ISO IRDS meeting in Tokyo

sowa@watson.ibm.com
Message-id: <9111210709.AA14524@quark.isi.edu>
Date: Thu, 21 Nov 91 01:56:19 EST
From: sowa@watson.ibm.com
To: SRKB@isi.edu, INTERLINGUA@isi.edu, KR-ADVISORY@isi.edu,
        JFULTON@atc.boeing.com
Subject: ISO IRDS meeting in Tokyo

During this past week, the ISO IRDS committee has been meeting in Tokyo.
One of the items of business is to consider the ANSI IRDS working paper,
which I circulated to this mailing list on October 18th.

Since no one from the X3H4.6 group who is conversant with conceptual
graphs was able to get to Tokyo this week, we wanted to have someone who
could answer any questions that the ISO committee might have.  It turns
out that Peter Eklund, who hosted the 1990 conceptual graph workshop in
Stockholm, is now a visiting research fellow at Hosei University in Tokyo.
Therefore, we arranged for David Gradwell, who is the chair of the ISO
IRDS committee, to meet with Peter over dinner in Tokyo.

Apparently, the meeting went quite well.  Before going to Tokyo, Gradwell
visited Boeing, where he talked with Jim Fulton, who has been developing
the Semantic Unification Meta-Model (SUMM) for PDES.  SUMM is based
on predicate calculus, and Fulton and I have agreed to work towards a
semantically equivalent convergence of the PDES SUMM and the ANSI IRDS
version of conceptual graphs.  At dinner last night, Eklund was able to
convince Gradwell that CGs and SUMM are compatible enough and expressive
enough to represent the kinds of things that the IRDS requires.

Following is a note from Peter Eklund about the meeting with Gradwell.
This outcome is very good for a first exposure, and there will be a
more formal presentation to ISO next spring, after the ANSI technical
report comes out in January.

John Sowa
_________________________________________________________________________
Date: Thu, 21 Nov 91 13:50:28 JST
From: peter@keiei.hosei.ac.jp (Peter Eklund)
To: sowa@watson.ibm.com
Subject: Meeting with David Gradwell

John,

I met David at his Hotel last night at 7:20 and we had supper together
along with Cliff Sundberg from Digital. David asked all manner of strange
and weird questions about conceptual graphs. I took a copy of the book
with me and showed it to him - he had never seen it before.

It seemed that his ambition was to sketch up a first "cut" of the IRDS
model using conceptual structures. A fairly ambitious goal over a two
hour meeting for someone who has never heard of conceptual structures
before.

The questioning took the form of two sessions in between which I managed
to eat a few bites. In the first, which I thought was totally unsuccessful
>From my point of view, David asked me about concepts. Is an instance a concept?
Is a relation a concept? How do we type structures? What about strings,
action-flows, contexts (domains of discourse) etc. I found these questions
difficult because I had trouble recalling all the representation detail from
the book. It wasn't enough to simply say, yes we can deal with all these
issues in CGs, he wanted to know how. Fair enough, but it put me on the
spot.

After a brief respite we started again. This time David took a concrete
example out and we coded it up. Apparently the example originates from
Jim Fulton, who had also authored a very informative Meta-model flow
diagram for the ISO IRDS which was shown to me (I can send you all the
material which he gave to me if you don't already have it). This seemed
like David's favourite and he kept trying to relate CGs to it. Fulton's
IRDS Meta-model is classical, with concepts like "terms", "predicates",
"sentences" relating to intensional objects and extensional "things".

This session was much more successful. In the process of coding up the
example, David was able to gain sufficient insight into CGs to realize
that they were powerful enough to clearly and simply represent his example.
As well as revealing the syntax of CGs, it also demonstrated something
of the ontology or meta-structure of CGs, how relations, concepts, contexts,
propositions etc hang together. I think working through the example turned
the tide of the meeting and created a favourable impression for CGs.

At about 9:30 they closed the restaurant and we left. David seemed relatively
pleased with what I had told him and promised to take up what he had drawn with
Sandra Perez. I have to say that I really wasn't prepared to give a "ground-up"
lecture to someone on conceptual structures. I would have assumed some knowledge
of them but there was none. Nor was there much commonality in terms of logic
or AI. This I hadn't realized beforehand. Still, I think David was sufficiently
experienced to make the most of the situation by asking the right questions.

Please let me know if there's anything else I can do (or inform you
about) regarding the ISO meeting. David told me that he was interested
in reconciling the ANSI and ISO groups but that the trouble was that
ANSI hadn't decided what it wanted regarding IRDS.

Peter Eklund
Department of Systems Engineering
Hosei University
Kajinocho, Koganei-shi
Tokyo 184

Tel: +81-423-876353
Fax: +81-423-876126

email: peter@keiei.hosei.ac.jp