Re: http://www.isolani.co.uk/blog/access/ScreenReadersAWebDeveloperFailure

On Fri, 14 Sep 2007, Sailesh Panchang wrote:
>
> You refer to "the algorithms for determining header cell / data cell 
> association, and the algorithms for determining section heading 
> associations" when you responded to Mike. Please can you point me to 
> these algorithms?

I'm not sure exactly which e-mail you are referring to, so I'm not sure 
exactly what the context is, but if I was referring to the HTML5 
specification then the algorithms in question are those given in the 
subsections here:

   http://www.whatwg.org/specs/web-apps/current-work/#processing
   http://www.whatwg.org/specs/web-apps/current-work/#headings
   

> Also one does see use of structural markup in layout tables at times… so 
> is there an algorithm to first distinguish layout tables from data 
> tables?

The HTML5 spec doesn't have such an algorithm currently, though there have 
been proposals along these lines. In HTML5 it is never valid to use a 
table for layout purposes, so this would effectively be a conformance 
check. It's not clear exactly how to determine what makes a layout table 
programatically.


> With regard to section heading associations, the problem is two 
> fold: absence of markup for headings and use of markup for pieces of 
> text that are not headings. Will the algorithm be useful to only detect 
> headings and content association when heading markup is used correctly?

Yes. That is, in fact, why conformance is important.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'