% notes from JC telecon 9/9/03
% by Benjamin Grosof

participants:
Mike Dean
Benjamin Grosof
Sandro Hawke
Harold Boley
one or two others

o news:
- OMG presentation today by Said and Gerd on rules in Meta Object Facility
[Harold] 
- W3C Semantic Web Services Interest Group has been proposed under
the Web Services Activity, there's a call
for review (i.e. comments) by W3C membership [Benjamin]
- SWSI (Semantic Web Services Initiative) requirements discussions 
during July and August 2003 have
identified nonmonotonic LP-based rules as important in addition to FOL
[Benjamin]


o discussion on the Rules proposal:

Benj: main concerns are layering on top of too much DL and compatibility.
Horn (with Datalog and binary-predicates) is fine as a 
Need a good warning label about layering on top of DL beyond its DLP 
subset will make later extensions to nonmonotonicity be problematic,
and is not well understood at a research level beyond just being FOL.  
Also, recommend not mixing highly expressive OWL-DL expressions into the
rules, since that can be problematic wrt back-compatibility/legacy later.   

Mike:  would be helpful to use similar format to Peter's and Ian's.

Action plan:  will defer to next week more discussion of tweaking the 
current proposal on rules.  

Mike:  Said has been looking at the new Jena 2 rule engine. 
Harold:  apparently it's somewhat hack-y.  
Mike:  it's a first version, so that's pretty common.  

Benj:  would be good if we can hear about that next week.  

Concrete syntax:  there's a current document.  Some revisions are in 
progress, including prefix not infix notation.  

Action plan:  update for the document for next week or two.  
Harold:  yes.  

Mike:  am working on test cases too, to be done before the DAML PI Meeting.
Would be nice to have demos for DAML PI Meeting.


Benj: SWSI (Semantic Web Services Initiative) Language committee's (SWSL's)
requirements discussions during July and August 2003 have identified
nonmonotonic LP-based rules as important in addition to FOL,
e.g., for representing inheritance or pre-/post- conditions in rules.  

Several:  that's interesting

Sandro:  how does that relate to the discussion at the last DAML PI Meeting
with Pat Hayes and Benjamin and others about published nonmon rules vs.
nonmon rules just within a given agent or for a particular kind of use?

Benj:  SWSL requirements discussions have not really made or focused on that
distinction.  They do envision publishing nonmon rules in service descriptions,
and several kinds of uses (e.g., discovery, contracts, composition, exeuction
monitoring) where a given agent can close off the ruleset and
in none of those have the discussions identified publishing nonmonotonic
rules as problematic.