Re: SUO: On the relevance of EXPRESS, EPISTLE, etc. to SUO -- RE: RE: RE: SomeProcedural Suggestions
I'm pretty much a novice with EXPRESS, but I'll offer a couple of comments.
EXPRESS comes with a graphical language (EXPRESS-G) which is fairly
"expressive,"
but doesn't include an obvious way to include constraints (that can be
expressed in
either declarative or procedural forms). As far as I know, pure EXPRESS
is acyclic,
so recursive relationships can't be described.
It seems to me that (as Mathew West said in another message in this
thread), EXPRESS
is suitable for expressing ontologies.
As I recall, Peter Wilson has written a paper on expressing set theory in
EXPRESS, but
I don't have the reference. On the other hand, a quick search turned up
this reference
which I haven't checked out yet:
Peter R. Wilson. On the Translation of KIF/Frame Ontologies to EXPRESS.
NIST Report NISTIR 5957, January 1997.
On a related subject, you can find an EXPRESS-G model of the HyTime
semantic grove, as developed by W. Eliot Kimber and Peter Bergstrom
at
http://www.hytime.org/papers/hytimesg/
For those not familiar with EXPRESS, this shows a small but
somewhat complex example, but will give an idea of what
EXPRESS G looks like.
Best,
John Velman
(john.r.velman@Boeing.com )
mfu@redwood.rt.cs.boeing.com (Michael Uschold)@ieee.org on 02/06/2001
06:04:29 PM
Please respond to mfu@redwood.rt.cs.boeing.com (Michael Uschold)
Sent by: owner-standard-upper-ontology@ieee.org
To: Matthew.R.West@is.shell.com, phil.jackson@computer.org,
standard-upper-ontology@ieee.org
cc:
Subject: Re: SUO: On the relevance of EXPRESS, EPISTLE, etc. to SUO -- RE:
RE: RE: Some Procedural Suggestions
EXPRESS is a very expressive language, but it was developed by a person who
did
not have a formal logic bent, so it has nothing approaching the kind of
solid
formal foundation that KIF does. It would no doubt be possible to put
EXPRESS
on such a foundation - which would presumably require making some changes
to
it.
EXPRESS also had different purpose than KIF did. It is designed to express
data
models, not knowledge per se. This is not to suggest that there is
something
intrinsically different about a data model and a knowledge base, but it is
to
say that there are different things that are emphasized, when coming from
one
viewpoint, vs another.
Mike