W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > April to June 2002

Re: FW: JSTOR and accessibility

From: Charles McCathieNevile <charles@w3.org>
Date: Tue, 28 May 2002 18:52:46 -0400 (EDT)
To: "Audrey J. Gorman" <agorman@megsinet.net>
cc: W3c-Wai-Ig <w3c-wai-ig@w3.org>
Message-ID: <Pine.LNX.4.30.0205281842420.9504-100000@tux.w3.org>
Presumably OCR works on a TIFF document to an accuracy of up to 99.95% - this
is the result that JSTOR claim they are getting by doing it. And it is indeed
complex to get better results.

The real questions are two - should they be commended for their approach to
providing a more accessible option, and could they easily do even better?

In both cases I think the answer is yes.

Providing a ramp to cover access to two out of three steps is enough for a
wheelchair to be lifted by a couple of large friends over the final one, so
it does increase acces, in the sense that some poeple who were completely
excluded are no longer completely excluded. The analogy is that not everyone
has an OCR system for recognising what the text is in the first place. But
helping the people who do happen to have that is better than ignoring

I also think they could clearly do better at more or less no cost. If they
can only get 99.95% accuracy using OCR systems it is a fair bet that their
users will only get that, or less in using OCR to read TIFF files. So they
are not really protecting the reputation of the publisher by not showing
typos, they are making the user go through the same hoops theydo, with the
likelihood that the user will end up with the same typos (and consequently
opinion of the publisher's editorial system) as if they made the text
available. And that is assuming the user has an OCR system as good as JSTOR.
Most users are not in a position to rival a commercial setup for software
quality. SO it would appear that maing the text available for download (with
the images retained as images) would be extremely low cost, and if it were
made clear how this text is generated and that it may contain transcription
errors the user can at least appreciate what has been made available to them.

My guess is that users willl accept the possiblity of an error, and may even
be prepared to propose editorial improvements that are simple to make but
cannot be systematically checked for by JSTOR.

SO I applaud the small step of makling tiff formats availale, but note that
it is a pretty small step, and I would be more inclicned to sing the praises
of an organisation that made their text available even if that required a
disclaimer pointing out the average error rate.

just my 2 cents


On Tue, 28 May 2002, Audrey J. Gorman wrote:

  This is from AXS-LIB, a libraries and accessibility discussion list. Two
  questions for the WAI-IG:
  1. How does OCR work on a TIFF document?
  2. Is Ron's assessment of the complexity of the process accurate?

  I think that accessibility fixes should be straightforward so that the
  "playing field" is at least a little more "level."  Answers? Comments? I'd
  like to help my colleagues in libraries get the right message to the folks
  at JSTOR. We're working on taking the accessibility message to all vendors
  who sell to libraries.

  Audrey J. Gorman
  Access for All
  Naperville, IL, USA

  Mobile: 630-661-9062
  Office: 630-717-7336
  www.accessall.net (under construction - temporary site)
  "The power of the Web is in its universality.
  Access by everyone regardless of disability is an essential aspect"
  Tim Berners-Lee, W3C Director and inventor of the WWW

  -----Original Message-----
  From:	Ron Stewart [SMTP:Ron.Stewart@ORST.EDU]
  Sent:	Tuesday, May 28, 2002 3:54 PM
  Subject:	Re: JSTOR and accessibility


  I guess my question is why bother. I appreciate the effort, but feel that
  is misdirected. While it does provide a better situation for the program
  doing a the document conversion it does nothing to improve access for the
  end user.

  Working with a TIFF file conversion can be more difficult that a straight
  OCR from print for the typical user, or using a reading machine which is
  our novice uses typically interact with non-accessible documents. This
  effort still will require that an intuitional process be established to
  convert the docs and does not provide the user with direct access to the
  content and as such does not meet the criteria of the law.

  Ron Stewart

  -----Original Message-----
  From: Schmetzke, Axel [mailto:aschmetz@UWSP.EDU]
  Sent: Saturday, May 25, 2002 9:02 AM
  Subject: Re: JSTOR and accessibility

  Ron, I would not say that the situation would not be better with at
  least TIFF files being made available. If it is the case that these can
  be OCRed (without having to go through the process of printing them
  first), then they can be converted into text files (which are, of
  course, of somewhat lesser quality because of the errors that occur in
  connection with the OCRing). It thus seems to me that having access to
  TIFF files is better than mere access to absolutely inaccessible GIF
  For me, the question is the following: How shall we respond to JSTOR's
  efforts? Shall we take the attitude that some access is better than no
  access and praise JSTOR for their attempts to make available a more
  accessible product--adding that they should strive to go still further
  by findings ways to make available top-quality text files? Or shall we
  simply snuff at their current efforts and insist that anything but
  text-based files is unacceptable?


  -----Original Message-----
  From: Ron Stewart [mailto:Ron.Stewart@ORST.EDU]
  Sent: Friday, May 24, 2002 3:02 PM
  Subject: Re: JSTOR and accessibility

  This is not any better than what we are dealing with now. TIFF files are
  graphics, not text, when JStor is willing to produce true text documents
  then they will not have to worry about compliance.

  Ron Stewart

  -----Original Message-----
  From: Schmetzke, Axel [mailto:aschmetz@UWSP.EDU]
  Sent: Friday, May 24, 2002 12:57 PM
  Subject: Re: JSTOR and accessibility

  I'm very hesitant to consider JSTOR's approach towards providing
  accessibility an acceptable solution, but the very fact that a
  database/e-journal provider publicly addresses accessibility issues
  pertaining to its graphics-based product is promising.

  I'm curious: How do you folks feel about JSTOR's approach towards
  providing some measure of accessibility? Should we, as librarians,
  consider graphics TIFF files, which can be OCRed and can thus be
  converted into a screen-readable text-file, to be sufficiently
  accessible? Or are we bothered by the fact that it takes an additional
  piece of technology (OCR software), and thus an additional step, to get
  to text-based information, and that the converted text is substandard
  because of the errors produced by current OCR technology?


  Axel Schmetzke
  University of Wisconsin-Stevens Point

  -----Original Message-----
  From: Coonin, Bryna R [mailto:COONINB@MAIL.ECU.EDU]
  Sent: Friday, May 24, 2002 1:23 PM
  Subject: JSTOR and accessibility

  Friends --
  Full-text e-journal provider, JSTOR, has a team actively working on some
  the accessibility issues in JSTOR that have concerned many of us over
  To keep users informed about developments in this area they have now
  included updates on this effort on the JSTOR web page at:


  Bryna Coonin
  Joyner Library
  East Carolina University
  Greenville, NC 27858
  E-mail: cooninb@mail.ecu.edu

Charles McCathieNevile    http://www.w3.org/People/Charles  phone: +61 409 134 136
W3C Web Accessibility Initiative     http://www.w3.org/WAI  fax: +33 4 92 38 78 22
Location: 21 Mitchell street FOOTSCRAY Vic 3011, Australia
(or W3C INRIA, Route des Lucioles, BP 93, 06902 Sophia Antipolis Cedex, France)
Received on Tuesday, 28 May 2002 18:52:47 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 13 October 2015 16:21:19 UTC