W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 2001

Re: More on 3.4

From: Charles McCathieNevile <charles@w3.org>
Date: Wed, 1 Aug 2001 19:54:30 -0400 (EDT)
To: Matt May <mcmay@bestkungfu.com>
cc: <w3c-wai-gl@w3.org>
Message-ID: <Pine.LNX.4.30.0108011926030.3826-100000@tux.w3.org>
On Wed, 1 Aug 2001, Matt May wrote:
  ..metadata is not
  commonly stored in images, and the tools to create, manipulate or view the
  images don't provide a standard mechanism for doing so, which limits the
  ability for images to self-identify in metadata, thus requiring the alt
  attribute.

CMN
Right. This is starting to change though, and we should be encouraging that
among other things. It has been in the non-normative Techniques document for
ATAG for a couple of years, and I find it hard to see how a graphic editing
tool could conform to ATAG without making this happen. As people satrt to
take accessibility seriously in large organisations, and apply it to large
collections of information, I think we will see acceleration  of these
trends.
[snip]
  MM
  I don't disagree with anything you're saying here. I may even agree with you
  on a great number of the "some cases" in which supplemental content is
  beneficial. I just don't see where success criteria are discoverable, or
  techniques can be broad enough to be effective.

Well, I don't see a lot of the success criteria yet either. I can see that
they are discoverable in discussion and experimentation, because some of them
turn out to be really obvious (I am about to propose one that hasn't been
mentioned as such yet, although it has come up in discussion a few times. I
even suggested it as easy to do without thinking further). In many cases the
first proposal won't work, but with refinement and thought it can work.
MM
  I've said before that there needs to be a non-normative document or
  documents for accessible web design, and even that it should be required
  reading. It's stuff like 3.3 and 3.4 that's continuing to lead me to believe
  that the idea still has merit.
CMN
Agreed absolutely. The document is likely to be called "techniques for WCAG
2"  and be some hypertext collection, and I think it is of crucial importance
even in the development process for WCAG 2. (Fortunately there are some very
smart people working on how to make such a document, like some Matt May guy
<grin/>).

Cheers

Charles

  MM
  Where I live, HTML-only authors are considered semi-skilled labor. Anyway,
  many new repair tools can be used by those who don't know HTML (for which we
  have ATAG to thank). But those HTML workers and people interacting with
  accessibility tools won't be able to do anything for 3.4, that I can detect.
  It's out of scope for either group.

Well, I think it is in scope, but we haven't gone a long way down the
development path. Here's an idea:

identify keywords (plenty of stuff chases this - it is common in authoring
systems designed for managing big sites).

Look them up in a dictionary (for example wordnet).

If there is an annotation pointing to some graphic representations, present
them as options (Amaya does most of this, and the Javascript work on Annotea
that makes it useable in most browsers to a lesser extent can also do it. An
authoring or repair tool could be extended without major technical
impediments).
 Otherwise ask the author for at least one, and make the annotation (several
systems available for that) in the database.

 Either way, annotate the page with the image/sound/etc.

Have a look through ATAG guideline 3 [1] imagining that you are trying to
make it about confomring to some falvour of 3.4 checkpoint. I just did it as
a thought experiment, and when the current "wombat" (our new ATAG version in
development) draft comes back for discussion on the working group agenda I
have some thoughts...)

[1] http://www.w3.org/TR/2000/REC-ATAG10-20000203/#gl-prewritten-descs
Also the newest thinking from the wombat draft:
http://www.w3.org/WAI/AU/wombat/#gl-prewritten-descs
Received on Wednesday, 1 August 2001 19:54:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:11 GMT