Re: @longdesc implemented as ARIA construct

Laura,

Laura Carlson, Wed, 15 Sep 2010 15:44:22 -0500:
>> Some of the @longdesc supporters (Laura, I'm looking at you) seems
>> to see ARIA solutions as a hack.
  ... snip ...
> Longdesc is native. It solves a problem. It is simple. Why reinvent
> the wheel? Isn't that a design principle...anyway I digress...

I don't really accept that I "reinvent the wheel", I think I extend the 
wheel. At least, that is what I try. I have studied how role="img" 
works. And I suggest that role="img", which do no not need to be used 
on a <img> but which can be used on e.g. a <div>, should take into its 
model, the longdesc idea/feature. As ASCII fish example in HTML5 shows, 
such an role="img" element will not necessarily even contain a <img>, 
and thus do not necessary contain any element with the @longdesc 
attribute.
 
> If aria-describedby can fix the problem, why not give the fix to
> longdesc and fix it natively?

aria-describedby can't fix the problem. _That_ was my message. (Many 
have thought that it could work, that one could e.g point to an anchor 
link via aria-describedby, however, the only thing the user would get 
from that, woudld be the link text, without the link! ... )

> Most authors are not going to learn two languages to make content
> accessible. We have a hard enough time teaching them HTML. Pile on
> another layer and it will complicate things more.

Well, HTML5 *will* have ARIA, and many authors *will* learn ARIA, and 
role="img" is *modelled* after how <img> works. So why don't you 
support me in saying that role="img" should therefore incorporate a 
long description link feature? That too me seems very logical if you 
would do. ;-) 

(I feel that it is only this last mobth that I began to understand ARIA 
...)

> But with that said, if we could make a simple, native, solution that
> is better and *gracefully* move to it, that would be great.

I think that role="img" has the possibility of being graceful - 
provided that the ARIA spec editors don't just pretend that @longdesc 
doesn't exist. However, it remains to be seen if the ARIA spec writers 
pick up the idea.
-- 
leif halvard silli

Received on Wednesday, 15 September 2010 21:04:15 UTC