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

Re: ISSUE 5.14 - Ontology versioning

From: Libby Miller <Libby.Miller@bristol.ac.uk>
Date: Wed, 11 Sep 2002 08:50:34 +0100 (BST)
To: Jeff Heflin <heflin@cse.lehigh.edu>
cc: Libby Miller <Libby.Miller@bristol.ac.uk>, WebOnt <www-webont-wg@w3.org>
Message-ID: <Pine.GSO.4.44.0209110848570.7037-100000@mail.ilrt.bris.ac.uk>


seems fair enough: my point was only that is was just a vocabulary, and
not really a fundamental point of either. But I guess there's no reason
why it shouldnt be in owl, and would definitely be useful.

My other point was that it can be complex to determine who says that
this is the new version of a vocab or ontology, and you might want to
describe that in detail in certain circumstances.

Libby

On Tue, 10 Sep 2002, Jeff Heflin wrote:

> Libby,
>
> I'm not sure you understand your question. Is it "does my versioning
> solution have to specific only to OWL, or could it apply to RDF?" or is
> it "why can't the RDF solution be sufficient for OWL?"  Personally, I
> think these issues are just as applicable to RDF as they are for OWL. My
> hope is that if we introduce them in OWL, they'll eventually work there
> way back down into RDF.
>
> Jeff
>
>
>
> Libby Miller wrote:
> >
> > Jeff, do we need owl-specific properties? Dan Brickley and I wrote a
> > brief document about versioning metadata vocabularies, and I think the
> > issues might hold here too:
> >
> > http://www.ilrt.bris.ac.uk/discovery/2001/06/process/
> >
> > Libby
> >
> > On Mon, 9 Sep 2002, Jeff Heflin wrote:
> >
> > >
> > > Here are my initial thoughts on the issue:
> > >
> > > We need various properties to handle elements of versioning with
> > > ontologies.
> > >
> > > My proposed solution consists of adding three new features to owl. These
> > > are for indicating prior versions, backward copatibility and
> > > deprecation.
> > >
> > >
> > > Prior versions:
> > > ----------------
> > > <url> priorVersion <url>.
> > >
> > > The second URL is an earlier version of the first. This has no meaning
> > > in the semantics, but could be used by software to organize ontologies
> > > by versions. Due to XML namespaces the identifiers in the two ontologies
> > > will be treated as distinct unless there are explicit statements of
> > > equivalence.  Thus, the two ontologies can be merged without problems,
> > > although there will be no "integration" unless specific mappings are
> > > defined.
> > >
> > >
> > > Backward compatibility:
> > > ------------------------
> > > <url> backCompatWith <url>.
> > >
> > > The first URL is a later version of the second, and is "semantically"
> > > backward compatible with it. This is basically syntactic sugar with the
> > > following effects:
> > >
> > > Assuming A backCompatWith B, then:
> > > * A priorVersion B.
> > > * all classes in B are the sameClassAs a class in A with the same ID.
> > > * all properties in B are the samePropertyAs a  property in A with the
> > > same ID.
> > > Note this depends on the resolution of the synonym issue (I prefer
> > > sameAs for both classes and properties).
> > >
> > > The sameAs statements essentially allow you to integrate data that
> > > commit to different (backward-compatible) versions of the same ontology.
> > >
> > > Note, that this approach does not address the problem described in
> > > Section 3.2 of the Requirements Document (under RDF(S) Support). There,
> > > we gave an example where we wanted to "fix" an incorrect definition of
> > > Dolphin.  Note, given the approach here, we cannot make the new version
> > > of the ontology backward-compatible with the old one because that would
> > > say that the class of Dolphins is both a sublcass of Fish and Mammal. I
> > > am still working on a clean way to solve this problem that meshes with
> > > the model theories we are working towards. We may have to save this for
> > > a "next version" of OWL.
> > >
> > >
> > > Deprecation:
> > > -------------
> > > <url> deprecates <classId>
> > > <url> deprecates <propertyId>
> > >
> > > This allows an ontology to deprecate a class or property. By deprecating
> > > the term, it means it still is sameAs a term with the same ID in the new
> > > ontology, but that the term should not be used in new ontologies. This
> > > allows an ontology to maintain backward-compatibility while phasing out
> > > old vocabulary. Deprecation should only be used in ontologies that are
> > > backward-compatible.
> > >
> > > This has no effect on the semantics, but authoring tools should use it
> > > in error checking OWL markup.
> > >
> > > Note that deprecation allows you to break the transitivity of the
> > > backCompatWith relation.
> > >
> > > Jeff
> > >
> > >
> > > References:
> > > ------------
> > > - Dynamics Research Corporation. DRC VES (Versioning) Ontology. At:
> > > http://orlando.drc.com/daml/Ontology/VES/3.2/drc-ves-ont.daml
> > >
> > > - Heflin, J. Towards the Semantic Web: Knowledge Representation in a
> > > Dynamic, Distributed Environment. Ph.D. Thesis, University of Maryland,
> > > College Park. 2001. At:
> > > http://www.cse.lehigh.edu/~heflin/pubs/#heflin-thesis (Sect. 3.4)
> > >
> > > - Heflin, J. and Hendler, J. Dynamic Ontologies on the Web. In
> > > Proceedings of the Seventeenth National Conference on Artificial
> > > Intelligence (AAAI-2000). AAAI/MIT Press, Menlo Park, CA, 2000. pp.
> > > 443-449. At: http://www.cse.lehigh.edu/~heflin/pubs/#aaai2000
> > >
> > > - Klein, M. and Fensel, D. Ontology Versioning on the Semantic Web. In
> > > First International Semantic Web Working Symposium (SWWS'01), 2001. At:
> > > http://www.semanticweb.org/SWWS/program/full/paper56.pdf
> > >
> > >
> > >
>
>
Received on Wednesday, 11 September 2002 03:52:14 GMT

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