micro-process on @headers and friends [was: Re: how to collaborate with the HTML WG (what works)]

Caveat: this is my interpretation of the state of the process.  I  
haven't reviewed the
HTML WG minutes to dot i's and cross t's.

On 7 Nov 2008, at 6:17 AM, Steven Faulkner wrote:

> Hi al,
>
> >Having removed some of the spurious sources of apparent  
> difference, we were able to report into the HTML WG meeting on  
> Thursday a status of the issue that had >some basic agreements  
> about requirements and a plan for next-steps action.  The HTML WG  
> agreed.
>
> Unfortunately I was not able to be there on thursday, after having  
> read the minutes of the html wg meet in relation to table headers,  
> I didn't get a sense that the HTML WG were any closer to agreement  
> on issues surrounding the implementaion of headers/id in the spec.

We weren't any closer after Thursday than after Tuesday in terms of  
stating the
feasible range for a solution.  Or any farther away.  What we worked  
out on
Tuesday survived intact.  What we had that was new was that we had  
_both_ groups on
board with where we wound up Tuesday.  The observers on Tuesday were  
not in a position
to commit the HTML WG, but they definitely helped us get to where we  
could take
out story into HTML WG and have it take its place in their working  
baseline without a ripple.

> Can you provide a clarification?
> Specifically was agreement reached that "@headers pointing to TD"  
> or "TH with @headers pointing to another TH" need to be made  
> conformant in the HTML5 spec in order to provide the required  
> accessibility hooks for current AT to be able to convey  
> relationships in complex data tables to users? Or one of these  
> mechanisms will still be required for irregular data tables, once  
> the 'smart headers algorithm' is implemented in popular browsers  
> and future versions of AT?

No.  The progress was in process dimension.  Establish next steps.   
Ben Millard got the action to put together a proposal that includes  
both markup and algorithm aligned.  He is supposed to circulate this  
to XTECH and public-html for a hives check, and then it goes to Hixie  
to craft spec language and integrate into the Editor's draft.  After  
which the normal review process within HTML WG with PFWG looking over  
their shoulder continues.

Al

> regards
> steve faulknert
>
> 2008/11/4 Al Gilman <Alfred.S.Gilman@ieee.org>
>
> Al had some hallway conversations with members of the TAG who  
> expressed concern
> as to how to interact with the HTML WG.  Since we think we have a  
> case study
> that reflects successful collaboration, we would like to share it  
> more widely.
>
> ** summary
>
> PFWG and HTML WG achieved constructive collaboration at TPAC.  The  
> two factors that allowed us to do this were:
>
> (a) framing a workable topic for discussion.  right-sizing (and  
> shaping) the bite-sized topic.  Find a semi-separable design  
> tradeoff, not necessarily a single markup feature or requirement.
>
> (b) separating demand issues (user and author factors) from supply  
> issues (markup-and-processing options).
>
> ** details
>
> (a) the topic:
>
> The topic we want to focus on here was "associating table cells  
> with the content context on which their interpretation depends."   
> Not just the @headers attribute; that's too small a topic because  
> the performance against user needs depends on client processing and  
> to some degree @headers competes with @scope.  But not all table  
> markup of interest to accessibility; that's too broad.  The  
> associations issue is sufficiently decoupled from other issues once  
> the @scope, @headers, and browser algorithms topics are included.
>
> (b) supply and demand:
>
> Separating what the user needs in operational terms (in the user  
> experience) from markup options allowed us to make incremental  
> progress working from the baseline started with the thread at  
> http://lists.w3.org/Archives/Public/wai-xtech/2008Sep/ 
> thread.html#msg304
>
> On Tuesday in the PFWG meeting, the HTML WG observers confirmed the  
> PFWG participants' sense that multiple levels of context  
> information are part of the requirement to cover real tables as  
> they are widely used.  We worked through differences in vocabulary,  
> realizing that 'nested headers' or 'chained headers' were both ways  
> of talking about this same phenomenon.  This was progress.
>
> The PFWG participants were able to confirm the HTML WG observers'  
> idea that "@headers pointing to TD" and "TH with @headers pointing  
> to another TH" were both markup patterns that, married with the  
> right sort of browser and authoring processing, could meet this  
> requirement.  This was progress.
>
> Having removed some of the spurious sources of apparent difference,  
> we were able to report into the HTML WG meeting on Thursday a  
> status of the issue that had some basic agreements about  
> requirements and a plan for next-steps action.  The HTML WG agreed.
>
> Al Gilman, co-chair, PFWG
> Janina Sajka, co-chair, PFWG
>
> Coordination note: Chris Wilson and Mike Smith, co-chairs of HTML  
> WG have seen this and
> agree it is accurate.
>
> </draft>
>
>
>
>
>
> -- 
> with regards
>
> Steve Faulkner
> Technical Director - TPG Europe
> Director - Web Accessibility Tools Consortium
>
> www.paciellogroup.com | www.wat-c.org
> Web Accessibility Toolbar - http://www.paciellogroup.com/resources/ 
> wat-ie-about.html

Received on Friday, 7 November 2008 16:27:21 UTC