Bill, (01)
Some comments below. (02)
Leo (03)
_____________________________________________
Dr. Leo Obrst The MITRE Corporation, Information Semantics
lobrst@xxxxxxxxx Center for Innovative Computing & Informatics
Voice: 703-983-6770 7515 Colshire Drive, M/S H305
Fax: 703-983-1379 McLean, VA 22102-7508, USA (04)
-----Original Message-----
From: uos-convene-bounces@xxxxxxxxxxxxxxxx
[mailto:uos-convene-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Bill
Andersen
Sent: Sunday, February 26, 2006 2:10 PM
To: Upper Ontology Summit convention
Subject: Re: [uos-convene] Other Approaches Too. (05)
Hi Leo... (06)
Wow. That was a long reply. Rather than respond to all of it in
detail, allow me to compress your remarks. The fault lies with me if
I've misinterpreted anything you meant. (07)
LEO: That's why I called it "longish", i.e., long-winded English ;) (08)
> LEO: One distinction between entity-relation (or extended
> entity-relation) and UML on the one hand and RDF and OWL on the other
> is that the latter are primarily logics and have a defined formal
> semantics. (09)
Granted - there is a semantics to E-R and UML but they are very
weak. That wasn't what I was getting at, however. Let me put my
reply in the form of a question which I would like the UOS to address: (010)
(1) Is there a difference between a formally-expressed logical theory
and an "ontology"?
(2) If you answered YES to (1), then what is that difference?
(3) If you answered NO to (1), then why is this situation any better
for integration than the current state of data models, other than the
fact that once you posit some correspondences between models, the
logic helps you identify logical inconsistencies? (011)
LEO: Ok, I answer (4) (i.e., no, don't force me into your decision
tree! ;). To me a logical ontology (as opposed to weaker models) is
both A) a logical theory, i.e., a theory expressed in a logic, and B) a
logical theory which purports to be something about the world (real or
possible, the latter grounded in the real world). We can for example
have completely consistent logical theories about arbitrary crap; one
"real world" example: the elaborate consistent models of paranoids.
Given magic premises and the usual logical apparatus, much is possible. (012)
LEO: However, the current state of data and object models is: 1) the
knowledge representation language they use isn't, 2) the models they
create aren't ontologies; 3) if they were concerned about real world
(and possible) semantics AND represented that in a logical language,
then, yes, welcome aboard! They have developed an ontology. But I don't
see (3)in the data or object worlds yet. (013)
> LEO: Data models primarily focus on "local" semantics, not "real
> world"
> semantics, i.e., the very narrow view from a need to organize data
for
> specific purposes and applications in a given part of an
organization. (014)
This is true. The job of a data model (as conventionally applied) is
to provide persistence support to applications. Of course there has
been a tremendous amount of work, dating back to the 70's (I'm
thinking of Bill Kent's "Data and Reality") that argue that data
models, too, need to pay attention to the "real world". Is the only
thing new with "ontologies" that they bring some extra logical muscle
to achieving this attention-paying on a larger scale? (015)
LEO: De facto, all data models are local. There may be those who have
promoted wider visions of those models, but I expect 99% of database
folks (theorists and practitioners) will say: only obliquely are we
concerned with the "real world". (016)
> LEO: I don't see any solution for semantic interoperability/
> integration
> that excludes the use of an upper ontology (or a lattice of theories,
> if you will, some of which are upper theories). (017)
As usual, I agree with you. However, if one counts up the projects
(and here I'm talking about *real* non-academic projects) that use an
upper level ontology to effect, that number is far outstripped by the
number of "beer ontologies" "wine ontologies", etc, that can be
easily demonstrated to be as rigid and non-extensible as data models. (018)
LEO: Agree. These are just well-defined domain ontologies using a
logical language. However, they are at least modelled in a logical
language, unlike those data models. There is value in logic. Even
conceptual models, when modelled in logic, are of more value than
conceptual models modelled in some non-logical, ad hoc language (and
even "language" here is often too strong of a term). (019)
If the purpose of the UOS is to achieve some kind of grand
unification of existing upper levels, I think we're shooting at the
wrong target. That's an interesting project, but I don't think,
given the relatively sparse use of ULO in general, that it's the most
pressing project for those who care about ontology vs logic or data
modeling. There's no or little disagreement among this crowd that
ULO makes sense for lots of reasons. (020)
LEO: I don't think the UOS is after that "grand unification" but in
fact after a commitment to map among the existing upper ontologies or
to identify some subset of all of those upper ontologies on which they
can agree, to promote 1) greater use of the individual upper
ontologies, 2) greater semantic interoperability. (021)
The real target, IMHO, is the view that so-called "domain
ontologies", or domain-specific "upper levels" using formal logics
constitute ontology at all. (022)
LEO: It's unclear to me what you mean by this last statement. The real
target should be to dissuade everyone that: "so-called "domain
ontologies", or domain-specific "upper levels" using formal logics
constitute ontology at all"? Is that what you mean? I think domain
ontologies, domain-specific upper level ontologies, and true upper
ontologies, all expressed in a formal logic, are exactly what we are
after. I don't buy extreme actualism. I am a profligate ontologist, a
scientist/engineer/semanticist not a philosophical ontologist: I want
logically formalized theories that characterize the world (real and
possible) for use by software, with the additional meta-addendum that I
would like it to be as concise as possible, thinking that like the best
scientific theories, the fewer constructs that cover (describe and
predict from) the same domain and map to other well-known theories
simply, the better the theory. (023)
.bill (024)
Bill Andersen (andersen@xxxxxxxxxxxxxxxxx)
Chief Scientist
Ontology Works, Inc. (www.ontologyworks.com)
3600 O'Donnell Street, Suite 600
Baltimore, MD 21224
Office: 410-675-1201
Cell: 443-858-6444 (025)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files:
http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files: http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit (026)
|