First International Semantic Web Conference 
9-12 June 2002, Baia Chia, Sardegna, Italia.  
http://iswc.semanticweb.org/

Not a Trip Report by Jacco van Ossenbruggen

Instead, just some things I learned about:

OWL (from a discussion with frankh)
===================================
- owl/daml/oil are based on description logics
- description logics are primarily suited for classification tasks
so if you need other types of reasoning, maybe you should look for
another language (like RuleML).


RDF (from the talk of Peter Patel Schneider)
============================================
- every RDF triple asserts a fact about a resource
- there are no variables, conditionals, disjunctions
So you cannot say 'a is true or b is true'
without asserting both 'a is true' and 'b is true' in RDF ... :-( 

In addition, all syntactic structures of languages built on top 
of RDF will be triples asserting facts. 
So if daml uses 'or', from a rdf perspective,
this can only by asserting a fact about the web resource 'daml:or'. 
So language constructs are in the same universe of discourse as the 
facts in the domain you are modeling, giving rise to all kind of problems. 
(In another way, this is could be handy sometimes as well,
e.g. when talking about different versions of 'daml:or' in different
specifications).

For more, see his paper "Layering the Semantic Web: Problems and Directions"
       http://link.springer.de/link/service/series/0558/papers/2342/23420039.pdf
and "Building the Semantic Web on XML"
       http://link.springer.de/link/service/series/0558/papers/2342/23420147.pdf


Web Services (questions from the audience after the Grid keynote talk 
by Carl Kesselman)
======================================================================
The (XML) Web community, the Semantic (KR) Web community and the (HPC)
Grid community all have similar, but slightly different notions of Web
Services. The first is focussing on the protocols to describe the
interaction with the (static) service (SOAP, WSDL) and to find
services by using registries that are human browsable (UDDI).  The
second is focusing on machine readable registries and composing
composite services out of existing services (DAML-S). The third is
focussing on life-cylce management and coordination of (large number
of) highly dynamic, short term services.  E.g. mobile compute services
that use office computers, traveling around the globe to stay in a
timezone where office hours didn't start yet).

FaCT (from a discussion with Sean Bechhofer)
============================================
In description logic terminology, FaCT has only a T-box, no A-box.
In Semantic Web terminology, this means that it can only reason about
ontologies, and not about instances.  It cannot classify instances, 
for example.


UK-based projects using anotated media (bar discussion after too much
alcohol)
=====================================================================
-Arkive project in Bristol (HP) (http://www.arkive.org/)
-Museum-related project in Bath (?)


also need to have a look at:

- the discourse ontology developed by kmi for scholonto/claimaker
  (http://link.springer.de/link/service/series/0558/papers/2342/23420436.pdf)

- Aidministrator's Sesame
  (http://link.springer.de/link/service/series/0558/papers/2342/23420054.pdf)
-