- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Thu, 21 Mar 2002 12:17:51 -0500
- To: ziv@unicorn.com
- Cc: www-webont-wg@w3.org
From: "Ziv Hellman" <ziv@unicorn.com> Subject: RE: SEM: semantics for current proposal Date: Thu, 21 Mar 2002 19:13:59 +0200 > I have a question regarding the interaction of uniquely identifying > properties and cardinalities in OWL. > > That is, suppose that a property p has min cardinality 3 and max > cardinality 5, is defined to be uniquely identifying, and we have two > instances i and j. > > Let the instances associated with i through p be { a, b, c } > and suppose now one wishes to associated with instance j, through p, the > instances { c d e f }. Is this problematic because c is in the range of > both i and j, or is this allowed because {a b c} != {c d e f} ? Yes, this is a contradiction, provided that i ~= j. uniquelyidentifying means that the converse is functional, so if i ~= j then c will have two values for the converse of p. peter
Received on Thursday, 21 March 2002 12:19:38 UTC