proposed addition to today's teleconference (was Re: Agenda/Logistics telecon June 4, 2003)

I have a list of the changes I have made to OWL since last call, including
a proposed change that I have described in messages sent out to the WG.  I
suggest that we spend a bit of time going over these changes.

peter


		Summary of post last-call changes to OWL (so far)

CHANGES prompted by Working Group decisions:

Change: Add owl:Nothing to OWL Lite
Approved: http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0017.html

Change: Added OntologyProperty axioms.
Approved: http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0402.html

Change: The semantic constraints on owl:intersectionOf and related
	properties changed from only-if to iff.
Approved: http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0402.html


CHANGES prompted by errors in the semantics:

Problem: Not all of the OWL vocabulary was required to be classes or
	 properties in the RDF semantics.
Noticed: http://lists.w3.org/Archives/Public/www-webont-wg/2003Apr/0046.html
Observable consequences: The empty ontology does not OWL-Full entail
	owl:Ontology rdf:type rdfs:Class . and other similar non-entailments.
Change: Require that owl:Class, owl:Restriction, owl:ObjectProperty,
	owl:DatatypeProperty, owl:AnnotationProperty,
	owl:OntologyProperty, owl:Ontology, owl:AllDifferent,
	owl:FunctionalProperty, owl:InverseFunctionalProperty,
	owl:SymmetricProperty, owl:TransitiveProperty be in CI and that
	owl:distinctMembers be in RI.

Problem: owl:OntologyProperty is not in the disallowed vocabulary
Noticed: http://lists.w3.org/Archives/Public/www-webont-wg/2003Apr/0064.html
Observable consequences: owl:OntologyProperty rdf:type owl:Class .
	is a valid OWL DL ontology.
Change: Add owl:OntologyProperty to the disallowed vocabulary.

Problem: The inverse mapping for EquivalentClasses requires the recognition
	of Hamiltonian paths, which is too difficult.
Noticed: http://lists.w3.org/Archives/Public/www-webont-wg/2003Apr/0003.html
	 http://lists.w3.org/Archives/Public/public-webont-comments/2003May/0052.html
Change: Changed mapping for EquivalentClasses from a path to a connected
	graph. 
Approved: http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0402.html

Problem: The mapping for Individual() results in no triples.
Noticed: http://lists.w3.org/Archives/Public/www-webont-wg/2003Jun/0004.html
Change: Require that anonymous Individual axioms result in a typing triple
	to owl:Thing if no other type is present.

Problem: Using rdfs:Literal results in rdfs:Literal rdf:type rdfs:Datatype
	 triple.  The status of rdf:XMLLiteral is unclear.
Noticed: http://lists.w3.org/Archives/Public/www-webont-wg/2003Jun/0011.html
Change: Change the abstract syntax so that data ranges include rdfs:Literal
	 as a special case and the mapping rules so that the offensive
	 triple is not generated.  Change wording to clarify that
	 rdf:XMLLiteral is a built-in OWL datatype.

Problem: Annotations of anonymous ontologies do not need to be on the same
	domain element in the direct semantics.
Noticed: http://lists.w3.org/Archives/Public/public-webont-comments/2003May/0050.html
Observable consequences: Ontology(Annotation(owl:versionInfo "1")) 
			 Ontology(Annotation(owl:versionInfo "2"))
	entails Ontology(Annotation(owl:versionInfo "1")
			 Annotation(owl:versionInfo "2"))
Change: Change the direct semantics for ontology annotations in an ontology
	to use a common domain element.

Problem: Annotations on non-individuals interact with individual in the
	 direct semantics. 
Noticed: http://lists.w3.org/Archives/Public/www-webont-wg/2003Jun/0004.html
Observable consequences: 
	AnnotationProperty(ex:a)
	ObjectProperty(ex:p Symmetric)
	SubClassOf(owl:Thing restriction(ex:p value(ex:x)))
	Individual(ex:x restriction(ex:p cardinality(1)))
	Class(ex:c1 annotation(ex:a ex:a))
		entails
	AnnotationProperty(ex:a annotation(ex:a ex:a))
Change (PROPOSED): Revise direct semantics to expand the domain of
	 discourse to include non-individuals.


MINOR CHANGES made for other reasons:

Change: Added optional rdf:Property typing for Annotation Properties in the
	mapping from abstract syntax to triples.
Rationale: To improve internal consistency in the mapping.

Received on Thursday, 5 June 2003 09:49:34 UTC