W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > July to September 2004

Is this FOO ok Re: HTML Table Markup

From: Charles McCathieNevile <charles@sidar.org>
Date: Thu, 16 Sep 2004 12:13:37 +0300
To: "Phill Jenkins" <pjenkins@us.ibm.com>, tina@greytower.net
Cc: w3c-wai-ig@w3.org
Message-ID: <opseez0zclw5l938@widsith.local>

Hi folks,

the basic problem seems to be that HTML and XHTML 1.x don't support enough  
information to describe whether they have been checked for accessibility.  
This shouldn't come as a great shock. After all, they are based on  
something designed 15 years ago to encode simple hypertext.

The people who originally designed HTML are mostly working on other tings  
now - in TimBL's case, the semantic web.

The people who are trying to drag HTML into the 21st century are for the  
most part working on XHTML 2.

It is possible to continue to add attributes, or try and add further  
menaing to the contents of overloaded attributes in HTML 4, but it seems  
like a mistake. An alternative approach is the use of metadata -  
information really genuinely meant to let tools know if something has been  
tested and is deemed OK... (among other purposes).

The Evaluation and Repair Tools group, when it is active, works on a way  
of doing just this, by using EARL. It's relatively straightforward to say  
that some table in a page is used for layout, and the process goes as  
follows:

Identify the table. You can do this by giving it an id attribute and using  
its URI, or you can use Xpointer to identify the Nth table in a page (or  
the Xth table within the Yth table, or whatever)

For the sake of example, there is a table at http://www.microsoft.com/  
which I calim is a layout table. It has the id attribute tMain. So I can  
talk about http://www.microsoft.com/#tMain

Define an appropriate test, and give it a URI. For this esxample, I have  
defined a rule - "a Given table is a data table not a layout table".  
Passing the rule means you have a data table, failing it means you have a  
layout table. The alternative is that you don't know or can't tell. This  
rule is described at  
http://www.w3.org/2001/sw/Europe/200407/earl/tests#DnotL (it actually  
exists and you can use it. There are a number of these that Chris Ridpath  
has described, too).

Then you record the data. Following  
http://www.w3.org/2001/sw/Europe/talks/0312-earl (a step by step guide to  
writing EARL), I make the following claim:

<?xml version="1.0" encoding="utf-8"?>
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
          xmlns:earl="http://www.w3.org/WAI/ER/EARL/nmg-strawman#">
   <earl:Assertion>

     <earl:subject>
       <rdfs:comment>I have said this is the Web Content that I got at a  
particular time, instead of just using the URI directly as the subject a la
          'earl:subject rdf:resource="http://www.microsoft.com#tMain"  
because I want to be able to check if this claim is valid, and the page  
may change
       </rdfs:comment
       <earl:WebContent>
         <earl:reprOf rdf:resource="http://www.microsoft.com#tMain" />
         <earl:date>2004-09-16T11:51:36+0200</earl:date>
       </earl:WebContent>
     </earl:subject>

     <earl:result  
rdf:type"http://www.w3.org/WAI/ER/EARL/nmg-strawman#Pass"/>
     <earl:testcase  
rdf:resource="http://www.w3.org/2001/sw/Europe/200407/earl/tests#DnotL"/>

     <earl:assertedBy>
       <foaf:Person>
         <foaf:name>Charles McCathieNevile</foaf:name>
         <foaf:mbox rdf:resource="mailto:charles@w3.org" />
       </foaf:Person>
     </earl:assertedBy>
     <earl:mode  
rdf:resource="http://www.w3.org/WAI/ER/EARL/nmg-strawman#manual"/>

   </earl:Assertion>
</rdf:RDF>

This is a methodology that is already used in practice - UBAccess record  
information this way to generate transformations of a resource, as does  
IBM's new service. It doesn't even rely on me being able to edit the  
table, so I can audit a site for someone else and record results. (Hera  
was developed to do exactly this, along with a n umber of other tools).

cheers

Chaals

On Mon, 13 Sep 2004 13:14:06 -0500, Phill Jenkins <pjenkins@us.ibm.com>  
wrote:

>> Let your tool flag any table without a th as a potential misuse of  
>> tables,
>> and move on.
>
> Tina, that's my proposal - to examine each table and determine if it  
> needs TH elements.  If the TH is needed it is added, if it is not (no  
> "potential misuse"), then the null summary is added as a flag to tell  
> that the
> examination has been preformed.  Again, my proposal is to flag those not
> needing TH elements with null summary attribute.  The only way I know to
> move on is to leave some change in the markup to indicate that TH is not
> needed versus it hasn't been examined yet.
>
> Regards,
> Phill Jenkins



-- 
Charles McCathieNevile         charles@sidar.org
FundaciĆ³n Sidar             http://www.sidar.org
Received on Thursday, 16 September 2004 10:14:24 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 23:39:44 UTC