8) Print, PDF (Adobe's Portable Document Format) and HTML</td></tr><tr><td>(HyperText Markup Language) are the three publication formats</td></tr></table>", | |
| "text": "", | |
| "num": null, | |
| "type_str": "table" | |
| } | |
| } | |
| } | |
| } |