Re: model theory for RDF/S

On Wed, 2001-09-26 at 16:21, Pat Hayes wrote:
> The RDF core working group has now published a model theory for (most 
> of) RDF/S, available at http://www.w3.org/TR/rdf-mt/. Comments are 
> welcome, as usual.

Very interesting! A few comments:

Section 3, §1:
"...every member of S also satisfies E'"
I guess this is E, not E'...

Section 3.1, §3:
"A tidy ground instance with respect to any set which is disjoint from
vocab(E)..."
I may be wrong but this sound redundant: how could possibly be tidy an
instance with respect to a set including URIs from vocab(E) ?

Section 3.1, Skolemization Lemma:
typo: sk() becomes SK().

Section 5:
in rules 3b and 4, bbb should be uuu
(or the produced triples may have a literal as subject or predicate)
I also think that in the example, some triples are missing from the
given closure (though these are triples which will be generated for
every schema closure):
  rdfs:Resource rdf:type rdfs:Resource
  rdfs:Resource rdf:type rdfs:Class
  rdf:Property rdf:type rdfs:Resource
  rdf:Property rdf:type rdfs:Class
  rdfs:Class rdf:type rdfs:Resource
  rdfs:Class rdf:type rdfs:Class
I guess that's all... :)

By the way, wouldn't that be as easy as generating all those triples to
automatically include the Schema for RDFSchema, which makes much of them
explicit directly ?

> This doesn't cover reification, by the way. Maybe in a future 
> version, but don't hold your breath.

I won't ;-)

  Pierre-Antoine

Received on Wednesday, 26 September 2001 11:10:51 UTC