From: Peter F. Patel-Schneider (pfps@research.bell-labs.com)
Date: 01/09/01
From: Lynn Andrea Stein <las@ai.mit.edu> Subject: Semantics questions Date: Tue, 9 Jan 2001 17:16:45 -0500 (EST) > These comments are based on > http://www.daml.org/2000/12/semantics.html, which may not be the > current version. Also, I'm not entirely sure what forum this should > go to, so my apologies if this is too narrowly sent; you may freely > forward it to a more appropriate forum. > > I have some questions about the relationships between IC, IR, and IO. > It seems clear to me that for this semantics to pertain to > DAML-in-RDF, the domains of these three functions cannot be disjoint. > (As I understand it, classes, restrictions, and relationships are all > individuals in RDF.) The domains of IC, IR, and IO are not disjoint. I find that this is unfortunate, but this is the way things are. :-( > Thus, it seems to me that equivalentTo constrains not just IC but also > IO. > > Also, if samePropertyAs is supposed to be a subPropertyOf > equivalentTo, equivalentTo ought to constrain IR as well. Actually, equivalentTo has to be semantics-free. I have changed the semantics document accordingly. > Also (reasonably unrelated), shouldn't the definition of inverseOf use > an iff?: > > <inverseOf, ?R, ?S> should be defined as > > <x,y> in IR(?R) *iff* <y,x> in IR(?S) > > > > Lynn This has been changed. peter
This archive was generated by hypermail 2.1.4 : 04/02/02 EST