W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > October to December 1999

Re: Stylesheet columnisation

From: Leonard R. Kasday <kasday@acm.org>
Date: Tue, 02 Nov 1999 13:38:25 -0500
Message-Id: <>
To: Kynn Bartlett <kynn-hwg@idyllmtn.com>, "Paul Bohman" <paulb@cpd2.usu.edu>
Cc: "GARETH P PARKINSON" <298gpp@tay.ac.uk>, <W3c-wai-ig@w3.org>
Here's my short and extended opinions about using tables for layout.

Short opinion:

Given the current browser situation, we can't use style sheets for layout.
Instead, use a table entirely for layout or entirely for data, but don't
mix these uses in one table or nest data tables inside layout tables.  (Or
avoid table layout altogether unless it serves a real purpose).

Also, help the reader identify when a table is used for data by:
1. Including a caption of the form "Table of blah blah blah..."
2. Using header cells in the top row.

Extended opinion:

1. Another advantage of tables is that with current tools it's easy to
control the order in which the contents are read, because it's directly
determined by the table layout.  Tools that use CSS for layout may produce
a reading order that's very different than the visual order. This is
because e.g. if you slide text blocks around on the screen, all the tool
does is change the coordinates, not the reading order.  This happens for
example with Microsoft Publisher.  Of course, it's straightforward to do
this if you're writing raw HTML and CSS by hand.  But you run into problems
with some visual type editing tools.

2. It's true that using stylesheets for layout instead of tables
theoretically gives the screen reader a way to deterine if it's really a
data table or layout control.  However, this could easily be done without
style sheets, e.g. by requiring a caption on all data tables (even a null
caption), or defining a class.

3. The current author guidelines permit tables for layout until browsers
shape up, but require that they make sense when read in the order of the
raw HTML ("linearized").

4. Therefore, perhaps we should permanently allow the use of tables for
layout provided that a standard way is agreed on to distinguish layout from
data use.  We've got time to think about this since it doesn't impact what
we're doing right now.


At 10:55 AM 11/1/99 -0800, Kynn Bartlett wrote:
>At 10:57 AM 11/1/1999 , Paul Bohman wrote:
>>I noticed that, although you are proficient at CSS layout, you are still
>>reluctant to use CSS for positioning. For example, the HTML Writers Guild is
>>built on table layouts and the Aware page (http://aware.hwg.org/) avoids
>>layouts that would require either tables or CSS positioning.
>>Even though I really like the concept of CSS, I have my doubts about its
>>usefulness until browsers give it better support.
>This is the crux of the matter.  CSS is not widely supported enough,
>nor reliably supported enough, to be able to use CSS reliably for
>layout.  In the case of the HTML Writers Guild, there's an extra
>design consideration involved in that while it's okay to look "different"
>in various browsers, we can't look "bad" in any of them, and if you
>use CSS for positioning you take a serious risk of looking "broken"
>in some browsers.
>(Most users, when they encounter a page that doesn't look right, will
>think the page is poorly designed, not that their browser is deficient.
>So the HWG site has to be created in a way that it will look "right"
>Kynn Bartlett                                    mailto:kynn@hwg.org
>President, HTML Writers Guild                    http://www.hwg.org/
>AWARE Center Director                          http://aware.hwg.org/
Leonard R. Kasday, Ph.D.
Institute on Disabilities/UAP, and
Department of Electrical Engineering
Temple University

Ritter Hall Annex, Room 423, Philadelphia, PA 19122
(215) 204-2247 (voice)
(800) 750-7428 (TTY)
Received on Tuesday, 2 November 1999 13:35:26 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 13 October 2015 16:21:06 UTC