Re: Please vote on the canvas accessibility proposal

On Wed, 24 Feb 2010, Steven Faulkner wrote:
> 
> do you have any evidence to the contrary?

The one <canvas> I use on a regular basis (not a demo) has accessible 
fallback and no adom="" attribute. Which is more common is essentially 
impossible to tell from purely anecdotal evidence.

   http://www.whatwg.org/issues/data.html?period=1


> >The point is that if the author doesn't care about conformance, there's 
> >the possibility that the author will specify adom="" even if the 
> >content is inappropriate for ATs, and equally a possibility that the 
> >author will _not_ specify adom="" even if the content _is_ appropriate 
> >for ATs.
> 
> how is the probability equal?

I didn't say the probability was equal.


> does any data support that attribute use follows this pattern of 50% 
> inappropriate use?

Actually data for similar attributes -- longdesc="" and summary="" come to 
mind -- show that the attributes are misused vastly more often than 50% of 
the time.


> there is data available to show that provision of accessible fallback 
> for canvas is pretty much zero.

In demos. It's unclear what the right fallback would be when the whole 
point of the canvas is to show off the canvas for its own sake. Demos are 
not really representative here. (Arguably, "you don't have canvas" is 
actually the right thing to say in this case, in fact.)


> >So adom="" is either redundant, or possibly inaccurate.
> 
> how so?

It's possibly inaccurate for authors who don't follow the spec, and it's 
unnecessary for those who do (since they can just make the page do the 
right thing in both cases).


> I can foresee instances where the developer provides an accessible
> alternative outside of the canvas (currently conforming no?) and wants to
> tell users of browsers that don't support canvas that they are missing
> something:
> 
> <canvas> you cannot see the graph because your browser does not support
> canvas </canvas>
> <!-- table containing data represented in graph -->

In this case, if canvas is available, then the author can trivially just 
remove the contents of the canvas in one line of code. adom="" isn't 
useful for hiding this text.

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

Received on Wednesday, 24 February 2010 10:07:03 UTC