W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 1998

Re: Tables - my thoughts

From: nir dagan <dagan@upf.es>
Date: Thu Jul 23 10:25:44 1998
Message-Id: <199807231030.MAA00513@sahara.upf.es>
To: jongund@staff.uiuc.edu
Cc: w3c-wai-gl@w3.org
Jon wrote:
> Has there been any discussions of labeling tables for their purpose?
>  For example indicating a table is for column formatting or for data. A
> crude way would be to use the "class" attribute with "wai-column" or
> "wai-data".  A user agent could then make some assumptions about
> presentation of the table information. 

In my view there is a certain balance of what user 
agents should do and what authors should do. W3C recommendations
allow authors to write informative accessible tabular information
using HTML4.0 tables, and arrange text-blocks side by side using CSS2.

Authors should write using these recommendations (both of which degrade 
gracefully to HTML3.2 browsers that do not support stylesheets) 
and user agents should support these recommendations.

When an author is using a table for layout justifying it by misusing 
the term "backward compatibility",  he is violating the 
first principle of HTML writing of separation of meaning
from presentation. An author who hacks in this way may counter-
hack in his *author's* aural stylesheet to get his layout-tables 
to be converted to normal blocks by converting TABLE, TBODY, TR, TD 
into DIV via CSS2 display properties.

One should keep in mind though that HTML browsers who support CSS2 may 
ignore (according to the CSS2 spec.) the display property. Also the CSS2 spec.
recommends to HTML authors not to use the display properties, as the HTML
elements have commonly agreed meanings. 

In other words, it is better not to hack, than to hack and then 
counter-hack, hoping that the so called disabled user have the latest 
XML-CSS2 browser.

Nir Dagan.
Received on Thursday, 23 July 1998 10:25:44 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:27 UTC