RE: [wcag2] Layout tables

Regarding the suggestion to ban the use of layout tables outright:

This is a difficult one for us working on techniques. We all agree that
layout tables are undesirable and we would like to "ban" them. But the
techniques aren't normative so we can't actually forbid the use of any
technology or technique that leads to WCAG 2.0 compliant content. If we try
to do so, authors can simply implement layout tables in an accessible manner
and say they achieved conformance using techniques not provided by W3C. This
is a practice that our conformance scheme will have to permit because we
probably won't think of every possible way to make a given technology
accessible, nor will we provide techniques for every technology used on the
Web. And as a matter of practicality, whether our techniques describe the
accessibility of layout tables or not, authors will use them. Failing to
provide techniques will simply leave those authors without guidance from
W3C. They will either learn about the accessibility issues of layout tables
from other sources, or will assume that since we did not comment on them
there are no accessibility issues.

So the approach we are taking is to say "we'd rather you don't use layout
tables, but if you're going to here's how to make them accessible." That's
more useful on the whole than saying "since we ban layout tables we don't
provide techniques to make them accessible", leaving authors in the dark and
potentially resulting in reduced accessibility in practice.

This is an example of a class of issue we face in the techniques, in which
we describe accessible uses of undesirable techniques. It is best
exemplified by the EMBED element [1], [2], [3]. EMBED is not part of any
formal HTML specification, therefore its use is a violation of the guideline
to use technologies according to specification. However, EMBED is widely
supported and is in fact better supported than the within-spec alternative,
OBJECT. Therefore we decided we have to provide guidance for authors on
accessible use of the element, even though we also want to encourage authors
to use alternatives, when user agent support is sufficient.

In spite of what I just said, I do agree that the techniques should oppose
the use of layout tables, even though we provide techniques for using them.
That opposition probably needs to be much stronger, more clear, and more
pervasive in our document.

Michael

[1] http://www.w3.org/WAI/GL/WCAG20/WD-WCAG20-HTML-TECHS/#noembed
[2] http://www.w3.org/WAI/GL/WCAG20/WD-WCAG20-HTML-TECHS/#embed-alt
[3] http://www.w3.org/WAI/GL/WCAG20/WD-WCAG20-HTML-TECHS/#embed

Received on Friday, 16 April 2004 13:06:41 UTC