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

Re: summary="" in HTML5 ISSUE-32

From: Robert J Burns <rob@robburns.com>
Date: Fri, 27 Feb 2009 12:32:48 -0500
Cc: "Ian Hickson" <ian@hixie.ch>, "Leif Halvard Silli" <lhs@malform.no>, HTMLWG <public-html@w3.org>
Message-Id: <9244DFA9-B071-40FB-A5C5-07690B614F6F@robburns.com>
To: "Simon Pieters" <simonp@opera.com>
Hi Simon,

On Feb 27, 2009, at 1:50 AM, Simon Pieters wrote:

> On Fri, 27 Feb 2009 02:40:16 +0100, Ian Hickson <ian@hixie.ch> wrote:
>
>>> > Unfortunately summary="" can't be made visible in Web browsers,  
>>> due to
>>> > the wide mis-use of the attribute.
>>>
>>> As with @title, it can.
>>
>> If Web browsers are willing to expose it, then that would definitely
>> change matters. Are they?
>>
>> Input from browser vendors here would be helpful.
>
> I've exposed summary='' in my user style sheet and then looked at  
> about 100 pages from this list
>
> http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/pages/tagattr/table/summary
>
> as well as pages from most sites in this list
>
> http://philip.html5.org/data/table-summary-gov.txt
>
> and concluded that, due to things like summary="For layout  
> purposes", we cannot expose summary everywhere. We can't just expose  
> it in standards mode, either, because there such pages using  
> standards mode. We *could* expose it if we had a reliable algorithm  
> to detect layout tables. Since such an algorithm would affect  
> rendering in browsers, it would have to be given in the spec.

I think the proposal to expose summary is forgetting entirely about  
the author and the user. The exposure of the summary attribute in the  
normal flow of the document is not something that should be imposed by  
the UA. Leif's proposal is geared at authoring practices as far as I  
understand it. So if a browser had a special authoring mode, then  
summaries could be exposed.

The 'summary' attribute value could also be exposed outside the flow  
of the document: in the chrome. This could occur even in those cases  
where improper labeling of layout tables occurred and other errant  
uses of summary. By exposing summary in the chrome, UAs help address  
the problems Leif sought to address, since authors can become aware of  
the summary and easily recognize when the value is incorrect, missing,  
or out of sync with the tables contents.


> While doing the above excersise, I noted that *most* pages were  
> using summary='' to say "for layout" and from the first list I  
> didn't find a single data table with summary, let alone any data  
> table with a useful summary. From the list of gov pages I only found  
> 2 or 3 tables with useful summary -- the rest either repeated a  
> heading or were "for layout".

It would be helpful if you could list some of the URLs you  
encountered. I took a stab at 7 different documents from that sample  
and found one proper use for a data table and 6 proper use of the  
'summary' attribute for layout tables[1].

>
>
> Now, I can't but conclude that summary='' values are useful so  
> seldom that it's not worth the cost and risk to break pages that use  
> it wrongly.

Indeed we cannot. We don't even have a research model in place that  
would allow us to make such conclusions.

> If there's a need to distinguish between caption and summary, it  
> seems better to use
>
>  <caption> Table 1. <summary> Foo bar baz. </summary> </caption>
>
> But it's not clear to me why
>
>  <caption> Table 1. Foo bar baz. </caption>
>
> is not good enough.

The reason it is not good enough is the same reason that summary  
should not be displayed by default in the normal flow of the document  
(as opposed to in the chrome or in the flow of the document in a  
special authoring mode). The 'summary' attribute provides a place for  
authors to include content directed at cognitively and visually  
disabled users to assist them in consuming the table. It should be  
left to the author and user whether they want that exposed in the flow  
of the document. Usually it will not be the case that authors and  
users want such information exposed in the flow of the document. This  
means that the two things  caption and summary  need two separate  
mechanisms or authors and users will not have that level of control  
over presentation.

In conclusion, we do not want the 'summary' attribute value exposed by  
UAs by default in the flow of a document. We need to leave that up to  
authors and users. However, authoring modes and in the chrome exposure  
of the 'summary' attribute helps ensure proper use of the attribute.

Take care,
Rob

[1]: <http://lists.w3.org/Archives/Public/public-html/2009Feb/0740.html>
Received on Friday, 27 February 2009 17:33:47 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:01 UTC