W3C home > Mailing lists > Public > www-webont-wg@w3.org > September 2002

Re: semantics layering document revision.

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Tue, 24 Sep 2002 10:16:13 -0400 (EDT)
Message-Id: <20020924.101613.39333692.pfps@research.bell-labs.com>
To: phayes@ai.uwf.edu
Cc: www-webont-wg@w3.org

From: pat hayes <phayes@ai.uwf.edu>
Subject: semantics layering document revision.
Date: Mon, 23 Sep 2002 20:09:16 -0500

> 
> The document has been revised in the light of Peter's comments and 
> can be found at
> 
> http://www.coginst.uwf.edu/users/phayes/RDFS2OWL-L.html
> 
> I have highlighted the important changes in red. They are mostly 
> technical tweaks to the basic semantic conditions on weak OWL in 
> order to support the required entailments. I believe that this now 
> handles all the points raised by Peter in his last message. 

Well there are still issues outstanding, including whether there are any
Large OWL interpretations at all. 

> There may 
> still be some slight glitches in the green tables, 

Here is my view of the status of Pat's claimed true triples in Weak OWL.
(Some of these are not correct with respect to the abstract syntax, but I'm
not concerned with that right now.  The incorrect ones include the domain
or range for owl:inverseOf, owl:oneOf, owl:allValuesFrom, and
owl:someValuesFrom.)

owl:Thing rdf:type owl:Class . 		No.  There is no reason in Weak OWL for owl:Thing to be an OWL class.
owl:Nothing rdf:type owl:Class .	No.  There is no reason in Weak OWL for owl:Nothing to be an OWL class.
owl:Nothing owl:oneOf rdf:nil .		No.  As owl:Nothing need not be an OWL class, it need not be an owl:oneOf even though it has the right extension.

owl:Class rdf:subClassOf rdfs:Class .			Only under an iff definition of rdfs:subClassOf
owl:Restriction rdf:subClassOf owl:Class .		Only under an iff definition of rdfs:subClassOf
owl:Property rdf:subClassOf rdf:Property .		Only under an iff definition of rdfs:subClassOf
owl:DataTypeProperty rdf:subClassOf owl:Property .	Only under an iff definition of rdfs:subClassOf
owl:ObjectProperty rdf:subClassOf owl:Property .	Only under an iff definition of rdfs:subClassOf
owl:FunctionalProperty rdf:subClassOf owl:Property .	Only under an iff definition of rdfs:subClassOf
owl:SymmetricProperty rdf:subClassOf owl:Property .	Only under an iff definition of rdfs:subClassOf
owl:InverseFunctionalProperty rdf:subClassOf owl:Property .	Only under an iff definition of rdfs:subClassOf
owl:TransitiveProperty rdf:subClassOf owl:Property .	Only under an iff definition of rdfs:subClassOf
        
owl:sameClassAs rdfs:domain owl:Class .			Only under an iff definition of rdfs:domain and rdfs:range.
owl:sameClassAs rdfs:range owl:Class .			Only under an iff definition of rdfs:domain and rdfs:range.
owl:samePropertyAs rdfs:domain owl:Property .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:samePropertyAs rdfs:range owl:Property .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:DisjointWith rdfs:domain owl:Class .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:DisjointWith rdfs:range owl:Class .			Only under an iff definition of rdfs:domain and rdfs:range.
owl:complementOf rdfs:domain owl:Class .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:complementOf rdfs:range owl:Class .			Only under an iff definition of rdfs:domain and rdfs:range.
owl:inverseOf rdfs:domain owl:Property .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:inverseOf rdfs:range owl:Property .			Only under an iff definition of rdfs:domain and rdfs:range.
owl:oneOf rdfs:domain owl:Class .			Only under an iff definition of rdfs:domain and rdfs:range.
owl:oneOf rdfs:range rdf:List .				Only under an iff definition of rdfs:domain and rdfs:range.
owl:unionOf rdfs:domain owl:Class .			Only under an iff definition of rdfs:domain and rdfs:range.
owl:unionOf rdfs:range rdf:List .			Only under an iff definition of rdfs:domain and rdfs:range.
owl:intersectionOf rdfs:domain owl:Class .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:intersectionOf rdfs:range rdf:List .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:onProperty rdfs:domain owl:Restriction .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:onProperty rdfs:range owl:Property .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:allValuesFrom rdfs:domain owl:Restriction .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:allValuesFrom rdfs:range rdfs:Class .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:someValuesFrom rdfs:domain owl:Restriction .	Only under an iff definition of rdfs:domain and rdfs:range.
owl:someValuesFrom rdfs:range rdfs:Class .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:hasValue rdfs:domain owl:Restriction .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:minCardinality rdfs:domain owl:Restriction .	Only under an iff definition of rdfs:domain and rdfs:range.
owl:minCardinality rdfs:range xsd:NonNegativeInteger .	Only under an iff definition of rdfs:domain and rdfs:range.
owl:maxCardinality rdfs:domain owl:Restriction .	Only under an iff definition of rdfs:domain and rdfs:range.
owl:maxCardinality rdfs:range xsd:NonNegativeInteger .	Only under an iff definition of rdfs:domain and rdfs:range.
owl:cardinality rdfs:domain owl:Restriction .		Only under an iff definition of rdfs:domain and rdfs:range.
owl:cardinality rdfs:range xsd:NonNegativeInteger .	Only under an iff definition of rdfs:domain and rdfs:range.

Similarly, the entailment rule

AAA rdf:type owl:Property .  ->  AAA rdfs:domain owl:Thing .

is only valid under an iff definition of rdfs:domain.

The entailment rule

AAA rdf:type owl:Class . AAA rdfs:subClassOf BBB . BBB rdfs:subClassOf AAA .
-> AAA owl:sameClassAs BBB .

is not valid because BBB might not be an OWL class, and similarly for the
entailment rule for subPropertyOf.

> and there are a 
> few places where I just don't know what the right answer is (eg must 
> the range of an OWL property be an OWL class? Are all datatypes OWL 
> classes?) but Im sure it can be tweaked to fit whatever people want.

These can probably be tweaked.  A set of answers can be found in the OWL
Abstract Syntax document.

peter
Received on Tuesday, 24 September 2002 10:16:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:52 GMT