Re: Agenda: HTML 5 Canvas Accessibility Meeting February 22, 2010

Ian, this is not a matter of trusting the author. It is simply providing a
vehicle for the author to convey their intent for the sub-tree. I am also
not making any assumptions about the intent of the fallback content. What I
am doing is allowing the author to make such a statement much the same way
that an author is now provided the ability to indicate if a form element is
required in HTML. Before it required guessing based on best practices.
Could the author get it wrong. ... sure.

I should point out that setting adom="true" is not a conformance claim. It
basically says that when an author is rendering <canvas> the accessibility
subtree is an accessible synchronized representation of what is rendered on
the canvas. Currently, the author has no vehicle to indicate whether there
is a correlation between what is in the sub-tree and what is being
rendered.

Could the fallback content be the same as what would be used for adom.
Sure, with the caveat that if there if keyboard focus can be had in it that
the script be enabled to ensure that focus be visibly drawn on the canvas.
It was your comments about this which caused me to use the attribute
approach as the content could in fact be the same. Thank you for those
comments.

Rich

Rich Schwerdtfeger
Distinguished Engineer, SWG Accessibility Architect/Strategist


                                                                           
             Ian Hickson                                                   
             <ian@hixie.ch>                                                
                                                                        To 
             02/22/2010 02:06          Richard                             
             PM                        Schwerdtfeger/Austin/IBM@IBMUS      
                                                                        cc 
                                       "public-canvas-api@w3.org"          
                                       <public-canvas-api@w3.org>          
                                                                   Subject 
                                       Re: Agenda: HTML 5 Canvas           
                                       Accessibility Meeting February 22,  
                                       2010                                
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           




On Mon, 22 Feb 2010, Richard Schwerdtfeger wrote:
>
> Without the adom approach we have no way of providing a directly
accessible
> solution for canvas that is testable for compliance.

If you trust the author to use adom="" correctly, why don't you trust the
author to write the accessible solution correctly? Or equivalently, why do
you trust that if the use has specified adom="", the fallback is
accessible? If you are saying that you don't assume this, but that the
accessible content will always be further tested by inspection with an AT,
then why can't you use that testing methodology regardless of the presence
or absence of the adom="" attribute?

You also seem to be assuming that the fallback content and the accessible
content cannot ever be the same. My apologies if you are not assuming
this. If you _are_ assuming this, then I must protest. I would posit that
in many cases (a more-or-less static picture with no semantic
interactivity), maybe the most common case, they will be the same.

--
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/,       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'