SDF and Blind Web Access (fwd)

[Hans Zoebelein <hzo goldfish cube net>]

---------- Forwarded message ----------
Date: Fri, 9 Jul 1999 10:31:20 +0930 (CST)
From: Dan Shearer <dan shearer org>
To: mgorse wpi edu
Cc: hzo leb net
Subject: SDF and Blind Web Access

G'day,

I have recently been doing quite a bit of work with the Simple Document
Format, which delivers right now what XML might deliver one day for the
boring bulk of documents. Then I hit on the idea of using it for disabled
users, and I have quite some interest in vision-impaired people.

I sent the following to the directors of various blind societies who have
R&D budgets. I expect it to come to nothing, but thought I would propose
it anyway.

I don't subscribe to any of the blinux lists, but if you wouldn't mind
passing this on to whoever might be interested I'll be grateful.

Regards,

Dan Shearer
dan shearer org

==================


Development Project Suggestion      Thursday 1 July 1999
------------------------------

This is a very brief outline only, and should not be taken as any kind of
formal proposal, technically exact - or even checked for spelling! Think
of it as a concept sketch, a precursor to a specification for a working
model should there be sufficient interest.

Preamble
--------

  The World Wide Web is built around the exchange of documents -
  frequently HTML documents - over HTTP connnections. The difficulty 
  of automatically extracting text from HTML and processing it for blind
  users continues to increase. This is because the search for excellence
  in presentation has resulted in closer coupling of the content and the
  presentation description. It has not been helped by the tangle of
  competing interests that have developed many different ways to solve
  presentation problems, resulting in a very large team being required
  to keep a parser up to date. (XML is designed to help change that, but
  more of that later.)

  As has been observed by many, however, there is a large class of 
  informative documents that are not at all well served by the 
  window-dressing that web publishing software puts on them. There 
  is a large number of document description languages that have been
  developed to address this. This problem is a major one for content
  developers who have little interest in the media employed to 
  make their information accessible and who prefer to use simplified
  logical descriptions of their content.

  The reason that none of these have become at all popular is because
  they are all very complicated to use. SGML (the parent of HTML) is an
  example, and XML (the sucessor to HTML) is expected to take a long time
  to become noticeably used, let alone dominant. Therefore ornate 
  HTML has continued to be the Internet publishing choice even for 
  documents whose structure can be very simply expressed and whose main
  purpose is to impart densely-packed information. Yet these are the very
  documents that vision-impaired people could have access to most simply! 

  There is therefore a potential triple benefit to developing a robust
  and simple scheme for universal presentation of simply-structured
  information:

	* to general developers of dense and/or plain content, from
          product manuals to white papers, reference documentation,
          academic-style publications and the equivalents through to plain
          books. This also covers those who have a simple message to
          impart (such as a terse description of a project) and
          no interest in fancy presentation

        * to sight-impaired web users, since their web interface is able
          to take a simple logical description and present it in whatever
          way the individual has chosen

        * to sight-impaired web content developers, for whom anything
          other than logical document description is largely irrelevant.

  (Note that such a scheme does not preclude the use of multimedia or 
  other technologies at the final stage of generating the information.
  It merely encourages a single source for production of all documentation 
  formats. Much more could be said on this topic.)

Simple Document Format
----------------------

  The Simple Document Format (SDF) has been developed within a largish
  company over the last five years or so to solve their internal
  documentation problems. The resulting specification and software
  has been made freely available, and as is usual in such cases a
  worldwide community of users has sprung up around the Internet. 
  SDF represents formally structured text in a bare
  minimum of description. It is designed to be handled by a single
  process that calls drivers depending on the output format desired.
  From SDF a very large number of outputs are available even currently,
  including at least ps, html, rtf, mif, pdf, sgml, txt. Styles for 
  structured documents from books to memos are available. Its programming
  capabilities allow construction of documents as complicated as the
  user desires. A very extensive and free suite of software exists
  to process SDF. There are many users of SDF around the world. The 
  home page is at http://www.mincom.com/mtr/sdf.

  One output that doesn't exist (except in a limited sense in plain text)
  is a "blind-friendly format". This is because the only context SDF can
  be used in at the moment is for static file generation runs. The sort of
  thing that would be done by this output format would be to offer a range
  of interactive indexing options (requiring much less text to be read to
  find information); present the exact logical structure of the document
  (eg tables of data would be clearly marked, whereas an HTML document
  frequently has many tables as part of its layout); allow extraction of
  certain parts of similarly-formated pages (eg just the abstracts) and
  so on.

  A classic application where this might make sense is the RLSB web
  site, where there are no hints on the screen-reader pages to
  optimise the browsing for a reader. In order to simplify the pages
  the logical structure has been largely lost.

Suggestion
----------

  It would not be at all difficult to recast SDF as a markup language
  designed for transport over HTTP. A single MIME type addition and
  development of client-side technology such as browser plug-ins would
  allow SDF to become instantly accessible to many millions of people.

  I would like to suggest that with quite a modest amount of funding 
  SDF could be prepared for submission as an Internet Engineering Task
  Force or World Wide Web Consortium standard. This would be no use
  without software developed to take advantage of the new standard,
  and so a set of reference tools would need to be developed and freely
  distributed to boost the application of SDF. It should be noted that
  in the Internet world it is the norm to proceed with a practical
  implementation of something and then discuss the initial results with
  and interested peer group: there would be no need to conduct some 
  kind of campaign to get support in order to do this.

  A very small number of blind-specific tags could be developed as part
  of the revised SDF. This would help in giving hints to screen readers.

  Specifically, the steps required would be:

    - refinement of existing and submission of a new SDF standard
    - development of client-side presentation software such as a 
      browser plug-in
    - integration of this into existing low-vision assistance 
      software
    - some refinements to the existing general parsing engine
    - development of showcase implementations. There are some high-profile
      companies who already use SDF who would quite possibly be very 
      pleased to participate. Furthermore blind societies around the
      world would surely want to help.

Desired Result
--------------

  The end result should be a new facility for web publishers which frees
  authors of much of the pain associated with maintaining bodies of 
  structured documents while giving the power back to end-users of
  defining the styles to be applied to the documents they are browsing.
  It just happens that this desirable facility is also perfect for 
  vision-impaired readers.

  This facility would by no means supplant existing and proposed 
  web standards, it would simply provide a tool for handling 
  documents which are not at all suited to current HTML technologies,
  and for which there is a clearly demonstrable need.

  If the resulting promotion of an SDF-like language and associated tools
  caused it to come into general use then all three potential
  beneficiaries would gain significantly.

  Of course the web is just a popular mechanism that sight-impaired need 
  to use. The document format would be useful in any general publication
  sense; so long as the software used had a good sense logical structure
  a File/Save As type filter should be quite easy to develop. Software
  companies often find it difficult to appear to be addressing the needs
  of the disabled and should the proposed language become a standard
  they may have a real PR-driven incentive to support it in their
  products.