RE: Agenda: August 27, 2015 WAI-PF ARIA Caucus (Correction I am still chairing this week)

More and more I'm in the camp of the removers .. BTW, what about colgroup?

- Stefan

-----Original Message-----
From: James Craig [mailto:jcraig@apple.com] 
Sent: Donnerstag, 10. September 2015 20:52
To: Joseph Scheuhammer
Cc: Richard Schwerdtfeger; PF; Alexander Surkov
Subject: Re: Agenda: August 27, 2015 WAI-PF ARIA Caucus (Correction I am still chairing this week)

One clarification: If an author explicitly add tabindex to an HTML rowgroup (thead/tfoot/tbody), it is exposed as a generic focusable (like a <div>) and therefore can invalidate the accessibility semantics of the table. The may require an authoring guide update, too.

James


> On Sep 10, 2015, at 11:43 AM, James Craig <jcraig@apple.com> wrote:
> 
> 
>> On Sep 10, 2015, at 6:24 AM, Joseph Scheuhammer <clown@alum.mit.edu> wrote:
>> 
>> There's another way to look at this that could get rid of the problem.
>> 
>> If there is no reason to ever expose an accessible for an element with role="rowgroup", then the semantics of rowgroup are non-existent, for all intents and purposes.
>> 
>> In that case, the rowgroup role can be deprecated and removed from the spec.
> 
> No. The only reason we added it (discussion logged at TPAC Sunnyvale in 2010?, 2011?) was so that presentation inheritance mapping to HTML elements could exist. 
> 
> Cite: http://www.w3.org/TR/wai-aria-1.1/#presentation
> """
> When an explicit or inherited role of presentation is applied to an element with the implicit semantic of a WAI-ARIA role that has required owned elements, in addition to the element with the explicit role of presentation, the user agent must apply an inherited role of presentation to any owned elements that do not have an explicit role defined. 
> """
> 
> We did this so that rows and cells would inherit the "presentation" role whether or not the rowgroup was explicitly included.
> 
> <table role="none">
>  <tr>
>    <td>presentational, no table semantics</td>
>  </tr>
> </table>
> 
> <table role="none">
>  <tbody>
>    <tr>
>      <td>also presentational, no table semantics</td>
>    </tr>
>  </tbody>
> </table>
> 
> The rowgroup discrepancy between browsers started when Alex (who was not involved in the TPAC rowgroup discussion 2011? 2012?) mapped rowgroup explicitly in Firefox a few years ago. Prior to this, it had not been exposed on any platform.
> 
> Since the consensus seems to be that there is no need to have focusable rowgroups, I believe the appropriate resolution is to change the implementation spec to explicitly mark rowgroup as "not exposed" for all platforms, like it always had been, and revert the Firefox change. 
> 
> James
> 
> 
>> The advantages are:
>> - no author could ever make the error of adding @tabindex to a rowgroup element.
> 
> <tbody tabindex="0">???
> 
>> - no author could ever make the error of adding click, keypress or other events to the element.
>> - no webapp would be affected since there are no real world examples, or, if there are, then we would find out pretty quickly.
>> - user agent mappings would be interoperable since no user agent would expose the role.
>> 
>> Any objections to removing the rowgroup role?
>> 
>> -- 
>> ;;;;joseph.
>> 
>> 'Array(16).join("wat" - 1) + " Batman!"'
>>          - G. Bernhardt -
>> 
>> 
> 
>

Received on Friday, 11 September 2015 07:48:03 UTC