W3C home > Mailing lists > Public > public-html-a11y@w3.org > April 2012

Re: 48-Hour Consensus Call: ARIA-DescribedAT & Longdesc

From: Charles McCathieNevile <chaals@opera.com>
Date: Tue, 03 Apr 2012 14:46:00 +0200
To: "Matthew Turvey" <mcturvey@gmail.com>
Cc: "HTML Accessibility Task Force" <public-html-a11y@w3.org>
Message-ID: <op.wb6weyhawxe0ny@widsith-3.local>
On Tue, 03 Apr 2012 14:21:53 +0200, Matthew Turvey <mcturvey@gmail.com>  
wrote:

> On 2 April 2012 17:13, Charles McCathieNevile <chaals@opera.com> wrote:
>
>> 1. aria-describedby *doesn't* provide a particularly usable way of  
>> getting to a description off the page, and more importantly
>
> In this context, putting a normal link on the image is the most usable
> way of providing a long description.

You appear to have simply ignored what I said about the cases where the  
link is an image, in which case you *can't* do that.

> How can aria-describedat be made as easy and intuitive to use as a
> normal link?

It can't. So what? "A normal link" and "a link whose purpose is to provide  
more information about a specific image" are not the same. So the question  
is irrelevant. The relevant question is how you can make  
out-of-normal-flow description easy to use. Longdesc, summary of tables,  
summary/details, alternative video tracks, etc. The problem is not  
trivial, but nor is it insoluble.

The answer will be primarily in development of user agent interfaces.  
Having worked on this for longdesc, I can confidently state that it takes  
much less time to iterate a number of solutions and progressively improve  
them than it does to have this argument. And that when I have deliberately  
not put a lot of time into this argument.

Dumbing everything down to a simple link is not the right answer for every  
question involving related information.

> The advantage of using aria-describedby as an enhancement *where
> required* is it still "works" in UAs that don't support ARIA, and for
> users that don't know to use ARIA, and when authors use ARIA
> incorrectly. As I understand it, aria-describedby was specifically
> designed to point to element(s) on the page to avoid repeating the
> same mistakes that made longdesc a disaster:
>
> http://lists.w3.org/Archives/Public/public-html/2009Sep/0596.html

As I recall it, the use cases were quite different (it was designed for  
web applications which were built using a pile of javascript and CSS that  
worked effectively to obfuscate any semantics or explanation which was in  
the content of what amounted to a single page). As I pointed out and you  
apparently ignored, the decisions made around aria-describedby meant that  
it solves certain problems, but does not solve certain prblems longdesc  
solves, which Geoff Freed explained.

>> 2. aria-describedby flattens everything to plain text, which is only
>> suitable for some of the use cases.
>
> See http://www.w3.org/WAI/PF/aria-practices/#Descriptions_external

Yes. That's one part of the first point above. aria-describedby is not a  
very suitable mechanism for solving some of the issues that longdesc (and  
the proposed aria-describedat) solves pretty well.

cheers

-- 
Charles 'chaals' McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg kan noen norsk
http://my.opera.com/chaals       Try Opera: http://www.opera.com
Received on Tuesday, 3 April 2012 12:46:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 04:42:57 GMT