Re: feedback requested on WAI CG Consensus Resolutions on Text alternatives in HTML 5 document

On Aug 17, 2009, at 2:17 PM, Leif Halvard Silli wrote:

> Maciej Stachowiak On 09-08-17 11.05:
>
>> On Aug 17, 2009, at 2:00 AM, Benjamin Hawkes-Lewis wrote:
>>> On 17/08/2009 09:42, Maciej Stachowiak wrote:
>>>> Would it be appropriate, in light of this, to add a user agent
>>>> requirement that an img with empty alt should not be mapped to
>>>> accessibility APIs at all?
>>>
>>> Given the variety of authoring practice around "alt", I don't  
>>> think that would be safe.
>>>
>>> In situations like:
>>>
>>> <a href="#"><img src="delete.png" alt=""></a>
>>>
>>> It's useful to AT for the "img" to be exposed and to be able to  
>>> access "src" attributes for the purpose of providing a substitute  
>>> for proper alternative text.
>> If that's so, then wouldn't it be better for authors to use alt=""  
>> instead of role="presentation", so that AT can decide whether it  
>> needs to expose the image anyway? In particular, if your example  
>> was marked up like this:
>> <a href="#"><img src="delete.png" role="presentation"></a>
>> Then isn't it equally necessary and appropriate for AT to expose  
>> that image?
>> It seems like, based on this example, images should always be  
>> exposed to accessibility APIs, and AT should make the call on  
>> whether it needs to override the author-provided semantics.
>
>
> You say that we have agreement to add ARIA to HTML 5 ... And yet you  
> argue for keeping it out of HTML 5 as much as possible?

I don't see where I argued for keeping ARIA out of HTML5. All I'm  
saying is this: if authors might incorrectly mark an image as  
presentational (whether with alt="" or role="presentation"), then it's  
better to let AT make the choice of whether to really hide the image.  
If role="presentation" has a hard requirement to hide content before  
AT even gets to see it, then we shouldn't blanket encourage its use.  
Or, alternately, role="presentation" on an image should give AT the  
choice of whether to show the image anyway, as in cases where it is  
the sole content of a link. That is all. I don't think any of this  
argues against ARIA.

> Do you simply want to make role="presentation" permitted because you  
> want to be polite?
>
> It has been argued that it must not be permitted to use @role in a  
> way that conflicts with semantics of the elements in the host  
> language.
>
> Just now Anne suggested to add a <main> element based on two  
> premises: 1) "main" and "content" are much used as class names. 2)   
> "main" is a role in ARIA. So, by introducing <main> we possibly  
> build a bridge to authors and to ARIA. (Though sometimes it is  
> probably easier to use role="main".)
>
> Now, it could be argued - perhaps - that the ARIA construct <element  
> role="presentation">,  when it comes to images, has an equivalent in  
> HTML, namely <img alt="">.
>
> If that is how you view it, then would it not be better to  
> completely disallow role="presentation" on IMG?
>
> I would further expect that you would say that an <object> element   
> which has no fallback content, should be considered as having  
> role="presentation".

That's some detailed inferences of my views you made there. I haven't  
proposed disallowing role="presentation" on anything. I'm not sure how  
to respond, because I don't think I've said any of the above.

Regards,
Maciej

Received on Tuesday, 18 August 2009 00:12:48 UTC