Re: Current PFWG thinking on ARIA integration in host languages

Yes, your distillation sounds correct, with one quibble. The statement
"(Invisible metadata is evil, and should be avoided in favor of visible
metadata whenever possible.)" is not something I would want ascribed to
me. I am not taking a position on that, and particularly am not
expressing a value such as "evil". I'm hoping you included that to
facilitate your own understanding only. I think your distillation does
not depend on that statement, but am pointing it out to be sure. Aside
from that, your distillation seems to be an accurate representation of
our current position. Michael

Tab Atkins Jr. wrote:
> I'm gonna boil down your email just to make things simpler for me;
> please correct me if I get something wrong.  Your current ideas are:
>
> 1. Host languages can define implicit ARIA 'state', 'property', or
> 'role' semantics for their elements.
> 2. If an element has an implicit ARIA 'state' or 'property' semantic,
> authors SHOULD NOT specify an ARIA version.  (Invisible metadata is
> evil, and should be avoided in favor of visible metadata whenever
> possible.)
> 3. If there is a conflict in a 'state' or 'property' semantic between
> ARIA and the host language's implicit semantics, the host language
> wins.
> 3. If there is a conflict in a 'role' semantic between ARIA and the
> host language's implicit semantics, ARIA wins. (This is to allow
> authors to fully repurpose elements in ways not anticipated by the
> host language.)
> 5. Host languages can define 'strong' implicit ARIA 'role' semantics.
> These still lose when put in conflict with an explicit ARIA 'role'
> semantic, but such conflicts can be flagged as an error by a
> conformance checker.
>
> Is this all correct?
>
> If so, sounds good to me.
>
> ~TJ
>
>
>   

-- 

Michael Cooper
Web Accessibility Specialist
World Wide Web Consortium, Web Accessibility Initiative
E-mail cooper@w3.org <mailto:cooper@w3.org>
Information Page <http://www.w3.org/People/cooper/>