W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > January to March 2002

Re: more CSS and tables

From: Charles McCathieNevile <charles@w3.org>
Date: Sat, 5 Jan 2002 20:35:39 -0500 (EST)
To: Joe Clark <joeclark@joeclark.org>
cc: WAI-IG <w3c-wai-ig@w3.org>
Message-ID: <Pine.LNX.4.30.0201052015150.15467-100000@tux.w3.org>
On Sat, 5 Jan 2002, Joe Clark wrote:

  >Well, you haven't demonstrated that the content being laid out in a
  >table would have any better identification,

  I guess <tr><td></td></tr> counts for nothing.

Well, in terms of being a better identification of the purpose of an element
on the page, it counts for nothing more than div.

  >In addition, the use of tables for layout purposes is an express
  >violation of the specification of HTML 4,

  Not exactly:

  <http://www.w3.org/TR/html4/struct/tables.html>

  >>Tables should not be used purely as a means to layout document
  >>content as this may present problems when rendering to non-visual
  >>media. Additionally, when used with graphics, these tables may
  >>force users to scroll horizontally to view a table designed on a
  >>system with a larger display. To minimize these problems, authors
  >>should use style sheets to control layout rather than tables.


  Unless we have entered a world of Newspeak where "should not" means
  "must not," the spec says no such thing. And in any event, if tables
  are used for layout and do not "present problems when rendering to
  non-visual media," the putative reason not to use them vanishes.

Well, should not is not as strong as must not, but it provides a clear
direction about the spirirt of the specification. However I had misremembered
the specification text.

So what I should have said is more like "it is strongly recommended not to do
this in the specification", or somesuch. And yes, if they don't present
problems, the putative accessibility reason does vanish. (As I said - I don't
think going against the spec on this one thing alone causes empires to fall,
but it doesn't augur well for following the rest of what the spec says -
notably some important stuff about marking up structure so its role can be
identified and used, rather than simply providing a lot of visual clues)

  Besides, every screen reader save for OutSpoken for Macintosh can
  handle tables. Mobile phones are not in the same category; they are
  not adaptive technology. And, in an irony unremarked by many, the
  only tables Lynx really mangles are data tables; typical layout
  tables linearize just fine.

Yes, it is a problem that lynx doesn't handle tables of anything well, and
data tables much worse than many simple layout tables. When WCAG 1.0 became a
recommendation tables were a much bigger problem, and it was suggested that
people could get a copy of lynx to linearise them, so the problem was of low
priority. It seems to me that the requirement to learn a new browser is a
pretty steep one (any new browser. I find it a challenge to figure out IE,
and something of a challenge with Netscape / Mozilla since I use other
browsers much more, and have to try to recall the special terminology that
those two apply for various odd things - sadly it seems to be the things for
which I decide I want to test something with them).
Received on Saturday, 5 January 2002 20:35:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:13:59 GMT