% notes from JC telecon 8/19/03
% by Benjamin Grosof

agenda: 

o news of various kinds

o more discussion of Peter and Ian proposal on OWL-ish Rules 

participants:
Mike Dean
Benjamin Grosof
Peter Patel-Schneider
Harold Boley 
Said Tabet
Ian Horrocks
Sandro Hawke
Deborah McGuinness
Pat Hayes
? perhaps one or two others


announcements:

o Sandro and Mike:
OWL has gone to candidate recommendation status as of today!

o Benj: 
nonmon inheritance seems an important requirement for service descriptions
and semantic web knowledge to support that, e.g., 
- see my recent paper 
- recent SWSI Language Committee discussions 

o IJCAI report by Ian:

not very interesting place, 
difficult to get to, lots of lost luggage
50% got Montezuma's revenge the second day

disappointingly light in substance on Semantic Web, tho' SW was mentioned
a lot as related

didn't attend workshops

structure of the conf:  lots of invited talks this time, good in themselves
but drained some energy from the technical sessions

invited talk by Mehran Sahami of Google (formerly of Stanford) on directions
for Web search:
didn't mention SW at all in main, was mainly on standard IR stuff; his 
A to Q at end about SW was basically non-committal in wait-and-see spirit,
partly since there's not a lot of markup, wasn't terribly encouraging, but... 
did invite folks to come to Google with ideas, they seem pretty open to it
- Benj:  I know Mehran, his background is more learning than logic-y
- Deborah:  another contact at Google is Peter Norvig

Deborah Q:  what about SW for query expansion?
A:  didn't get asked

they bought 
Applied Semantics, co with big linguistics-oriented ontology (Wordnet-y),
which did/do placements of ads by words on search engines
- perhaps to hurt Google's competitor Overture

Alon Halevy and Craig Knoblock and some other invited talks on data
integration mentioned SW but didn't really give substance on
connection to that

session on AI and Internet (Ian didn't attend)

several DL sessions on supporting technologies 
- nice paper by Franz Baader on tractable alg for expressive subset with 
conjunction and existentials and some cyclic definitions;
some large ontologies (e.g., _____ [? Snomed ?]) actually fall in this subset
- some papers on role inclusion axioms and keys
- papers by Heiner ___ and ___ on 
reasoning about multiple distributed ontologies on the web;
still preliminary

%%%

More discussion of Peter and Ian proposal on OWL-ish Rules:

discussion around RDF and OWL-Full expressiveness -- Pat msg (under
name of "Brandon Amundson") from earlier today

discussion around bnodes and existentials
- Benj:  worried about extensibility to LP not just FOL, thus want to 
stay Horn; to represent bnodes as existentials is problematic,
so should give up either binary or Datalog or Horn;  

discussion around whether or not in Rules Lite to go for something in the
expressive intersection between FOL and LP, thus extensible to it
- a split among the group in this regard, when considering the combined
{OWL ontologies + rules} KR's expressiveness: 
 . some want {(all of OWL-DL+ + Horn}
 . some want {DLP + Horn}

one approach:  permit the rules to be Horn, or reducible to Horn,
even if when combined with OWL-DL 

suggestion/plan:  by early Sept., someone make a new iteration of 
proposal for LP-based onto + rules