SWSL: Minutes for the SWSL Telecon - 28th January, 2004 #38

[Note - these are past minutes from recent SWSL telecons]

=============================================
Minutes SWSL telecon - 29th January, 2004 #38
=============================================

Participants:
   Deb McGuinness
   David Martin
   Steve Battle
   Michael Gruninger
   Daniela Bernardi
   Terry Payne
   Ian Horrocks

Appologies
   Michael Keifer
   Bijan Parsia
   Austin Tate


Notetaker: TP
Next Week: tbd (MG if no takers)

Action Items
------------
Terry:
   To make final adjustments to discovery section and send to Dave who 
will
   integrate into requirements doc

Dave:
   Think about conversations as part of the process model - should be a
   bullet point in the Requirements Document
   To send a note to Michael G. about this
   Send minor feedback to Terry
   Send References to Situational Calculus
   Follow up Mark Burstein on Architecture Group.

Editors of the Requrements Doc:
   Should provide short comments on how the requirements relate to the
   models in the WS Architecture Doc.

Minutes
-------
Discussion on planning the F2F
Most people expect to be attending WWW
Peter/Frank/Ian thinking of running an OWL developers meeting on the WWW
conf at some point.  Tbd?

Discussion Points
- Requirements Doc
- Semantics of Process modelling

No comments on the requirements doc, other than minor ones to be sent by
David Martin, who will move glossary to the end of the doc.
Vote planned on Requirements Doc on 12th Feb 2004
Conversations should be definable in the process model.
DM: What is the conceptual basis for our modelling framework, based on
Service Oriented Model as proposed by Web Services
Architecture Group; http://www.w3.org/TR/2003/WD-ws-arch-20030808/

DM: I have asked Mark Burstein about what the Architecture Committee has
thought about the framework (above) - currently awaiting a reply.

MG: Is the Services Oriented model a strawman Model - which we then 
expand?
SB: Seams too course grained for our needs.
DM: Message-Oriented model is necessary (implicitly in OWL-S, but needs
bringing out)
DM: Policy Model also of central interest
DM: How are we going to use these?
MG: Could be used to form a suite of ontologies.  How do semantics fit
into this architecture.  Agents should reason about services - what 
parts
of this architecture should they reason about?
For example - should the agent reason about the messages - the content
etc?
Deb: I would think they would, including subclassing.
MG: Need to ensure anything we reason about is in the ontology
DM: Looking at the service oriented model (Section 2.3.2)
TP: Agent plays a role of providing several orthogonal(?) services.
Each service may have associated or ancillary services, e.g. "Cost of
doing thing" or "Making a reservation to do thing sometime in the 
future",
where 'thing' is the core service.
DM: Is Agent playing a role?  Is this the right term - could be a local 
of
state?
DM: Want to avoid discussion of Agents and agency for this discussion.
DM: 6 key issues on this diagram
DM: Agents perhaps should have a goal - our view is different to the
model, and hence we'd need to abstract away.
DM: Should our language mention goals?  An early critique of OWL-S was
that the implied architecture of OWL-S omitted goals.
MG: Implicit in the language is that there are relationships between
agents and goals.  Goals could be specified in the language, but not
necessary - it is a modelling design decision  But we could want to have
goals to talk about the goal of an agent, etc  Might need to reason 
about
conflicting goals (though would want to avoid deontic notions if
possible).
DM: this might tie into contracting etc.???
MG: Contracting/Negotiation would use goals.  This is something that 
seams
to be missing from the W3C models...
DM: They have notion of a request - subclass of profile
TP: An agent request typically looks for a service that achieves a goal
MG: Need to think what we need for SWSL - shouldn't be too bound to
whether or not OWL-S has goals, etc.
DM: Need to think about how goals relate to matchmaker requests etc.
MG: We probably need for now to just decide if Goal should appear in the
language.
MG: Use diag as a starting point, but should look at requirements doc.
TP: Do we want to produce our own diag/explanation (based on this and 
the
requirements) and ultimately produce an OWL representation?
MG: Primary work is on requirements - use diag as a reality check
DM: Should therefore proceed on this.

DM: Background on future work.  OWL-S committee have been revisiting
formal semantics of process model.  Consensus is that we map OWL on
to situational calculus.  But not to be an action item for now.  Mebbie
this is what we should do here.  Some discussion on www-rdf-rules.


_______________________________________________________________________
Terry R. Payne, PhD.		| http://www.ecs.soton.ac.uk/~trp/index.html
University of Southampton	| Voice: +44(0)23 8059 8343 [Fax: 8059 2865]
Southampton, SO17 1BJ, UK | Email: terry@acm.org / trp@ecs.soton.ac.uk

Received on Monday, 23 February 2004 22:04:27 UTC