Re: additional role of 'hint' or 'contexthelp' -- please you do it

Hi, Al-

We have removed all mention of "contexthelp" from the SVG Tiny 1.2 spec,
along with that proposed functionality [1].  We may reintroduce the
functionality somehow in a later specification, or in the accessibility
module, after further examination of the best way to provide it (such as
with a dedicated element as originally proposed, or indeed as a role).
We will keep you in the loop.

For what it's worth, I disagree with one aspect of the design philosophy
of ARIA... specifically, that it assumes that the only use of ARIA
should be scripted.  This does not take into account the likely
proposition that some UAs will provide native behavior for ARIA-tagged
elements.  A good way to future-proof this might be to expose an API
that authors can use to detect if there is such support, so that authors
could provide fallback (scripted) behavior only for those roles that are
not supported, thus providing a means to transition.  Another point is
that ARIA should be useful even without scripting; in a way, ARIA roles
are themselves an API to underlying AT, so even without behavior, adding
a 'contexthelp' role would expose that functionality to AT that
understands it.

[1] http://lists.w3.org/Archives/Public/public-svg-wg/2008OctDec/0331.html

Regards-
-Doug

Al Gilman wrote (on 11/20/08 10:03 AM):
> 
> BTW, congratulations on SVG 1.2T to PR!
> 
> .. back on topic...
> 
> We (PFWG) and SVG have been discussing the use of a new @role
> value for on-the-spot, tailored context help for widget
> instances.
> 
> http://www.w3.org/WAI/PF/Group/track/issues/75
> 
> On November 11 the Working Group asked me to tell you that
> we prefer you define this @role value in the SVG documents.
> 
> http://www.w3.org/WAI/PF/Group/track/actions/302
> 
> While we can see the benefit to SVG, we don't need this role
> where the behavior is scripted.  Only once it is built into
> the browser or processor does it become necessary.  There it
> is a matter of the native semantics of the host language,
> not a WAI-ARIA overlay.  So it makes sense that it be defined
> in the document controlling the host language.
> 
> This is not meant to cut off discussion, but as this is
> technically feasible, we hope you will give it your serious
> consideration.
> 
> Al
>

Received on Saturday, 22 November 2008 20:30:53 UTC