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

Re: ISSUE 5.14 - Ontology versioning

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Tue, 10 Sep 2002 12:21:39 -0400 (EDT)
Message-Id: <20020910.122139.121855705.pfps@research.bell-labs.com>
To: hendler@cs.umd.edu
Cc: www-webont-wg@w3.org

From: Jim Hendler <hendler@cs.umd.edu>
Subject: Re: ISSUE 5.14 - Ontology versioning
Date: Tue, 10 Sep 2002 11:58:46 -0400


> ******* chair hat off **********
> It is my personal opinion that treating 5.6 and 5.14 as extralogical 
> would address our requirements and be sufficiently described so as 
> not to require treatment within the semantic documents.   I am happy 
> with owl:ontology being something defined within the semantics, as 
> long as we can insert the extralogical tags (4.4) into the definition 
> in a way that is comfortable to that semantics.  I feel it is more 
> important to be able to put the various metadata, versioning, and 
> descriptive features into the ontology header than it is to have 
> owl:ontology be an overly constraining definition.  I also worry that 
> being constraining in our definition of the "boundaries" of an 
> ontology, we would hurt our various use cases, objectives and 
> requirements that involve ontology sharing, mapping, 
> interoperability, etc.
> ********** chair hat back on *********

Well some of this is definitely doable.

In particular, attaching non-logical tags to classes and properties can be
done using RDF properties in an embedding of OWL into RDF.  It might even
be possible to have RDF resources corresponding to ontologies, which then
could have attached non-logical tags.

However, an acount of importing, of necessity, involves logical consequences.
Capturing only the syntactic portions of it is insufficient.  Ontology
versioning also has logical consequences.  If OWL is to be layered on top
of or even embedded into an extension of RDFS, then these consequences must
be captured somehow.  For if not then what use is a partial layering or
embedding?  We might as well not bother further.

Received on Tuesday, 10 September 2002 12:21:50 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:56:47 UTC