Re: URIs : How to find the ontologies behind them

To: Eli@SemanticWorld.Org
Cc: www-rdf-interest@w3.org
Subject: Re: URIs : How to find the ontologies behind them
From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Fcc: +Outgoing
In-Reply-To: <015a01c2feae$9a872d30$5c14a8c0@unicorn.co.il>
References: <013e01c2fe7f$ea98cf60$5c14a8c0@unicorn.co.il>
	<20030409.090133.87611393.pfps@research.bell-labs.com>
	<015a01c2feae$9a872d30$5c14a8c0@unicorn.co.il>
X-Mailer: Mew version 2.2 on Emacs 21.1 / Mule 5.0 (SAKAKI)
----
From: "Eli Israel" <Eli@SemanticWorld.Org>
Subject: Re: URIs : How to find the ontologies behind them
Date: Wed, 9 Apr 2003 18:42:18 +0300

> ----- Original Message -----
> From: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
> Sent: Wednesday, April 09, 2003 4:01 PM
> Subject: Re: URIs : How to find the ontologies behind them
> 
> 
> > From: "Eli Israel" <Eli@SemanticWorld.Org>
> > Subject: URIs : How to find the ontologies behind them
> > Date: Wed, 9 Apr 2003 13:08:04 +0300
> >
> > > A question about URIs:
> > >
> > > The URI for a class does not have to point to a particular resource on
> the
> > > web, it just has to be unique.  An ontology describing that entity may
> be
> > > located somewhere else entirely.
> > >
> > > If an ontology refers to this class by its URI, how is additional, or
> even
> > > primary, information about that class supposed to be found?
> >
> > Well, just about anywhere, at least in the general case.  For example,
> > suppose that the class is rdfs:Class.  Information about rdfs:Class can be
> > found in just about any RDF document.
> >
> > Of course, there are very many cases where a lot of information about a
> > class (or any other property) should be found (maybe not now, but when the
> > Semantic Web actually gets going) by dereferencing a URI related to the
> URI
> > references of the class.  Of course, this would only be one organization's
> > information about the class, and other agents might reasonably have
> > different views.
> Thanks for the quick response.
> 
> I'm primarily interested in the 'when the SW gets going' case (it's about
> time we get it going, no? ;).  You seem to be saying that a best practice
> would be to put the OWL describing the resource in the place that the URI of
> the resource refers to.

Not necessarily in that exact place.  However, most RDF URI references have
fragment identifiers, so it would be possible for example to place
information about http://foo.ex/bar#bax in a document located at
http://foo.ex/bar.  

Is this going to be the way it is often done?  I don't know, but it is a
possibility.

I certainly not advocating that this be the only place that shared
information about a resource should be placed.  Neither am I advocating
that the use of a resource would automatically commit an agent to this
information.

> If OWL documents are named seperate from their namespaces, and an agent
> can't find the document by following the URI, it would have to rely on an
> index of documents describing entities.  How would these documents be found,
> registered, etc.?  They wouldn't be naturally interlinked through the
> ontologies, they would have to be dug up on the www, or registered in a
> repository.

Another possible way to obtain information about resources would be through
distributed ontologies.  OWL has an imports construct that allows
ontologies to to be connected together.

> It seems like the easiest way is to place the ontology describing the
> concept at the URI for the concept, no?

Well, I don't think that there should be one ontology for each concept, so
I don't think that this is going to work well.  I think that the
relationship between ontologies and URI references is many to many (and
many to many in interesting ways).

peter

Received on Wednesday, 9 April 2003 13:07:59 UTC