Re: summary: acyclic class hierarchy

From: Peter F. Patel-Schneider (pfps@research.bell-labs.com)
Date: 08/29/01


The situation may not be as bad as Frank makes it out.

It would be possible for DAML+OIL to use rdfs:subClassOf as the main stated
relationship between classes.  This would be ``proper subclass''.  The
relationship that is inferred between classes, would not, however, be this
relationship, instead being ``improper subclass''.

I do not advocate this approach, however, as it would require significant
changes to the ``natural'' way of expressing subclass relationships between
defined classes.  One would have to take care that the relationship between
two defined classes was indeed ``proper subclass'' and that the two defined
classes do not happen to be co-extensional.  This would be particularly
burdensome for DAML+OIL because of the distributed definition of classes in
DAML+OIL, as additional information could cause two classes to be
co-extensional.  

Because of this burden, I would expect that most subclass relationships
would be expressed using ``improper subclass''.

peter


This archive was generated by hypermail 2.1.4 : 04/02/02 EST