RE: Example canvas element use - accessibility concerns

Rob Sayre wrote:
> 
> 
> As I have explained once already, the terms we're using have
> definitions
> (RFC 2119).
> 
> 3. SHOULD   This word, or the adjective "RECOMMENDED", mean that there
>     may exist valid reasons in particular circumstances to ignore a
>     particular item, but the full implications must be understood and
>     carefully weighed before choosing a different course.

Right, and I'm glad to see that the BeSpin developers were fully informed
and aware, and "carefully weighed" the implications of the inaccessible
application they were releasing on the world.  This is not a criticism of
them per-se, but surely proof that the current should/recommendation is
clearly not enough.  And I believe *I* was the first in this thread to
specifically reference RFC 2119, and it's definition of "Must", so I
understand full well what I am suggesting.

> 
> This is not different than Ian's approach. In his draft, @alt text can
> be omitted if you have a valid reason.

Rob, you know full well that this remains a contentious issue, and is in no
way resolved. There has yet to be a truly valid reason why not, simply some
conditions where it becomes increasingly difficult to achieve the desired
outcomes - not impossible, just increasingly difficult.  Sam Ruby last week
suggested that keeping @alt mandatory was likely the best baby-step forward,
no?  Finally, according to the current finalized specification
(HTML4/XHTML1), @alt is mandatory, and the jury remains out on the final
place for @alt in the draft spec.

> >    For every image that has alt="picture"
> > there is also one that has alt="useful alt text" - all simply because
> we
> > (currently) *insist* that<img>  contain @alt.
> >
> 
> That data I've seen doesn't support that assertion. Here's one example:
> 
> <img>  elements with no alt attribute: 1104466 (47%)
> <img>  elements with zero-length alt: 530687 (23%)
> <img>  elements with non-empty whitespace-only alt: 11943 (1%)
> <img>  elements with non-empty non-whitespace alt: 702702 (30%)
> 
> <http://lists.w3.org/Archives/Public/public-html/2008Apr/0536.html>

I would hardly call that a good exemplar, as it is but "130,000 pages from
the list on dmoz.org" (out of how many millions on the web today?)
Disappointing to be sure, but I think a survey of web pages less than 5
years old, from multiple domains, would likely be a better representation.

However, I tire of this squabble - if you really believe that leaving things
up to the good graces of "suggestion" will make a more accessible web, then
you are entitled to your opinion.  Myself, I point to all examples of
<canvas> I've seen in the wild, and not one of them is currently accessible
to Adaptive Technology (and specifically screen readers), so I know for sure
that currently your method does not seem to be working - "carefully weighed"
considerations notwithstanding. 

JF

Received on Tuesday, 24 February 2009 05:18:09 UTC