SKOS and imports


 From the record of yesterday's call [2], I guess that this topic may  
been discussed. However, as Alistair included it in his list of
updates, I wanted to post a message detailing some of my
thoughts. This message does not provide answers, but is intended to at
least record the issue (without raising it as an WG issue at this

In the current document, skos:Concept and skos:ConceptScheme are
owl:Classes. This seems like a sensible position -- SKOS is really an
OWL Ontology with the particular vocabularies as instantiations of
that vocabulary.

However, this means that any use of owl:imports to express a
relationship between concept schemes pushes us straight into OWL Full
-- owl:imports has owl:Ontology as domain and range, so this results
in a violation of the constraint on separated vocabularies (OWL DL
requires that URIs are not used as more than one of Ontology, Class,
Property or Individual. Now, the WG agreed to postpone ISSUE-38 and
decided to build on OWL Full semantics, but this feels a little
uncomfortable to me. The OWL Ref [1] says:

[[ NOTE: The ontology-import construct owl:imports and the
ontology-versioning constructs owl:priorVersion,
owl:backwardCompatibleWith and owl:incompatibleWith are defined in the
OWL vocabulary as instances of the OWL built-in class
owl:OntologyProperty. Instances of owl:OntologyProperty must have the
class owl:Ontology as their domain and range.  ]]

As I read it, this means that even if we live in an OWL Full world,
this means that instances of skos:ConceptScheme that import (or are
imported) will then, by necessity, be instances of owl:Ontology.

I guess I'm wondering whether owl:imports is really the "right thing"
to use here. From the record of the telecon [2], I think Antoine also
raised a question here of issues related to the implementations of
imports.  In my opinion, owl:imports is a rather tricky beast -- it's
one aspect of the OWL documents which I think many people agree could
be clearer. Of course, not using owl:imports would require us to
provide some alternative mechanism, which is itself not an ideal

So, as promised, no answers, but some questions.....



Sean Bechhofer
School of Computer Science
University of Manchester

Received on Wednesday, 16 April 2008 14:55:37 UTC