multiple tools contributing to a result

Hi folks,

(here's an insight into Hera development :-) We're looking now at how to  
code up results when multiple tools have been involved. The example case  
here is the checkpoint about valid markup - Hera uses the W3C validators  
to check it.

So a simple approach would be to note that the validators gave the result:

    <earl:Assertion rdf:about="#wcag1cp3.2">
     <earl:subject rdf:resource="#subject" />
     <earl:message rdf:parseType="Literal">
      <p xmlns="http://www.w3.org/1999/xhtml" xml:lang="es">La página es
<strong>XHTML 1.0 Transitional</strong> válido.<br />
      El código de las hojas de estilo es correcto.</p>
     </earl:message>
     <earl:result
rdf:type="http://www.w3.org/WAI/ER/EARL/nmg-strawman#pass"/>
     <earl:mode
rdf:resource="http://www.w3.org/WAI/ER/EARL/nmg-strawman#automatic"/>
     <earl:testcase
rdf:resource="http://www.w3.org/TR/WCAG10/#tech-identify-grammar"/>
     <earl:assertedBy rdf:parsetype="Collection">
       <earl:Tool rdf:about="http://validator.w3.org/">
         <dc:title xml:lang="en">W3C Markup Validator</dc:title>
         <dc:location
rdf:datatype="http://www.w3.org/2001/XMLSchema#URI">http://validator.w3.org</dc:location>
       </earl:Tool>
       <earl:Tool rdf:about="http://jigsaw.w3.org/CSSvalidator">
         <dc:title xml:lang="en">W3C CSS Validator</dc:title>
         <dc:location
rdf:datatype="http://www.w3.org/2001/XMLSchema#URI">http://jigsaw.w3.org/CSSValidator</dc:location>
       </earl:Tool>
     </earl:assertedBy>
    </earl:Assertion>

An aternative is to say that Hera (the tool) is itself using some tools.  
We already say that when a test result is manual it is asserted by a  
Person using Hera, and when it is automatic it is asserted directly by  
Hera.

    <earl:Assertor rdf:about="#assertor">
       <rdf:type rdf:resource="http://xmlns.com/foaf/0.1/Person"/>
          <foaf:name>charles</foaf:name>
          <sidar:usando>
             <earl:Tool rdf:about="http://www.sidar.org/hera/">
                <dc:title>Hera</dc:title>
                <dc:location>http://www.sidar.org/hera</dc:location>
                <sidar:usando>
                   <earl:Tool rdf:about="http://validator.w3.org/">
                      <dc:title xml:lang="en">W3C Markup  
Validator</dc:title>
                      <dc:location
rdf:datatype="http://www.w3.org/2001/XMLSchema#URI"
                       >http://validator.w3.org</dc:location>
                   </earl:Tool>
                </sidar:usando>
                <sidar:usando>
                   <earl:Tool rdf:about="http://jigsaw.w3.org/CSSvalidator">
                      <dc:title xml:lang="en">W3C CSS Validator</dc:title>
                      <dc:location
rdf:datatype="http://www.w3.org/2001/XMLSchema#URI"
                       >http://jigsaw.w3.org/CSSValidator</dc:location>
                   </earl:Tool>
                </sidar:usando>
             </earl:Tool>
          </sidar:usando>
    </earl:Assertor>

So in this case all results asserted by Hera are the version of Hera using  
the validators to help it. This seems a bit inexact, because for 64 of the  
65 checkpoints tested Hera doesn't consult the validators at all (we  
haven't yet automated the test for whether W3C markup is being used) but  
where the overall tool is a framework that might have various sub-tools  
working or not (e.g. the options on the W3C markup validator to show an  
outline or ignore encoding errors) this might make good sense. Anyway, I  
don't think it is wrong per se, although this case shows where it might  
start to seem odd...

Finally, we could write an OWL restriction that it meets checkpoint 3.2 if  
it meets a validity test for each technology it uses, and define a number  
of validity tests for different technologies.

Then we do a more complete analysis, and determine (or claim we determined  
:-) the page uses XHTML and CSS and PNG, and validate each of those using  
the relevant test.

Then we claim the thing meets 3.2 in mode "heuristic", showing as evidence  
the OWL rule, the claim that the page uses those technologies, and the  
results for each validity test.

-- 
Charles McCathieNevile                      Fundacion Sidar
charles@sidar.org   +61 409 134 136    http://www.sidar.org

Received on Wednesday, 13 April 2005 02:51:49 UTC