Re: From the HTML-WG about aria-hidden

Jason White, Fri, 17 Feb 2012 15:10:17 +1100:
> Leif Halvard Silli wrote:

>> But now I understand that what you really mean is that it would be 
>> simpler to add longdesc=link directly on the <img>. And I won't deny 
>> that you are right.
>> However, authors might find it simpler to test how @hidden works 
>> compared with testing how @longdesc works. Because, as you know, 
>> authors often don't get those things right that they don't see any 
>> effect from. [And I realize that now I talk as if the @longdesc support 
>> won't improve.] 
> And that's where your case founders.
> If @longdesc is specified properly, then there will be a mechanisms that
> allows authors to see its content, so the above argument doesn't apply.

Yes, if 'specified properly' means 'specified in a way that vendors 
will implement'.

  ... snip ...

> I don't like @longdesc because I think it's just a hack to deal with 
> a bad design decision made prior to HTML 2.0, whereby IMG is not a 
> container element and therefore cannot have alternative content 
> supplied. Unfortunately, decades later, we have to live with the 
> legacy of this and other decisions. In that context, @longdesc solves 
> a problem, but could be better specified on the user-agent side. 
> There may be superior solutions, but the proposal discussed in this 
> thread isn't one of them in my opinion.

First, I will say that ARIA can also be described as something that we 
unfortunately must live with.

Second: W.r.t. alternative content as child element, then I have news 
for you, in that regard: Support for <object> is improving. And I have 
recently written a change proposal for ISSUE-158 which will change the 
content model of <object> to be more similar to what it was in HTML4. 
This would allow authors to wrap an <object> in the anchor element 
*and* place a link as child of <object>. This should be, quite frankly, 
the HTML2.0 solution. You can read more about my proposal at the end of 
this letter. I love to get feedback that could help me improve it. In 
particular no screen reader user has wetted what I say there yet. But - 
danger, danger: I *think* that there is a risk that my change proposal 
for OBJECT could impact negatively on @longdesc ... Not because I want 
to weaken @longesc [I would like it to be valid and also would like 
support to incrase], but because it does provide another option for 
doing what @longdesc is supposed to do.

Leif Halvard Silli

Received on Friday, 17 February 2012 04:43:40 UTC