Main content

Alert message

American Foundation for the Blind Textbooks and Instructional Materials Solutions Forum

Paper Presented at March, 2001, CSUN, by Jim Allan.

Joint Technology Task Force - History 

Collaborative project between American Association of Publishers, Recording for the Blind and Dyslexic, and American Foundation for the Blind. Initial Meeting - June 15, 2000. 

Participants

American Association of Publishers, Recordings for the Blind and Dyslexic, American Foundation for the Blind, Duxbury Systems, American Printing House for the Blind, Library of Congress - National Library Service, Computer Applications Specialties Company, Texas Education Agency, Harcourt-Brace, McGraw Hill, Houghton Mifflin, Addison Wesley, Scholastic, National Braille Press, ESC 4 & 20, and others...

Goal

Ensure students with disabilities receive instructional materials (textbooks) at the same time as their non-disabled peers.

The goals of the new task force include analyzing the National Information Standards Organization (NISO) XML (Extensible Markup Language) file format to determine its suitability for converting textbook content into braille and other accessible formats and promoting and demonstrating to accessible book producers the efficiency and benefits of using publisher files in NISO XML format.

Problem

Original documents created in Word. Publishers develop presentation files in Quark (mostly) or Pagemaker. Changes are made until press time. These files not easily converted to usable braille translation files. Currently, files are converted into ASCII file. Publishers do not use ASCII in publishing process, hence conversion takes 60-90 days. Braille producers then take 60-90 days to convert ASCII (with proper formatting) into braille.

Publishers want to re-purpose content for different media streams: paper, web based, e-book, braille, etc. Quark/Pagemaker files are not suitable for this task (although software is beginning to change). 

Solutions (work to date)

Joint Technology Task Force formed to. Explore potential use of XML, specifically DIASY/NISO Digital Talking Book 3.0 Document Type Definition. Test using XML for production of files for braille translators in transition period. Publishers moving toward creating content in XML as day to day file format. Goal: publishers provide XML (DAISY/NISO)  file (can be created faster than ASCII file) which is easily usable by braille producers and students receive braille and digital talking books faster.

Why XML?

XML- eXtensible Markup Language. - allows creation of own special purpose tags. It is flexible and separates content from presentation. Allows for special tags, such as side bars, headings, etc. Allows creation of presentation rules for media output (large prints, synthetic speech, digital talking book, braille, etc.) from the same content.

Current Tasks

Conversion Processes Being Explored

  1. Utopia. The ideal is automated conversion from Quark to DTB 3.0 with minimal intervention.
    •  Not all Quark files are equal. K-12 books are "high format." All publishers use different tag sets, even within the same publishing house.
    • 20-60% of a book is imported art (i.e. Illustrator) with embedded text. Software is in development to extract text from image. Currently must be rekeyed. 
  2. Reality Version 1: paper to DTB 3.0 DTD - send of conversion house (usually off-shore) book is rekeyed to specified DTD.
    • no electronic file (seems to be cost effective and working) 
  3. Reality Version 2: electronic file conversion to DTB 3.0 DTD
    • extract text (ASCII),  then send off shore to be marked up in DTD. Must supply proofs to check print vs. XML structure created.
      • 20-60% of a book is imported art (i.e. Illustrator) with embedded text.  Image files sent off-shore for rekeying. 
    • Quark to RTF conversion. Send off-shore to have RTF converted to DTB 3.0 DTD. Used elementary science book, 2 chapters (60 pages), took 30 days. Communication, first time process, understanding DTD and task took time. 
  4. Reality Version 3: transformation of publisher DTD to DTB 3.0 DTD
    • no report yet.
    • Possible problem: style sheet generated content. 

Current Testing Process

  1. Publishers send files to RFBD - validity and well-formed check
  2. RFBD check against paper copy (feedback loop with publishers)
    • is tag set (DTD) complete and correct
    • are conversion houses tagging appropriately
  3. RFBD send file to braille producers
  4. Braille producers using Duxbury and Braille2000 (in development) for braille formatting of file. 
  5. Braille producers check braille output against paper copy. Translation software translates text easily, interprets XML for appropriate formatting. Translation software is very dependent upon how well and which conventions were used in the XML markup. If tagging was done correctly and tagging is correctly interpreted by translation software, then braille formatting is pretty accurate. (feedback loop with RFBD and braille translation software manufacturers)
    • is braille translation software making proper use of tags
    • are tags used consistently
    • tweaking rules for formatting.

Status

  1. Not as easy to do as first conceived in June 2000
  2. Conversion houses not as fast as stated (new process for all involved)
  3. Cost determination for conversion still uncertain (process still too new)
  4. STILL HOPE that technology will ease/speed production process of files and braille
  5. Developing matrix of subjects vs. grade levels vs. conversion process (no time line)
  6. ALL COMMITTED to more testing, more feedback, more learning, more... 
    • already  updated the DTBook3 DTD based on things we learned in early in-house tests with the Duxbury import facility. Notably, DTD now allows entries in a table of contents to be subdivided into that entry's components, i.e., the entry itself ("chapter 12") and the corresponding page number. That way, the braille software can format each entry properly, inserting guide dots, etc. 
    • Added an attribute called "showin." that can be used to mark sections of a document (a transcriber's note, for example) and specify that it show only in a braille version. A different wording of the note could be tagged to appear only in a large print edition.

Examples of textbook pages

Resources