International Semantic Web Working Symposium
SWWS

sponsored by the
National Science Foundation, Information and Data Management Program

Stanford University, California, USA
July 30 - August 1, 2001

Semantic Web Working Symposium
30-31 July 2001
Tutorial Track
http://snrc.stanford.edu/~petrie/agents/tutorial.html

Demonstrations

We had 7 demonstrations in 2 hours!
The good news is that there seems to be lots of commercial development though some of the demonstrations seemed shallow.

Ontology Engineering

Natalya Noy
Stanford Medical Informatics
This was an excellent overview of basic issues.
It emphasized the need for application-specific ontologies.
Somewhat surprisingly, it did not emphasize formal semantics, which is what distinquishes ontologies from glossaries.

Semantic B2B Integration

Christoph Bussler
Oracle Corporation
This was an extemely comprehensive overview of evolving standards and issues.
Chris mentioned that there were over 200 emerging standards and this was no joke.
He distinguished between semantic and technical integration and also did not emphasize formal semantics.

Models and Languages
for Describing and
Discovering E-Services

Fabio Casati
Ming-Chien Shan
Hewlett-Packard Labs, Palo Alto
This was a good overview and discussion of a few important emerging standards.
The tutorial focussed on interoperable process standards.
Ontologies wrt formal semantics was not mentioned as an issue.
The issue was raised by the audience: can standards initiatives overcome tendency of commercial interests to have proprietary formats?
The answer was: follow stock price of non-compliers.

Last words:

"Pragmatic Unification".
This is my personal view, as expressed during the conference.
We should do applications and promote applications.
In fact, I believe that ontologies require applications to be meaningful, and that ontological integration is meaningful only with the interacton of applications.
I advise working with commecial E-Commerce folk to use academic expertise.
Otherwise, formal ontology work will go the way of academic software agent technologies.
We should try standards, but modify them based on experience and success.
Let ontological integration depend upon required interoperation of applications.