Re: ARIA roles added to the a element should be conforming in HTML5.

Tab Atkins Jr. On 09-10-22 17.00:

> On Thu, Oct 22, 2009 at 9:49 AM, Leif Halvard Silli:

>> To make <h1 role="button"> non-conforming *is* to restrict how it can be
>> used and *is* to treat ARIA different from CSS.
> 
> Indeed, I think it should be treated differently.  Making it
> non-conforming is one step in that.
> 
>> The only likeness between CSS and ARIA that I suggested, is that ARIA should
>> over-rule the default semantics, the same way that CSS should overrule
>> default styling.
> 
> I still don't see why this should be allowed.  Just use the closest
> proper element.


We agree that it should be disallowed. But we perhaps disagree 
about whether it should work.

 
> Philip` pointed me to some of his raw data, at
> http://philip.html5.org/data/role-attributes-raw.txt, which is very
> helpful.  This is every usage of @role across 425k pages.  Scanning
> the list, you find only a handful of uses which are ARIA-related, and
> none of them are gross mispurposings of default element semantics.


Well, HTML doesn't allow ARIA yet, so that's fine, isn't it? ;-) 
It would also be interesting to check whether people actually use 
CSS in order to misuse <h1> ...

I don't know if a handful of pages can be used to formed an 
opinion. But if we had info about a handful of UAs that support 
ARIA, about whether they always let ARIA over-rule the default 
semantics, then that would definitely be helpful.

I think simple rules are better. Why don't you agree?
-- 
leif halvard silli

Received on Thursday, 22 October 2009 15:18:58 UTC