W3C home > Mailing lists > Public > public-html@w3.org > March 2009

Re: Draft text for summary attribute definition

From: Robert J Burns <rob@robburns.com>
Date: Sat, 28 Feb 2009 20:27:59 -0500
Cc: HTMLWG WG <public-html@w3.org>, Leif Halvard Silli <lhs@malform.no>, Steven Faulkner <faulkner.steve@gmail.com>, W3C WAI-XTECH <wai-xtech@w3.org>, Janina Sajka <janina@rednote.net>, Sam Ruby <rubys@intertwingly.net>
Message-Id: <6482762C-8823-4E16-9FA1-553B54B230B7@robburns.com>
To: Gez Lemon <gez.lemon@gmail.com>
Hi Gez,

On Feb 28, 2009, at 7:51 PM, Gez Lemon wrote:

> 2009/2/28 Robert J Burns <rob@robburns.com>:
>> The substantive differences between version A and version B is that  
>> version
>> B follows WCAG in recommending summary="" for layout tables and  
>> also directs
>> authors to include a caption within 'summary' if the author elects  
>> to omit a
>> caption entirely (though prohibiting the use of 'summary' for  
>> caption when
>> the author includes 'caption' content). The idea here is that non- 
>> visual
>> users have a strong need for a caption even though captions can  
>> often be
>> omitted due to context for visual users.
>
> WCAG 2.0 does not recommend using summary="" for layout tables. There
> is a note in one of the WCAG 2.0 techniques that states that a null
> summary is acceptable on layout tables [1], but it is not a WCAG 2.0
> recommendation.

I'm sorry. I didn't mean to say recommend (in the RFC 2119 sense). I  
more meant to say in keeping with WCAG advice. I should be more  
precise even in my email messages.

> HTML 5 should not encourage the use of layout tables by describing how
> to create layout tables in the specification.

I think we should discourage the use of layout tables, but that we  
shouldn't avoid guiding authors who use them (unless we're prepared to  
prohibit their use and say "authors must not use layout tables").

> The purpose of the caption element and the summary attribute are
> completely different (caption titles the table, summary describes how
> the table should be read by people who cannot see how the table has
> been rendered), and they shouldn't be confused. Version A doesn't
> encourage layout tables, and it doesn't confuse the purpose of the
> summary attribute. Most of the debate around the summary attribute has
> been about misunderstanding its purpose, so trying to merge its
> purpose with another element's purpose is probably not the best way of
> getting the summary attribute into the HTML5 specification.

I certainly don't want to contribute to the confusion between caption  
and summary. Version B tries to reduce the confusion by being explicit  
about how to craft summary and captions (and with Leif's section also  
a more complete specification for crafting a caption). My view is that  
we only contribute to the confusion by not addressing these issues. I  
would certainly be happy to see HTML5 say "authors should not use  
tables for layout", but I don't think we have reached a level of  
specification nor support to be able to say "authors must not use  
tables for layout" which is why WCAG addresses the situation. My  
feeling is that version A is too ambiguous and will therefore  
contribute to the confusion we're all trying to avoid. CSS  as  
specified would have probably eliminated the need for layout tables  
entirely. CSS as implemented makes some layouts simpler with a single  
layout table (though that too can be avoided, but not as easily as the  
way CSS was specified).

Also since we're trying to craft this in keeping with RFC 2119, its  
important to draft the prose according to the harm done. If using a  
single table for layout causes harm for AT users and other users,  
could you say something more about that? To me the biggest harm from  
layout tables is the intense nesting of tables and the combined use of  
non-breaking elements and spacer gifs which created an unmaintainable  
tangle of markup and impeded AT UAs. We might say something elsewhere  
to discourage such layout table use or even all layout tables, but I  
don't think we have a situation where we should prohibit the use of a  
single layout table.

Take care,
Rob
Received on Sunday, 1 March 2009 01:28:53 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:32 GMT