W3C home > Mailing lists > Public > public-html@w3.org > October 2009

Re: ISSUE-30 (Longdesc) Change Proposal

From: Charles McCathieNevile <chaals@opera.com>
Date: Thu, 29 Oct 2009 10:29:53 +0100
To: "Jonas Sicking" <jonas@sicking.cc>, "Leif Halvard Silli" <xn--mlform-iua@xn--mlform-iua.no>
Cc: "public-html@w3.org" <public-html@w3.org>
Message-ID: <op.u2j1z3olwxe0ny@widsith.eng.oslo.osa>
On Thu, 29 Oct 2009 07:29:52 +0100, Jonas Sicking <jonas@sicking.cc> wrote:

> On Wed, Oct 28, 2009 at 6:19 PM, Leif Halvard Silli
> <xn--mlform-iua@xn--mlform-iua.no> wrote:
>> Jonas Sicking On 09-10-29 00.57:
>>
>>> On Wed, Oct 28, 2009 at 7:59 AM, Leif Halvard Silli:
>>>>
>>>> Jonas Sicking On 09-10-27 20.15:
>>>>>
>>>>> On Tue, Oct 27, 2009 at 7:09 AM, Leif Halvard Silli:
>>>>
>>>>>>> I agree that @longdesc and @aria-describedby aren't exactly the  
>>>>>>> same. However they are very similar.

Yes, they are.

>>>>> If two features are designed to solve the same problem, then I think
>>>>> they are similar enough that having both is a loss for all involved
>>>>> parties.

In theory, this is true. In practice, it sometimes is and sometimes isn't.

>>>> Fact is: @alt and @aria-labelledby are somewhat related [*].
[...]

>> It would not break anything: "removing" as per HTML 5 means that authors
>> should not use the feature anymore, but that user agents continue to
>> implement it.
>
> True, depends on how we "remove" it, and what you mean by "break". See
> more below.
>
>>> That's the only reason why I think we should not let
>>> @aria-describedby replace @alt.
>>
>> Ahem ... Either you did not notice that I spoke about aria-LABELLEDBY.  
>>  Or,
>> you are saying that you would like aria-DESCRIBEDBY to replace
>> aria-LABELLEDBY ... Aria-describedby is not a solution to everything ...
>>
>> For the record, "@aria-label" (not @aria-labelledby) is what is equal to
>> @alt. But since HTML has @alt, we don't use aria-label (on IMG at  
>> least).
>
> Ok, point taken. I don't actually know enough about accessibility to
> say if aria-label and/or aria-describedby could replace alt. Reading
> the spec it actually doesn't seem like either does....

Which points to something the specs do badly, since as Leif says they  
really are doing equivalent jobs (in slightly different ways, with some  
overlaps and some uniqueness)

What should happen is that the specs for ARIA should clarify where  
aria-labelledby is like alt - in part simply because people have an idea  
of how to use alt.

One reason not to do so is that many (perhaps most - I haven't done the  
detailed analysis that getting accurate numbers would require, but clearly  
it is a very significant proportion) people's idea is actually more or  
less wrong.

But then, it's also like label (for forms), so we don't have to be  
piggybacking entirely on bad legacy.

> Ultimately I still think that removing @alt at this point is bad,
> irrelevant of if there is a collision with aria features or not. There
> simply is too much content out there that uses it. Asking all the
> authors that filled out an alt attribute would not make us any
> friends.
>
> There's another good reason too which I forgot to mention in my
> previous email, but which I've mentioned on this list before. @alt is
> one of very few success stories for accessibility-specific features.
> There's more correctly authored @alt attributes out there than
> @longdesc, @summary, @headers and @aria-* put together. Messing with
> @alt would be trying to fix something that isn't broken.

Guessing, as I already mentioned, but with some clear basis, I strongly  
suspect there is way more broken alt content out there than there are  
longdesc, summary, headrs and aria-* put together. Discounting completely  
absent alt, which I think is still the overall majority case.

And yet, alt is indeed a success story. It genuinely helps people in many  
many cases, and many many authors genuinely understand how.

> The same can not be said for @longdesc ... Yes, there is
> more than zero uptake, but I don't think there is enough to warrant
> having duplicate (or near-duplicate) features.

There is significant uptake in the ancillary materials - the books,  
courses, how-to pages and so on, that already use longdesc. Stripping it  
out doesn't always make us friends either.

Since it isn't that complicated (Opera recently implemented it, in trivial  
time, iCab has had it for ages as have screen readers that look for it on  
their own, there is a nice plugin for Mozilla) and since it has a lot of  
supporting material already, we can take advantage of that fact to explain  
more of aria to authors. It is possible that we should deprecate longdesc  
- point out that it is obsolete, *if* aria-describedby can do the same job  
in the real world. But I don't think that day has yet arrived, and until  
it does, leaving an existing and well-described HTML feature that is  
useful (admittedly not to the whole universe, but to some people in some  
cases) in the spec seems a more rational choice.

cheers

Chaals

-- 
Charles McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg lærer norsk
http://my.opera.com/chaals       Try Opera: http://www.opera.com
Received on Thursday, 29 October 2009 09:30:30 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:09 UTC