W3C home > Mailing lists > Public > public-owl-wg@w3.org > December 2007

Re: ISSUE-55 (owl:class) (and ISSUE-56 repairsomerdf)

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Mon, 03 Dec 2007 04:19:00 -0500 (EST)
Message-Id: <20071203.041900.22256472.pfps@research.bell-labs.com>
To: alanruttenberg@gmail.com
Cc: public-owl-wg@w3.org

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Subject: Re: ISSUE-55 (owl:class) (and ISSUE-56 repairsomerdf)
Date: Sun, 2 Dec 2007 16:52:20 -0500

> 
> I'm not sure I consider this a best practice or UDF issue. It is  
> related to ISSUE-56, which I have been thinking of as possibly  
> related to fragments.
> 
> Here is another way to think about it.
> 
> Suppose that we added an axiom (in OWL-Full, for the moment) in the  
> spirit of
> (1) rdfs:Class rdfs:subClassOf owl:Class

Umm. This is already true in OWL Full, 
see http://www.w3.org/TR/owl-semantics/rdfs.html#5.3
where an OWL Full interpretation must have
	IOC = CI
(recall that IOC is defined as CEXTI(SI(owl:Class)) in OWLS&AS Section
5.2 and CI is defined as the CEXTI(SI(rdfs:Class)) ).

The triple is false in OWL DL.

> Let's also assume that if one substituted owl:Class for rdfs:Class  
> everywhere in the ontology, the ontology would be considered OWL-DL.

Which ontology?  You mean in any particular ontology?

So something like:

   Consider an RDF graph that would be an OWL DL ontology if all
   occurences of rdfs:Class were replaced by owl:Class.

> With the addition of (1), then, would we not have a "fragment" of OWL  
> Full that had could be considered to have the same expressivity as  
> OWL-DL?

I'm not sure what you are getting at here.

But perhaps you mean

  Take all the RDF graphs that satisfy the above condition, and add the
  triple (1).  Is this set of RDF graphs co-expressive with OWL DL?

As the graphs are not OWL DL ontologies in RDF graph form, you have to
use the OWL Full semantics, where (1) is an axiom and thus can be
disregarded.

You have to ignore certain OWL DL graphs (namely those that use
rdfs:Class as a type), but this doesn't really reduce the number of
ontologies.  So, yes, syntactically you get the "same" ontologies, as
far as I can tell.

However, you are definitely changing the semantics of the ontologies.
For example, OWL DL ontologies that must have a finite domain become
unsatisfiable.  

> -Alan

peter
Received on Monday, 3 December 2007 09:36:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:13:29 GMT