RE: Canvas 2D API specification update - defining the element or not


I actually think that the relative lack of maturity of the accessibility pieces of Canvas is one of the reasons it should be broken out of the HTML 5 spec.  That would allow it to proceed on its own timeline, with enough time to design the required accessibility features.


-----Original Message-----
From: [] On Behalf Of Adrian Bateman
Sent: Thursday, October 22, 2009 12:59 PM
To: Maciej Stachowiak; Richard Schwerdtfeger
Cc: Eliot Graff; Frank Olivier;;; Doug Schepers
Subject: RE: Canvas 2D API specification update - defining the element or not

On Wednesday, October 21, 2009 3:25 PM, Maciej Stachowiak wrote:
> On Oct 21, 2009, at 2:57 PM, Richard Schwerdtfeger wrote:
> > Adrian,
> >
> > If we are going to place this in a separate specification then we will
> > need to include support for accessibility. How well we address that
> > depends on runway. For example, I believe we need a combination
> > of:
> >
> > - DOM/ARIA support
> > - equivalent alternative content
> > - Accessibility API support.
> >
> > What is the timeline being projected for a separate canvas spec.,
> > relative to HTML 5, if we were to separate it out?

Rich, I absolutely agree with the need for improved accessibility support and your
summary makes sense to me. As you know Frank Olivier has been participating in
some of the discussions you've already had on this. The working group will decide
the timeline. We will continue to contribute to the discussion and together we can
make these improvements.

> The broken out spec does not include the parsing rules, the content
> model elements, and the definition of allowed markup attributes. So it
> can't define those things, they would have to be defined in HTML5
> itself. And indeed they are. This is one reason I suggested that the
> breakout spec should be just CanvasRenderingContext2D (and related
> interfaces), while the full element definition should remain in HTML5.
> As it is, the split out spec partially defines the element, leaving
> gaps. I don't think breaking out part of the element definition makes
> sense.

Maciej, as I mentioned in my follow-up to Eliot's mail, I think the way that the
spec is broken out needs attention. I actually agree that just specifying the
CanvasRenderingContext2D interface as a separate document is best, provided that
it is possible to adequately address the accessibility concerns now and for future
versions with that division. My view remains the same as when I mentioned this
back in August [1].

I think the best way to break out the canvas API makes an excellent discussion
topic for the group. I propose that we take this to the canvas API mailing list
and try to find a consensus about the best way to proceed. Once the group reaches
a decision Eliot can then update the draft to that effect. There are a number of
issues in the "to do" list at the end of the document. Finding the best way to
reconcile the relationship to HTML5 is one of them.




Received on Friday, 23 October 2009 00:04:51 UTC