RE: Drop longdesc, get aria-describedat?

Schnabel, Stefan, Fri, 9 Mar 2012 12:45:02 +0100:
> Therefore,
> * For  ARIA 1.1 (not 1.0 anymore), the aria-describedat 
> property should be verbalized/formalized ASAP
> 
> * There should be high pressure already during 2012 on talks 
> with the AT and UA vendors for early support implementations 
> available until end of the year
> 
> * For the time being, longdesc could  remain as part of the 
> HTML5 spec with unchanged usage statements (to address Laura's 
> warning issue)

By 'unchanged', I suppose you referred to Laura's CP: That it should be 
adopted, unchanged w.r.t. the normativity that it suggests: Longdesc as 
fully conforming, right now.

> * When solid AT/UA support for aria-describedat is there, the 
> longdesc  "conforming but obsolete declaration"  will be appropriate 
> and pointing to the newer recommended technique (using aria- 
> describedat) then makes sense

+1 Sounds like a plan. It could allow us to keep the CP roughly as is.

W.r.t. what PF should do in ARIA 1.1, then I would give them permission 
to - and give them permission to - define @longdesc as an ARIA 
attribute. That is: It should be up to PF whether to EITHER turn 
@longdesc into an ARIA attribute, OR whether to go for a a shiny new 
attribute [and simply define how @longdesc should be handled from an 
ARIA perspective]. If reused directly, then I suppose @longdesc would 
become like the @role attribute: Without any prefix. Whether it would  
be wise to reuse it, probably relies on an evaluation of the legacy. 
Priority should be an attribute that works the way we want it to work.
-- 
Leif Halvard Silli

Received on Friday, 9 March 2012 12:26:02 UTC