roadmapping checking for HTML+ARIA. [was: Re: conformance checker for HTML+ARIA?

On 30 Sep 2008, at 9:11 AM, Michael(tm) Smith wrote:

>
> Hi Steven,
>
>> @2008-09-30 14:00 +0100:
>> The idea in regards to the HTML WG was brought about by david  
>> dorwards
>> suggestion that it is something that the WG charter covers.
>>
>> I certainly understand if this is not something that the HTML WG  
>> to be
>> actively involved in , but it is good to get your feedback so if  
>> there
>> is any development, its result does not upset any applecarts within
>> the HTML WG.
>
> I think having a normative spec for ARIA is HTML 5 is absolutely
> something that the HTML WG needs to produce. But having a spec of
> HTML 4.01 + ARIA is not. And if the need to produce a HTML 4.01 +
> ARIA DTD and have support in validator.w3.org for it is urgent,
> then I really don't think taking a discussion about it to the HTML
> WG as whole to try to get consensus on it is what you want to do.

** summary:

* PFWG will take the next turn; please join me in finding the right
HTML WG observers to attend PFWG sessions at TPAC.

* WAI-ARIA can deploy with 'experimental' checking of HTML+ARIA  
documents.

* HTML WG is responsible for publishing HTML specs including an HTML  
that implements WAI-ARIA and is the normative reference (with a lot  
of include-by-reference provisions citing the WAI-ARIA spec).  PFWG  
will work on clarifying HTML semantic interaction with ARIA  
semantics, but we need HTML WG to work with us on this.

** at greater length:

I want to subscribe to most of what you and Aaron have said, here.

Let me start by saying what we in PFWG are doing.

* TPAC agenda gives snapshot

Our agenda for our face-to-face on 20, 21 October at the TPAC  
includes sessions that touch on this.

The Monday session 16:00 - 17:50 on implicit ARIA semantics of
host languages, we would especially welcome observers from HTML, SVG.
Here we look at "what are the appropriate constraints in a host+ARIA
profile?"

The Tuesday session 08:45 - 10:30 on ARIA Implementation
we would welcome observers from HTML, Backplane XG, SVG
This includes checking technology, how to afford authors a
checking capability.

We hope SVG and AUWG between them can come up with some meeting
time Thursday or Friday to address validation of SVG+ARIA documents

If HTML were to put this on their agenda, we would support you with
observers.

* roadmap view

Yes, the checking we need now is not the same timeframe as W3C  
checking to W3C Recommendation.

I anticipate publishing some sort of checking schema or reference  
document for authors as an adjunct to the WAI-ARIA Test Suite.  This  
would be something that runs in validator.w3.org or validator.nu or  
some widely available tool.  It would not be part of the normative  
specification.  You could call it 'experimental.'

While HTML WG clearly has the responsibility for the normative  
statement on HTML5+ARIA, individual participants in HTML WG have  
asked PFWG for PFWG to be more propositional about the interaction of  
HTML and ARIA semantics, and I put the above agenda item to treat  
this request seriously.  In my view,
we will be saying things about the interaction of HTML markup and  
ARIA markup in the binding to accessibility APIs.  This will be  
offered to HTML WG as a courtesy to aid them in their Last Call  
review of WAI-ARIA, to give them a better handle on the technical  
risks associated with creating said normative statement on HTML5+ARIA  
with the proposed definition of WAI-ARIA.

* demurrur, separation of concerns:

This is not to say HTML WG doesn't need to pay more attention to  
providing documentary and checker support to the evolution of HTML.   
HTML WG has been incautious in allowing features important for  
accessibility in today's HTML to be removed from HTML5 drafts and  
hence checking in validator.nu to an HTML5+ARIA profile will reject  
current best practice for accessibility.  But the 'checking HTML 
+ARIA' topic doesn't need to become captive to ratholing on that  
issue.  I claim our roadmap for WAI-ARIA adoption is not contingent  
on having a normative W3C profile for HTML+ARIA.  Interoperable  
browser implementations and experimental author/checker tools should  
suffice to get a critical mass of successful (cross-browser for  
authors, delivers all functionality usably for people with  
disabilities) early-adopter content out on the Web to where everyone  
else can see that it's doable and begin to figure out how to do it.

Al

PS: distribution note: dropping www-archive as www-validator and wai- 
xtech are already publicly archived.


>   --Mike
>
> -- 
> Michael(tm) Smith
> http://people.w3.org/mike/
>

Received on Tuesday, 30 September 2008 15:16:51 UTC