W3C home > Mailing lists > Public > www-html@w3.org > May 2007

Re: Complex Table Examples

From: Ian Hickson <ian@hixie.ch>
Date: Mon, 14 May 2007 21:35:21 +0000 (UTC)
To: Laura Carlson <laura.carlson@gmail.com>
Cc: www-html@w3.org, "Patrick H. Lauke" <redux@splintered.co.uk>
Message-ID: <Pine.LNX.4.62.0705142111490.11553@dhalsim.dreamhost.com>

On Mon, 14 May 2007, Laura Carlson wrote:
> > 
> > I've already asked 3 times for examples to be provided that would 
> > support the addition of the headers attribute (and provided some 
> > myself), trying to help you strengthen you case for adding it, but 
> > each time you have bypassed question entirely.
> 
> [...]
>
> http://www.jimthatcher.com/webcourse9.htm#wc9.3
> http://www.webaim.org/techniques/tables/data.php#id
> http://www-03.ibm.com/able/guidelines/web/webtableheaders.html#complex
> http://tinyurl.com/qmm8b
> http://tinyurl.com/2fsnkk
> http://www.ittatc.org/training/webcourse/WAWCSection9.php#WebCourse9.3
> http://tinyurl.com/yozae9
> http://www.w3.org/TR/WCAG20-TECHS/#H43

I'm not sure exactly what Lachlan meant, but for what it's worth the 
examples that I personally (as the editor of the spec) am interested in 
are examples of actual real-world usage.

The cost of _including_ headers="" in the spec is that if people misuse 
it, the attribute becomes worthless, and we actually end up *hurting* 
accessibility rather than helping it, by having the tools that are 
intended to help accessibily actually result in *worse* performance when 
they take these attributes into account.

A recent study I'm aware of that looked at actual use of the "summary" 
attribute, for example, suggests that "summary" is almost universally 
abused:

   http://canvex.lazyilluminati.com/misc/summary.html

(For instance, saying summary="This table is for layout only" is a 
complete waste of time -- if the author cares about accessibility, then 
he should not use the table for layout in the first place. Effectively, 
that value just says "Hi, I'm too lazy to bother making my page usable for 
you, but I'm going to pretend I care".)

The same kind of study for headers="" would be useful. In particular, it 
would be extremely useful to see if, in a sample of several thousand or 
million pages, the pages that used headers="" used them in (a) a way that 
was compliant (e.g. pointing to IDs actually in the same table, no 
circular references, etc), and (b) a way that wasn't redundant with the 
algorithm defined in the HTML 5 spec.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 14 May 2007 21:35:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:10 GMT