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

Re: Adaptive Image Element Proposal

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Thu, 30 Aug 2012 18:46:03 +0200
To: Peter Winnberg <peter.winnberg@gmail.com>
Cc: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>, Mathew Marquis <mat@matmarquis.com>, HTML WG <public-html@w3.org>, public-respimg@w3.org
Message-ID: <20120830184603992109.cc75c474@xn--mlform-iua.no>
Peter Winnberg, Thu, 30 Aug 2012 18:33:09 +0200:
> 2012/8/30 Leif Halvard Silli:

Regarding permission to use img@aria-labelledby="picture" instead of 
img@alt:

>> <picture alt="Alernative text" id="pict" >
>> <img src=file aria-labelledby="pict" />
>> </picture>
> 
> What a browser / AT that lacks support for ARIA and the picture
> element would "see" (i.e., what parts of this markup is supported) is
> most likely this:
> 
> <img src=file />
> 
> Which means that it has an image with no alt text.

Except that if it supports ARIA, the picture support is not required: 
Agreed. 

And another, minor, issue that came to mind is that the presence of 
aria-labelledby on the img element currently never makes it valid to 
omit the alt attribute.

The question is how important these drawbacks are. Keep in mind that 
the proposed draft does not make the img element an requirement and 
neither does it operate with MUST rules for the alt attribute of the 
element.

The simplest rules for the img element as child of picture, would be to 
say that it should be validated/authored as if the parent picture 
element did not exist - but for the, of course, very important detail 
that it should contain the same image and the same alternative text.

> I have previously expressed concern that people would abuse this
> element to show images that don’t share the same content [1]. Having
> an alt attribute on the picture element makes it fairly clear that
> they should share the same alternate text and therefor also content.
> On the other hand, repeating the alt text for both the img and the
> picture elements is problematic (for example, what if the alt text
> isn’t the same on both attributes?) just like you pointed out.

I have not thought much on the same content problem.

>>  <picture aria-labelledby="pict" >
>>   <img src=file id="pict" alt="Alernative text" />
>>  </picture>
> 
> Placing aria-labelledby on the picture element instead of the img
> element seems like a possible solution.

It is of course the most back-compatible solution. And perhaps also the 
method that is most in line with the 'augmentative' authoring praxis?

> Just not made up my mind if
> this still sends that "signal" that all images inside the picture
> element should share the same content or not.

Question regarding letting picture@aria-labelledby=id point to img#id:

* Might it create the impression that the img element is part of the
  picture element construct?
* Might it complicate the rules for when one must include the 
  alt attribute on the picture element?
* Might it make it impossible to have exactly same alt attribute
  rules for the picture element as we already have for the img
  element? (Though this might not be possible anyway?)

If we say that it is the picture that should point to the img, then I 
feel it tempting to say that the img element should be a required child 
element ... 

> [1] https://www.w3.org/Bugs/Public/show_bug.cgi?id=18384#c5

-- 
leif halvard silli
Received on Thursday, 30 August 2012 16:46:36 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:33 UTC