W3C home > Mailing lists > Public > www-rdf-logic@w3.org > December 2003

Re: relation between ontology

From: Drew McDermott <drew.mcdermott@yale.edu>
Date: Thu, 4 Dec 2003 13:54:58 -0500 (EST)
Message-Id: <200312041854.hB4Isvu02058@pantheon-po04.its.yale.edu>
To: www-rdf-logic@w3.org

   [Yongchun Gao]
   Can OWL "unite" all the ontologies? Or OWL is just a way to represent =
   ontology on the Web?

The latter.

   In OWL, there are subClassOf, subPropertyOf, etc. From first sight, it =
   seems that OWL can unite at least some ontology. But giving an simple =

   Suppose someone developed an ontology by OWL, in which "humans" is a =
   class and has "hasGender" as a property (value=3Dmale/female). A man =
   could be an instance of "humans" which "hasGender" of "male". It can =
   work well.

   Suppose anther expert developed an ontology by OWL too, in which =
   "humans" and "animals" are classes, but "females" and "males" are =
   classes too (can be attached to both "animals" and "humans"), and "men" =
   is just two subclass of both "humans" and "males". It may work too.

   But the problem here is HOW to unite these two different OWL files which =
   tell the same ontology?

Import them into a new ontology, using namespace prefixes to keep
'male' in the first ontology a distinct symbol from 'male' in the
second.  The new ontology (which we call a _merged ontology_) then has
a property T1:hasGender and an individual T1:Male that is a possible
value of T1:hasGender, and it also has a class T2:Male. 

Add an axiom, or "rule," that says

   forall (x) ((x member T1:humans) & (x T1:hasGender T1:Male)
               (x member T2:humans) & (x member T2:Male))

Make sure you have an inference engine that can handle these rules.
Many of them handle equivalences by expressing them as a conjunction
of implications, so you may have to break this into two rules.

                                             -- Drew McDermott
                                                Yale University CS Dept.
Received on Thursday, 4 December 2003 13:54:58 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:41 UTC