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

Re: LANG?/SEM?: using resources

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Wed, 18 Sep 2002 12:50:19 -0400 (EDT)
Message-Id: <20020918.125019.63238296.pfps@research.bell-labs.com>
To: welty@us.ibm.com
Cc: www-webont-wg@w3.org

From: "Christopher Welty" <welty@us.ibm.com>
Subject: Re: LANG?/SEM?: using resources (was Re: LANG: owl:ontology)
Date: Wed, 18 Sep 2002 12:07:07 -0400

> 
> Let me try to clarify my case hopefully using something closer to y'all's 
> hearts.  I realize most of the people in this group are not strictly 
> ontology researchers, so I'll use a KR language example:

I like this example, but it is very different from the other kind of
example.

> Someone comes along and sees the OWL spec.  "Oh," they say, "I like this 
> SameClassAs thing, it's transitive.  I just need something transitive, I 
> don't need the rest of this OWL monster."  So they build a system that 
> uses the OWL:SameClassAs tag, and they use it in cases like this:
> 
> Car OWL:SameClassAs Engine
> Engine OWL:SameClassAs SparkPlug
> 
> The fact is that OWL:sameClassAs does not mean Part.  

Correct.

> If you borrow this tag and use it FOR SOMETHING ELSE, then you shouldn't
> be using this tag.  If you don't agree with the full semantics of
> OWL:sameClassAs, then don't borrow that tag.  Create one of your own.

What *is* the full semantics of owl:sameClassAs?  Does using
owl:sameClassAs commit one to the existence of owl:samePropertyAs (a
consequence of the OWL RDF schema document)?  Does using owl:sameClassAs
commit one to the existence of all legal OWL restrictions (not a
consequence of any RDF or OWL document)?  Does using
owl:sameClassAs commit one to the separation of classes and instances (also
not a consequence of any RDF or OWL document)?  

Does this mean that an incomplete OWL reasoner cannot use owl:sameClassAs?
After all, it is not completely abiding by the ``meaning'' of
owl:sameClassAs.

Even for well-defined, formal resources, it is problematic to abide by
their intended meaning.

> The same goes for borrowing symbols from ontologies.  If you don't want to 
> commit to the full semantics or meaning of a symbol as specified in some 
> ontology, then don't borrow it from that ontology.  
> Make up your own, this 
> tells everyone that you may mean something different.  Borrowing the 
> symbol from the ontology should tell people that you intend the same 
> thing.

I agree that it can be suspect to abuse someone else's terminology.
However, there are cases where one wants to use terminology without
committing to its definition.   (Going outside of OWL, how could one talk
about someone else's beliefs if one had to commit to entire ontologies just
by mentioning terms from them?)

> -Chris


peter
Received on Wednesday, 18 September 2002 12:50:31 GMT

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