On Feb 4, 2010, at 8:40 AM, Richard Schwerdtfeger wrote:

Maciej,

This is in contradiction with your earlier statements.

My previous statements were that I agree an accessible DOM as immediate children of the <canvas> is practical to implement and a good solution. I have been consistent in opposing a special <accessible> element and having multiple accessible alternatives declared through markup.

Regards,
Maciej


Rich


Rich Schwerdtfeger
Distinguished Engineer, SWG Accessibility Architect/Strategist

<graycol.gif>Maciej Stachowiak ---02/04/2010 10:16:39 AM---On Feb 3, 2010, at 5:49 PM, Ian Hickson wrote:

<ecblank.gif>
To
<ecblank.gif>
Ian Hickson <ian@hixie.ch>
<ecblank.gif>
cc
<ecblank.gif>
Richard Schwerdtfeger/Austin/IBM@IBMUS, Jonas Sicking <jonas@sicking.cc>, "public-html@w3.org" <public-html@w3.org>
<ecblank.gif>
Subject
<ecblank.gif>
Re: Integration of HTM
<ecblank.gif><ecblank.gif>


On Feb 3, 2010, at 5:49 PM, Ian Hickson wrote:

> On Wed, 3 Feb 2010, Richard Schwerdtfeger wrote:
>>
>> We are calling it the accessible DOM for canvas. It starts and ends with
>> the <accessible></accessible> tags and it is not visually rendered.
>
> I really don't think this is a good idea, as explained in the following
> e-mails:
>
>   http://lists.w3.org/Archives/Public/public-html/2010Jan/0488.html
>   http://lists.w3.org/Archives/Public/public-html/2010Jan/1151.html
>   http://lists.w3.org/Archives/Public/public-html/2010Jan/0931.html
>
> I do not think it is necessary to have multiple inline alternatives for
> <canvas>, nor do I think it is necessary for widgets that represent the
> graphically-rendered widgets on a <canvas> to be marked up separately from
> an inline alternative representation. The existing features of HTML
> already allow us to have multiple alternatives. Adding more features for
> this is IMHO a mistake.

I agree. I don't think the <accessible> tag is an improvement. In the common case, the same content can work as an accessible DOM and as fallback content. And that's also the model for other elements that use fallback content partly for accessibility purposes (e.g. <object>). I don't see the case for making canvas accessibility intrinsically more complicated.

Regards,
Maciej