W3C home > Mailing lists > Public > public-wai-ert@w3.org > March 2005

Re: EARL and describing tests Re: Agenda for F2F at TP

From: Charles McCathieNevile <charles@sidar.org>
Date: Tue, 01 Mar 2005 11:56:03 +0100
To: "Chris Ridpath" <chris.ridpath@utoronto.ca>, shadi@w3.org
Cc: public-wai-ert@w3.org
Message-ID: <opsmygnpvzw5l938@saturne.cust.hotspot.t-mobile.com>

On Tue, 1 Mar 2005 11:01:32 -0500, Chris Ridpath  
<chris.ridpath@utoronto.ca> wrote:
...
> Charles said:
>> But we do have same basic ability to describe tests. We could use Dublin
>> Core (for example) to give more detailed information about who is the
>> author of a particular test, when it was developed, etc. To make this
>> easy, all we do is keep using RDF.
>>
> Yes, but it may be easier just to reference a stable test.

Well, this is where it gets interesting.

The WCAG tests are described already. In english. Sidar uses Hera  
primarily in spanish, and then in Portuguese, French, then english. So we  
want the explanations available in those languages. The test itself, and  
code that runs it, might not change. But descriptions of it for building  
user interfaces, or for messages that automatic tools put into reports, do  
need to change according to language.

Rather than assume that the developers of any given test I want to use  
will add my translations, part of the value of RDF is that I can provide  
the human-readable information in another language, and build on that in  
my interface, without needing this kind of coordination.

>> Nor does it describe the relationship between a group of tests, which I
>> think should be done using OWL restrictions rather than inventing
>> EARL-specific stuff that does the same thing.
>>
> We can now reference an accessibility guideline that contains all the  
> tests that are required for conformance to that guideline. Example:
> http://www.w3.org/WAI/GL/WCAG20/tests/wcag-2-0-aaa.xml
>
> This is just a simple XML file but perhaps we could improve it using  
> another technology as you suggested.

Yeah, I think that is important. If we use the XML there the first problem  
is that there is no apparent specification, so I have to make up an  
implementation by guessing from the code. More to the point, this relies  
on implementing a specific XML vocabulary in any tool.

The idea of using OWL is that you can work with generic tools like Jena to  
build the tools, and that it saves us from the work of discussing how we  
should design an XML or RDF vocabulary in the first place :-)

> Here's how our checker says you have an image without an alt attribute  
> in EARL:
>
> - <earl:Assertion>
>   <dc:date>2005-3-1T10:26:16-5</dc:date>
>   <earl:Subject element="img" line="9" rdf:resource="#subject"  
> xpath="/HTML/BODY/P/IMG" />
> - <earl:Testcase  
> rdf:resource="http://www.w3.org/WAI/GL/WCAG20/tests/test1.html">
>   <earl:testId rdf:resource="1" />
>   <earl:message>IMG missing ALT attribute.</earl:message>
>   </earl:Testcase>
> - <earl:result  
> rdf:resource="http://www.w3.org/WAI/ER/EARL/nmg-strawman#Fail">
>   <earl:confidence  
> rdf:resource="http://www.w3.org/WAI/ER/EARL/nmg-strawman#high" />
>   </earl:result>
>   <earl:Assertor rdf:resource="#assertor" />
>   <earl:mode  
> rdf:resource="http://www.w3.org/WAI/ER/EARL/nmg-strawman#automatic" />
>   </earl:Assertion>

I think you should check your stuff with the RDF validator, and make sure  
that the code is correct for general tools. (A major problem we have is  
that the schema we have is not valid XML, RDF or anything else, but I  
think it is obvious what it should be :-( I'll send an email with the  
version that we rely on in Sidar, which just corrects things that are  
obvious typos leading to validity errors.)

Clearly, from your example, you have a use case for picking a detailed  
part of a resource as a subject - which is I think a common request in  
EARL for people who are interested in actually doing repair work.

 From Sidar's perspective, most of our reports are in a spanish language,  
so we would be likely to publish some minimal further description of the  
test cases to provides labels (such as the earl:message in your example)  
in different langauges. The reason we like RDF is that this is easy for us  
to do, and makes it easy for us to share our translations with others....

Anyway, I too am looking forward to making progress, and seeing what  
others are doing :-)

Cheers

Chaals


-- 
Charles McCathieNevile - Vice Presidente - Fundacion Sidar
charles@sidar.org                      http://www.sidar.org
     (chaals is available for consulting at the moment)
Received on Tuesday, 1 March 2005 17:05:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:18:25 GMT