Re: <canvas> and the 2D context API (was RE: Begin discussions for pushing Last Call into 2010)

On Aug 12, 2009, at 5:09 PM, Adrian Bateman wrote:

> On Wednesday, August 12, 2009 4:37 PM, Shelley Powers wrote:
>> Sam Ruby wrote:
>>> I will also say that while I truly did not have an idea how the
>>> current poll would end up (and even though there does appear to be a
>>> trend at the moment, I'm still not certain), I do have a strong
>>> intuition on what the consensus would be on canvas at the moment.
>
>> The current working group may consider that the current charter can  
>> be
>> stretched to include Canvas, but I'm not sure others outside of the
>> group would agree. (Though I'm not sure that the working group of  
>> today
>> would necessarily vote the same -- the makeup of the group is  
>> different.
>> Views about the Canvas element are also different. )
>
> My understanding is that the WG decision was that canvas was in  
> scope for
> the working group. The mail documenting the decision [1] indicates  
> that
> there was some support for the 2D context API being documented  
> separately:
>
>  We also note support for splitting the immediate mode graphics API
>  out of the HTML 5 spec and inten to pursue that option by
>  recruiting writing resources.
>  http://www.w3.org/2002/09/wbs/40318/tactics-gapi-canvas/results#xq3

I don't think we absolutely ruled out putting the API itself (but not  
the <canvas> element) in a separate document. But I think it's  
potentially misleading to characterize a 30-5 vote the other way as  
"support". In any case, there was a clear desire to make the work an  
HTML WG deliverable working via public-html: <http://www.w3.org/2002/09/wbs/40318/tactics-gapi-canvas/results#xwhowhere 
 >.

> I don't know whether having the <canvas> element defined in the HTML  
> 5 spec but the graphics API in a separate document solves the  
> accessibility issue at the root of this thread.

I don't believe it would address accessibility issues. Refactoring the  
spec would not make the element any ore accessible. It may even make  
it harder to do solve the accessibility issues, since some of the  
approaches being considered rely on existing HTML facilities.

>  I suspect there is some relationship since a different API would  
> likely require different accessibility support.

That seems fairly unlikely to me. Many native operating systems have  
the same accessibility API for a custom drawing area regardless of the  
graphics API used. For example, on Mac OS X, whether you use  
QuickDraw, CoreGraphics, CoreAnimation or OpenGL, you would use the  
same accessibility APIs to reflect the content to assistive  
technologies.

> Nonetheless, I think Microsoft would support that separation since,  
> amongst other things, it would allow the graphics API to proceed in  
> the longer term on a different rhythm to the HTML 5 spec itself.

If anyone makes a high-quality split out spec for just the  
GraphicsContext2D part of Canvas, then we could probably find  
consensus on normatively referencing such a document and publishing it  
separately. So far, no one has stepped up.

Regards,
Maciej

Received on Thursday, 13 August 2009 00:32:11 UTC