Re: missing (and incorrect) RDFS axioms

At 02:33 PM 11/9/02 -0600, pat hayes wrote:

>>which would then also carry
>>
>>   rdf:Alt rdfs:subClassOf rdfs:Container .
>>   rdf:Bag rdfs:subClassOf rdfs:Container .
>>   rdf:Seq rdfs:subClassOf rdfs:Container .
>>   rdfs:member rdfs:domain rdfs:Container .
>>
>>
>>rdfs:isDefinedBy rdfs:subPropertyOf rdfs:seeAlso .
>
>I will put this in, but it strikes me as a damn silly thing to say, since 
>it is like saying 'burble subproperty of babble'. Neither of these things 
>actually are properties of anything, in fact, so how can one be a 
>subproperty of the other? Ah well, never mind.

It may not mean much formally, but I think there are situations where it 
can be useful.

For example, when generating graphical views of RDF documents, Dan 
Connolly's work [1] uses cwm rules to generate new graph-property 
assertions from the RDF statements in a graph, and in so doing defines the 
graphical view to be constructed.  If one interprets the link "seeAlso" as 
meaning that more information about the subject can be found at the object 
resource, then it seems reasonable that definedBy also satisfies that 
description.  Thus, by defining rules that seeAlso links are represented in 
one's graphical presentation, the definedBy links get included when 
evaluating the graph under RDFS closure rules.

#g

[1] Circles and arrows diagrams using stylesheet rules. Dan Connolly's page 
on generating node and arc diagrams from RDF: http://www.w3.org/2001/02pd/.



-------------------
Graham Klyne
<GK@NineByNine.org>

Received on Sunday, 10 November 2002 06:47:31 UTC