Re: incompleteness of rdf-closure

From: pat hayes <phayes@ihmc.us>
Subject: Re: incompleteness of rdf-closure
Date: Mon, 9 Jun 2003 11:33:07 -0500

> >The RDF entailment lemma is still not valid in the 6 June 2003 version of
> >RDF semantics.
> 
> That might well be the case: I have not yet checked the proof in 
> detail after the many editorial changes.
> 
> >
> >For example, the empty RDF graph rdf-entails
> >	rdf:subject rdf:type rdf:Property .
> >but this is not part of the rdf-closure of the empty RDF graph.
> 
> It is:
> 
> rdf:subject rdfs:range rdfs:Resource . (rdfs axiom)
> rdfs:range rdfs:domain rdf:Property .  (rdfs axiom)
> rdf:subject rdf:type rdf:Property .  (rdfs2)

I don't understand how two rdfs axioms and an rdfs rule can be used in the
determination of a rdf-closure.

> The rdf:types of all the rest of the RDFS class and property 
> vocabulary can be derived similarly from the domains and ranges of 
> domain and range plus the exhaustive listing of the domains and 
> ranges of the vocabulary in the RDFS axiomatic triple table.
> 
> Clearly, however, the text should provide details of derivations of 
> this kind (and many others) rather than merely hinting at them.
> 
> Pat

peter