- From: Gez Lemon <gez.lemon@gmail.com>
- Date: Fri, 26 Sep 2008 09:32:08 +0100
- To: "Simon Pieters" <simonp@opera.com>
- Cc: "Ben Millard" <cerbera@projectcerbera.com>, "James Graham" <jg307@cam.ac.uk>, "W3C WAI-XTECH" <wai-xtech@w3.org>, HTMLWG <public-html@w3.org>
2008/9/26 Simon Pieters <simonp@opera.com>: > # [15:46] <zcorpan_> it doesn't work with > http://projectcerbera.com/web/study/2007/tables/clark2006/06-gui/minimal > # [15:56] <zcorpan_> > http://projectcerbera.com/web/study/2007/tables/clark2006/19-movies/minimal > is another case where it fails > > It worked correctly for those tables without the newly implemented rule. > > Some observations: > > * Those two were columns -- Gez had rows. > > * Those two had the relevant header cells at the top of the table -- > Gez had data cells to the left of the relevant header cells. > > I guess we need a bigger sample of tables to know what to do for this > particular case. Could this pattern of cells be detected by a script? If the direction could be determined reliably, this could be a great algorithm. I noticed that "Property" wasn't reported as a header in the last example [1], but I'm sure it was reported properly in earlier versions of the table inspector. [1] http://juicystudio.com/wcag/tables/nestednoheaders.html Gez -- _____________________________ Supplement your vitamins http://juicystudio.com
Received on Friday, 26 September 2008 08:32:51 UTC