Re: Comments on latest EARL Schema draft

Michael A Squillace schrieb:

> 2. regarding the Assertor class: First, the last example markup looks 
> suspect:
> 
> <foaf:Agent rdf:about="#assertor">
>   <dc:title xml:lang="en">Bob using Cool Tool</dc:title>
>   <dc:description xml:lang="en">Bob doing semi-automated 
> testing</dc:description>
>   <earl:mainAssertor rdf:resource="http://www.example.org/people/#bob"/>
>   <foaf:member rdf:resource="http://www.example.org/tool/#cool"/>
> </foaf:Agent>
> 
> I think we want to define a foaf:group, not foaf:agent here.

ACK.

> This makes 
> more sense intuitively and is also consistent with the definition of the 
> earl:mainAssertor property, a subproperty of foaf:member which can only 
> have a subject of foaf:group.

Slightly different: A resource acting as subject in a triple with a 
foaf:member property (or subproperty of foaf:member) by RDFS _has_ the 
type foaf:Group. Specifying a domain does not specify a constraint. It 
adds a type.

> Related to this point, the "Related Classes" 
> section states that, "Rather than using the generic earl:Assertor class 
> directly, it is recommended that one of the following refinements be 
> employed," and goes on to list foaf:person, foaf:agent, foaf:organization, 
> foaf:group, and earl:software as the possible "refinements." Actually, we 
> want to use some combination of these and wrap them in a foaf:group, which 
> represents the earl:Assertor; these are not subclasses in any way of the 
> earl:Assertor class. (Part of the problem is with the use of the word 
> 'refinement', which in the description of earl:mainAssertor seems to imply 
> subProperty of, whereas here we do not want to indicate any inheritence 
> relation.) Some of this might be clarified by requirement #4 in the 
> "Conforming Reports" section, but I think there is still some confusion 
> between what counts as an assertor in general and what is allowed by the 
> spec.

Taking into account the RDFS information, the resource <#assertor> has 
the following types:

* foaf:Agent (by using the foaf:Agent element type in RDF/XML)
* earl:Assertor (by domain of earl:mainAssertor)
* foaf:Group (by domain of foaf:member, inherited domain of 
earl:mainAssertor)

I think the wording "Rather than using the generic earl:Assertor class
directly" sounds too XML-like (XML element types). We want to define an 
RDF-based vocabulary. So we should think in RDF, not in XML.

What about the following?

"Rather than specifying only an earl:Assertor type, it is recommended 
that one of the following types be added:"

-- 
Johannes Koch
Fraunhofer Institute for Applied Information Technology FIT
Web Compliance Center
Schloss Birlinghoven, D-53757 Sankt Augustin, Germany
Phone: +49-2241-142628    Fax: +49-2241-142065

Received on Tuesday, 23 June 2009 08:49:15 UTC