Re: Path to Last Call (was closing various issues)

On Aug 23, 2009, at 5:07 AM, Steven Faulkner wrote:

> Hi maciej,
>
> >I not heard any claims the ARIA integration work done so far is  
> insufficient for Last Call.
>
> as far as i can tell the ARIA spec text was only added in the last  
> few days, and its presence was not signalled on the html wg or wai- 
> xtech mailing lists (unless i missed it)  so it is not surpising  
> little feedback has been forthcoming, hopefully the spec text and  
> Ians questions will be discussed in this weeks PF meetings.

It's definitely too soon to declare that we have consensus - I think I  
said that in the rest of my message. But it also seems too soon to  
declare we don't. I would say the point we're at is that it's now time  
to review the draft, identify issues, and determine if we have  
consensus. If I understand you correctly, I think we're in agreement  
on this.

> In regards to the canvas issues, while i generally agree that good  
> progress is being made,
> I may be misunderstanding the discussions so far, but I do not get a  
> sense yet that a solution will be forthcoming without some additions  
> to the canvas API involving the creation of shadow W3C dom objects  
> and focus management.

I believe nearly everything that's needed can be added without API  
changes, and if API changes are needed, they will be fairly minor. On  
the specific points you mention: What we loosely referred to as a  
"shadow DOM" is in fact just the DOM children of the <canvas>, and can  
be added with the ordinary DOM APIs. That part doesn't need new API at  
all. Maybe a demo would show this best (working on it.) For focus  
management, one proposal on the table is to simply let descendants of  
the <canvas> with explicit tabindex (including tabindex="0") still  
receive focus and participate in the tab cycle, even though they are  
not rendered. I'd be glad to explain further offline.

It may be that we'll find we need API additions - the ones I can  
imagine include a way to draw a standard focus indicator, and a way to  
expose a caret or selection range in custom-drawn text. I expect we'll  
learn more about what we need from the prototyping/mockup work. My  
estimation is that the additional APIs needed will be modest.

Regards,
Maciej