ISSUE 5.26 OWL DL Syntax: Proposal to resolve

[This is the proposal as formulated during the editors meeting at the 
Tech Plenary to resolve the OWL DL Syntax issue]

ISSUE 5.26 OWL DL Syntax

PRPOSAL: The following changes to the current syntax are proposed to 
resolve this issue:

A. Annotations & types:

A.1 types required on all non-builtin urirefs
A.2 types optional for builtin urirefs
     [builtin means something defined in our documents]
A.3 object of annotation property can be any uriref (see 1) or literal
     or blank node [xsd datatypes are builtins]
     [pfps: user-defined data types are another issue]
A.4 types requires on blank nodes [includes Lists]
A.5 top level directive in abstract syntax for annotation properties
A.6 permit annotations on annotations

B. Descriptions and bnodes:

B.1 bnodes at descriptions form directed acyclic graphs (DAGs)
B.2 the mapping rules permit T(description1) = T(description2)
     globally, whenever description1=description2
     [doesn't work for Individuals]
B.3 bnodes of lists, all different, individuals, dataranges unchanged
B.4 no acylicity constraint in owl:disjointWith

C. Restrictions:

C.1 OWL Lite excludes multipart restrictions

D. Orpahn descriptions:

D.1 Allow equivalentClasses with a single description, which maps to
     the  description. Result is that ontologies with
     "orphan"descriptions are in DL. They are NOT in Lite.

E. Versioning properties:

E.1 owl:priorVersion, owl:backwardsCompatibleWith and
     owl:incompatibleWith are of syntactic category "ontologyPropertyID" in
     the abstract syntax
E.2 and to add the following production:
     directive ::= Annotation(ontologyPropertyID ontologyID)
E.3 The domain and range of these properties are owl:Ontology
E.4 owl:versionInfo is an annotationPropertyID that may be
     used like any other annotationPropertyID.

F. Named datavalue lists discussion.

F.1 add owl:DataRange as syntactic category for data oneOf
F.2 no naming mechanism for these in DL.  If you name it, you are in Full.

G. Datatypes:

G.1 allow unsupported data types such that they depend on
     identity of RDF literals, but no more.

H. RDF compatibility

H.1 include explcit list of excluded terms in DL/Lite

A. Th. Schreiber, SWI, University of Amsterdam,

Received on Tuesday, 11 March 2003 11:00:33 UTC