Re: Comment on WAI-ARIA Role: Needs "hint" Role

Hi, Rich-

Richard Schwerdtfeger wrote (on 9/4/08 5:26 PM):
> 
>> > Information on how to interact with an object would use the
>> description role.
>>
>> No, that's a misuse of the "description" role.  A "description" should
>> be suitable as a textual replacement for the image, in case the image in
>> not available or perceivable.
> 
> That is not how we designed it. Description is designed to replace
> longdesc, provide additional help information, and
> link to prose on the page if necessary.

Everything I've read about "longdesc" matches the use I described above:

* http://www.w3.org/TR/REC-html40/struct/objects.html#adef-longdesc-IMG
* http://www.w3.org/TR/WCAG10-HTML-TECHS/#long-descriptions

Maybe there's some subtle point I'm missing.  Are you saying that
"description" should or should not provide "hint" information?  I
contend that it should not.


> I see. The problem with that is the platform accessibility API have no
> single hint. Looks like you modeled hint
> after XForms or perhaps the other way around. 

I suspect you're right that the <hint> element planned for SVG 1.2 Full
was informed by XForms, and in turn informed our desire to use a role
instead of an element for that purpose.

It seems a useful distinction to me... am I wrong here?


> The one problem I see is that with script you could
> potentially have multiple actions and consequently multiple "hints" -
> one corresponding to each action. Have you
> thought about this for SVG or do you always expect there to be one action?

It would be up to the author to make judicious use of hints.  We aim to
provide the infrastructure to enable them to do so in a structured and
well-defined way.  Having a "hint" role defined in ARIA is the logical
way to effect that.


> We do have, in some accessibility api, is an action description which in
> effect would be your "hint."

Ok, that seems like a place to start... how would this relate to what's
specified in ARIA?  Will there be a role attribute value for "action"?
Otherwise, where would the action description be placed?


> So, for SVG there should not be an overlap. We relegate desrcliption to
> "A brown horse, blah blah."
> ARIA, consequently has no equivalent for hint so for SVG the mapping
> should be to the corresponding action. This is typically short though.
> 
> Make sense?

Actually, not really.  What do you mean by, "the mapping should be to
the corresponding action"?  In what way?  What form would the
information about the "corresponding action" take?  How will it be
exposed to the UA or user?  Is this a text description, as I described,
or is it somehow inferred from key mappings, script listeners, etc.?

Maybe you could provide an example, to clarify?


> We can clarify this in our spec that description should not be used to
> describe an action.

That sounds good, but this seems to potentially conflict with how you
defined the "description" role above.

Regards-
-Doug Schepers
W3C Team Contact, WebApps, SVG, and CDF

Received on Thursday, 4 September 2008 22:41:53 UTC