W3C home > Mailing lists > Public > www-webont-wg@w3.org > June 2003

Re: Proposed response to Golbeck regarding imports issue

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Mon, 16 Jun 2003 23:56:29 -0400 (EDT)
Message-Id: <20030616.235629.125122709.pfps@research.bell-labs.com>
To: connolly@w3.org
Cc: www-webont-wg@w3.org

From: Dan Connolly <connolly@w3.org>
Subject: Re: Proposed response to Golbeck regarding imports issue
Date: 16 Jun 2003 13:10:43 -0500

> On Mon, 2003-06-16 at 13:00, Jonathan Borden wrote:
> > >
> > > From: Dan Connolly <connolly@w3.org>
> > > Subject: Re: Proposed response to Golbeck regarding imports issue
> > > Date: 16 Jun 2003 12:05:35 -0500
> > ...
> > > >
> > > > Yes, namespace declarations do imply that ontological terms
> > > > are being imported! At least in the tools that I build, they
> > > > do. I don't mind if the WG doesn't endorse that position, but
> > > > I do mind if the WG specifies that it's not so.
> > > >
> > > > Please strike that text.
> > >
> > > I strongly disagree.
> > 
> > I also most strongly disagree.
> > 
> > Are you suggesting that the _presence_ of an XML Namespace declaration in an
> > RDF/XML document indicates that the 'namespace' ought be imported.
> 
> More precisely: that the use of the term http://...foo#bar
> implies assent to the contents of http://...foo.

This is contradicted by test Imports 002.

> Namespace declarations are an incidental syntactic detail.
> 
> >  If, so
> > this would suggest that an XML Namespace is to be _identified_ with an OWL
> > Ontology -- if that is what we are saying, then let's say that clearly.
> 
> That's pretty much what I'm saying. I accept that the WG doesn't
> endorse this view.
> 
> I don't accept that the WG has decided to specify that it doesn't work.

See test Imports 002.

> > Since the XML namespace declarations in the RDF/XML source don't end up in
> > an N-Triples representation of the RDF graph, I think we *should* say
> > affirmitively that XML namespace declarations do not imply OWL importation
> > of the namespace document (if any exists) -- otherwise we'd certainly not
> > need owl:imports eh?
> 
> Indeed, I don't believe we need owl:imports. I objected to the WG
> decision, you may recall.
> 
> http://www.w3.org/2001/sw/WebOnt/webont-issues.html#I5.6-daml:imports-as-magic-syntax
> 
> 
> > 
> > >
> > > It is definitely the case in OWL that ``the namespace reference does not
> > > imply that all (or even any) ontological terms from that namespace are
> > > being imported.''  You may write whatever tools you want, but this does
> > not
> > > change the fact that OWL namespace references do not imply any importing.
> > >
> > > In my opinion the removal of that text will leave a mistaken impression.
> > 
> > I'd go further to say that Dan's tools are behaving in an extra-OWL fashion.
> 
> Yes, they're doing more than the spec requires.
> They're not doing anything that the spec should prohibit.

If they are making more inferences than licensed by the spec, then they are
not fully OWL compliant.  Indeed, they are unsound with respect to the OWL
spec, which is generally considered to be much more of a violation of a
logical spec than being incomplete with respect to the spec.

> > > > > > Therefore, it is common to have a
> > > > >  > corresponding namespace declaration for any ontology that is
> > > > >  > imported."
> > > >
> > > > --
> > 
> > It certainly is convenient -- and we could entertain a discussion about
> > whether namespace declarations ought imply importation, but that isn't the
> > current situation -- unless I'm seriously mistaken.
> > 
> > Jonathan
> 
> -- 
> Dan Connolly, W3C http://www.w3.org/People/Connolly/

peter
Received on Monday, 16 June 2003 23:56:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:01 GMT