Minutes W3C WS Choreography WG con call held on 6^th May 2003, 1pm PDT.

Agenda:


1.            Role Call


2.            Confirm scribe


The following is a list of recent scribes (in order): Mike Champion, Abbie Barbir, David Burdett, John Dart, Carol MacDonald, Yaron Goland, Daniel Austin, Jim Hendler,  Peter Furniss, Ed Peters, Greg Ritzinger, Leonard Greski


3.             Approve minute

4.


5.             Minutes29April.htm


6.             Action Item Review


7.            Dynamically changing participants in a choreography (hot topic from public mailing list)


Various emails plus some usecases


1.                         1.   FinancialServices Use case


2.                         2.   Use case on choreography negotiation


8.                 Formal models and choreography (hot topic from public mailing list)


a.                      The value of the value of the pi-calculus


b.                   What about petrinets


c.                   What is a process (hot topic from the mailing list)


9.             AOB.


10.        Summary of New Actions

Role Call:

Chairs:
Martin Chapman	Oracle
Steve Ross-Talbot	Enigmatec

W3C Staff Contacts
Hugo Haas

Members:

Yaron Goland	BEA Systems
Mayilraj Krishnan	Cisco Systems Inc
Francis McCabe	Fujitsu Ltd
Yoko Seki	Hitachi, Ltd.
Assaf Arkin	Intalio Inc.
Ravi Byakod	Intalio Inc.
Richard Bonneau	IONA
Sanjay Patil	IONA
Eunju Kim	National Computerization Agency
Abbie Barbir	Nortel Networks
Greg Ritzinger	Novell
Jeff Mischkinsky	Oracle Corporation
Nickolas Kavantzas	Oracle Corporation
Kevin Liu	SAP AG
Michael Champion	Software AG
Carol McDonald	Sun Microsystems, Inc.
Jon Dart	TIBCO Software
Daniel Austin	W. W. Grainger, Inc.
Jean-Jacques Dubray	Eigner

Raw IRC log at: http://www.w3.org/2003/05/06-ws-chor-irc

Appointment of scribe:

Francis McCabe, Fujitsu, kindly volunteered to scribe for the meeting.

Minutes Approval:

Draft minutes at http://lists.w3.org/Archives/Member/member-ws-chor/2003May/att-0008/Minutes29April.htm

Minutes approved with minor comments sent to list and role call updates.

Action Item Review

ACTION : F2F planning.

IN PROGRESS

New ACTION: on Yves should ensure a registration page for f2f

New ACTION: Daniel to post logistics information for June f2f

ACTION : Harvesting of use cases by Steve

Done See is http://lists.w3.org/Archives/Member/member-ws-chor/2003May/0019.html

10 potential use cases

Steve needs comments, brick-bats and other bats

new version of requirements in approx 2 weeks from Daniel

missing callback use case (two use cases in Mchapmans doc)

ACTION : Everyone should review requirements doc and provide feedback to editors. Editors consider response satisfactory.

IN PROGRESS

ACTION : Everyone should read CSF part of WSA spec. Daniel sent list of CSF references.

IN PROGRESS

CSF sounds good

we need consensus? on the CSF approach for requirements

Group agreed to work on CSFs

What is our mission statement ?

New ACTION: Daniel to kick off discussion on mission statement

WSA is commenting on choreography

New ACTION: John Dart to summarize WSA position on choreography wrt mission statement and goals

New ACTION: summarize use cases from privacy discussions

ACTION : (clarifying) SRT will harvest use cases, URIs, and publish them – NO PROGRESS

ACTION: SRT will continue to monitor use cases -- He spent 5 hours going through emails and summarized to list.

ACTION: ALL, feedback on usecase summary to SRT.

ACTION : Collect list of technologies we should harvest from (Open)

Daniel: What are existing technologies we need?

Martin: key input technologies

JJ: I volunteer for BPSS

CarolMcD :subject experts reporting sounds like a good idea

Yaron: Gee, I'm guessing I could take BPEL

SRT: Leveling the playing field on new technologies based on appointed experts gets my vote.

Daniel: Need volunteers to represent and explore technologies.

Examples: BPEL, WSCI,

Steve: more examples: BPSS, BPML, DAML-S

JJ: PIPs are probably covered by BPSS

Arkin: XPDL?

CarolMcD: bpml damls pips

Yaron: I thought RosettaNet hadn't officially adopted BPSS?

John takes on Rosetta-net?

Jj: Yes but BPSS was created by RN people

Main features, why are we interested, and maybe why not

WSCI is a significant input

We need to know what is good and what is less than good

New ACTION: Carol to look into WSCI and to pester Assaf

ACTION : SRT will capture requirements from this discussion OPEN

ACTION : SRT let's capture these use cases in our documents open

ACTION: DavidB should send these use cases in to the list open

Dynamically changing participants in a choreography

In financial services, need to participate with different partners chosen on a dynamic basis

Rules of engagement can be seen as choreography

Which attributes are relevant to choreography?

Modeling of time and observable state

Participants change based on the logic of the choreography

Can you dynamically bind different endpoints to role during a choreography

Do we need this power?

Can I dynamically introduce new role/types on the fly?

What about a trader changing roles during a trade?

Can you change roles during a choreography

What is a role?

Agents must? be able to take on different roles?

JJ: role is a place holder for a binding

MChapman: A role defines responsibilities of a particular partner in a choreography

JJ: Can you have changing roles if the choreography is static?

JJ: more than one level of role.

JJ: E.g. is buyer a role

JD: Delegation allows one to enter a choreography and delegate to another the completion of the choreography

Mchapman: when a choreography is instantiated, endpoints have to be bound. (Buyer = A, Shipper=C, Seller=B) Are these immutable - that is the question.

CarolMcD: bound or not bound - that is the question

Frankmccabe: sees two questions: do we want to allow dynamic re-assignment of roles, and techniques for doing this

Steve: if a role is public then anonymity becomes very difficult.

Steve: a trader does need to change roles dynamically

overlapping choreographies active simultaneously

MChapman: buyers and seller, credit card agencies represent roles.

MChapman: a role refers to a participant in particular MEP

Trader has to has selling interface, trading interface, price fixing interface ...

mchapman: roles are statically defined, participants are not defined?

we want to be able to define choreographies to permit participants to change roles dynamically

jd: Requirement to permit participants to dynamically join and leave choreographies. This could

likely fulfill the use cases

that call for role re-assignment. E.g. a trading scenario can be

modelled by having a participant

enter one or more choreographies (possibly overlapping in time) as a

buyer or seller. It is not

necessary to have a participant bound into a buyer role morph into a seller.

Frank McCabe followed up by remarking: "an auctioneer isn't ever a

punter",

JD: making (I think) the same point; roles are distinct, transformation of role isn't necessary.

Formal models

SRT:most has been said about pi calculus

JJ: two comment on formal models

JJ: pi-calculus good model for theory, but it has many concepts missing

SRT:I like the idea of the notion of auctioneer never being a punter. The should be true in wholesale banking. This is what Eliot Spitzer wants from investment banking. It's case of enforcing chinese walls.

JJ: chor is an expression of a sequence of message exchanges.

JJ: can be modeled in pi-cal but modeling is difficult

??: (Do I hear intractable?)

jj: Expressivity is low

jj: all sides need to be understood together.

Nick: Agree in parts A formal model describes an abstract computation not a complete language

nick: Expressiveness is different to programmabliity

Mchapman: pi should be seen as an assembly langauge type thing. we need to derive the appropriate abstractions useful to define choreography

SRT: +1 to nick - he says showing his bias

nick: Pi-calculus's composition gives a global view?

nick: maybe there is a subset of pi; e.g., CCS.

Frankmccabe: (PI is not enough - my comment)