Re: ISSUE-30 longdesc - Chairs Solicit Alternate Proposals or Counter-Proposals

The web is polluted with bad code period. I led a large middleware
transcoding project for seniors back in 2000 and you can't imagine all the
hacks browser manufacturers do to repair really bad code. The fixes
propagate more bad code. One of the largest pieces of our project was
repairing and dealing with badly written and poorly formed web content.

What I am saying is bad coding practices are not limited to ARIA.

Rich Schwerdtfeger
CTO Accessibility Software Group



From:	Matthew Turvey <mcturvey@gmail.com>
To:	HTML WG <public-html@w3.org>
Date:	06/17/2011 06:33 AM
Subject:	Re: ISSUE-30 longdesc - Chairs Solicit Alternate Proposals or
            Counter-Proposals
Sent by:	public-html-request@w3.org



On 16 June 2011 18:21, Laura Carlson <laura.lee.carlson@gmail.com> wrote:

> It gets back to what Cliff was talking about and what you said about
> people having to read another large spec besides HTML5. I too
> experience the same thing all too frequently here at my job. Content
> authors may know basic HTML and be willing to put in longdesc for
> complex images but they are not going to delve into other specs. ARIA
> is not an option for these authors.

Cliff also thought the longdesc attribute contained the image
description text not the link to the image description until quite
recently. This was despite longdesc being specified in the much
shorter HTML4 specification for 13 years.

(no disrespect to Cliff who is an experienced and respected
accessibility/usability expert; it's a common misunderstanding due to
the misleading name).

So I think the problem is more with AT-specific features generally,
rather than which particular technical document the feature is
specified in. I don't think content authors generally read technical
specs, they generally use authoring guides, books, references, blogs,
in-tool contextual help etc to find out what to use, or just copy and
paste, in my experience.

I think ARIA features will take off, but some may be misused to the
extent that they become effectively useless to users, and we may even
end up regarding some of them as a disaster that should be buried in
the back yard. For example, here's some recent tweets from screen
reader users about their experience with role application:

md_curran: ARIA folk: could someone please explain to me why a large
chunk of this page is within an ARIA application? [page is now fixed]
http://www.microsoft.com/enable/news/marcexpansion.aspx

jcsteh: Wow! That's one of the worst abuses of ARIA application I've
ever seen. Stupid.

DomasoFan: @jcsteh here at our banking site too. painful. i always
have to switch to jaws because nvda keeps activating it all the time.

I think our efforts would be better directed towards ensuring WAI-ARIA
markup does not become polluted with the kind of widespread misuse
we've seen with the longdesc and summary attributes, to the extent
that AT users ignore it or switch it off altogether. I'm not sure how
we can do that, but ensuring WAI-ARIA is as clearly specified and
unambiguous as possible to authors and implementers may be a good
place to start:

http://www.w3.org/2011/06/15-html-a11y-minutes.html

-Matt