Re: HTML-A11Y Task Force Consensus on Issue-204 (Updated)

On Thu, May 24, 2012 at 8:54 AM, Jonas Sicking <jonas@sicking.cc> wrote:
> On Wednesday, May 23, 2012, Edward O'Connor <eoconnor@apple.com> wrote:
>>> While allowing UAs to expose the accessibility tree of @hidden content
>>> is great, I think we should make it a MUST level requirement.
>>
>> My understanding is that some AT implementors believe it would be
>> difficult to comply with such a requirement. That said, I agree that we
>> should encourage UAs to expose the tree to AT if they are able to. Would
>> a SHOULD work for you?
>
> So far I haven't seen such feedback from any implementor. I've only seen me
> and Maciej speak up on this and we've both said that implementation-wise
> this is similar to exposing an accessibility tree for the contents of
> <canvas> elements. Certainly not trivial, but doable and something that
> needs to be solved in order to make canvas accessible (which I hope we agree
> should be a MUST level requirement).

The accessibility approach for <canvas> involves creating bindings
between the accessibility tree and regions of the canvas, allowing not
only totally blind users but also users with motor disabilities or low
vision _who need to see something on screen_ to navigate and
understand the canvas. An accessibility approach for subtrees in
@hidden would need to allow for the creation of a *visual* user
interface with which users with motor disabilities or low vision could
interact.

So it's simply not true that if we can solve <canvas> then we can
solve @hidden. In the absence of use cases for hidden subtrees, it's
hard to say what solving it would mean anyway.

I agree with Ted that if we can _solve_ @hidden we should add spec
text elaborating our solution, rather than just assuming it will be
solved … somehow.

--
Benjamin Hawkes-Lewis

Received on Thursday, 24 May 2012 09:44:24 UTC