Re: model theory for RDF/S

Some remarks on chapter on the RDFS entailment in http://www.w3.org/TR/rdf-mt/ :

During the development of Sesame [1] we also have been struggling with the RDF
Schema interpretation. I think writing a RDF model theory document is a really
good move to clear things up, especially concerning RDF Schema.

In chapter 5 on "RDFS entailment" some rules are given on how to derive the
schema-closure of RDFS graphs. These rules are then applied to an example RDF
graph consisting of a single statement:

foo bar baz .

According to the document, the schema-closure of this graph is:

   foo bar baz .
   foo rdf:type rdfs:Resource .
   baz rdf:type rdfs:Resource .
   bar rdf:type rdf:Property .
   rdf:type rdf:type rdf:Property .
   rdf:Property rdf:type rdfs:Class .
   rdf:Property rdfs:subClassOf rdfs:Resource .

However, I think a number of statements are missing from the closure. First of
all, the three triples from step 1 are missing:

   rdfs:Resource rdf:type rdfs:Class .
   rdfs:domain rdf:type rdfs:ConstraintResource .
   rdfs:range rdf:type rdfs:ConstraintResource .

(BTW: there's a typo in the document; it mentions rdf:domain and rdf:range
instead of rdfs:domain and rdfs:range.)

Second, strictly applying the rules from step 2 will also yield the following
statements:

   bar rdf:type rdfs:Resource .
   
   rdf:type rdf:type rdfs:Resource .
   rdfs:Property rdf:type rdfs:Resource .
   rdfs:Class rdf:type rdfs:Class .
   rdfs:Class rdf:type rdfs:Resource .
   rdfs:Class rdfs:subClassOf rdfs:Resource .
   rdfs:subClassOf rdf:type rdfs:Property .
   rdfs:subClassOf rdf:type rdfs:Resource .
   rdfs:Resource rdf:type rdfs:Resource .
   rdfs:Resource rdfs:subClassOf rdfs:Resource .

And this doesn't even include the expansion of the three triples from step 1
yet. The first statement is specific for this RDF graph, the rest will be
generated in every schema closure.

I'm certainly no expert on model theory, so forgive me if I'm completely
missing the point, but aren't the last two statements troublesome? These
triples clearly state that rdfs:Resource is both a member and a subset of
itself.

Regards,

Arjohn Kampman


[1] http://sesame.aidministrator.nl/


--
arjohn.kampman@aidministrator.nl
aidministrator nederland b.v.  -  http://www.aidministrator.nl/
prinses julianaplein 14-b, 3817 cs amersfoort, the netherlands
tel. +31-(0)33-4659987   fax. +31-(0)33-4659987

Received on Thursday, 11 October 2001 05:56:12 UTC