Re: FORMAL OBJECTION (was RE: Working Group Decision on ISSUE-204 aria-hidden)

On 8/13/2012 7:28 PM, John Foliot wrote:
> Sam Ruby wrote:
>> >
>> >    * Assertion of "harmful behaviors" is also not sufficiently
>> >supported
>> >      by evidence.  In particular, it makes a claim that has previously
>> >      been disputed, and without a single example of such markup.
> It is impossible to demonstrate harmful behaviors in a technology that has
> yet to surface - this is a ludicrous statement.
> Placing a <a href>Link</a> inside an @hidden container MUST still take tab
> focus for a non-sighted user and their Assistive Technology to interact with
> it, yet at the same time hides that focus from the sighted user: this is
> simple logic. Proving the harm of this is currently impossible due to lack
> of implementation anywhere: conversely however no one has been able to prove
> that this is NOT harmful.

There is a question as to whether or not existing browsers would or 
should be supported.
By changing the semantics here, in the ARIA binding, we lose some 
compatibility with HTML4 and existing HTML5 browsers and ATs.

As John has pointed out, there are no browsers in existence currently 
supporting focus on @hidden elements.
The only analog is the canvas tag, which has strong reasons for focus 
and has a spatial surface which authors are encouraged to bind to.

@hidden elements would have no spatial surface. This would mean that 
some portions of the page would only be accessible via the web rotor on 
the iphone.
It's unlikely that authors would be able to test this easily. With 
hidden lacking any spatial bindings, mainstream authors are less able to 
test for semantics.

The Canvas tag is already a departure, but in some sense, we expect that 
Canvas applications are supposed to be HTML5 only,
whereas other HTML5-centric applications have strong HTML4 fallback -- 
further, Canvas applications when written well can have full HTML4 fallback.
That's because we have strong feature testing with Canvas.
As of yet, I've not seen any proposals for authors to do feature testing 
on the new semantic.

The greatest burden is a departure from all existing browsers and there 
have been no accompanying specs as to how to feature test for support in 

I have not followed this thread in awhile, it seemed to me that there 
was greater consensus amongst members around this document:


Received on Tuesday, 14 August 2012 03:44:19 UTC