Minutes W3C WS Choreography WG con call held on 15th July 2003, 1pm PDT.

Agenda:


1.            Role Call

2.            Confirm scribe

3.            Approve minutes

 

4.              Action Item Review

                        

5.              Standing tracking items (a section designed to ensure that longer running items are properly tracked)


·        Mission Statement 

·        Requirements (progress/review)

·         Issue tracking (progress/review)


·         Task forces (progress/review)

 




6.            Threads (on the archive not covered elsewhere in the agenda)

·         Composition

·         Choreography protocol

·          Choreography State Definition

·         Proposed text for Choreography concepts

·         Decision Points Requirement Proposals

7.            AOB.

 

Role Call:

 

 Martin Chapman

Oracle

Steve Ross-Talbot

Enigmatec

 

 

W3C Staff Contacts

 

Yves Lafon

 

 

 

 

 

 

Members:

 

 

Yaron Goland

BEA Systems

Anthony Fletcher

Choreology Ltd

Mayilraj Krishnan

Cisco Systems Inc

David Burdett

Commerce One

Francis McCabe

Fujitsu Ltd

Yoko Seki

Hitachi, Ltd.

Eunju Kim

National Computerization Agency

Greg Ritzinger

Novell

Kevin Liu

SAP AG

Ugo Corda

SeeBeyond Technology Corporation

Michael Champion

Software AG

Monica Martin

Sun Microsystems, Inc.

Jim Hendler

University of Maryland (Mind Lab)

 

Raw IRC log at: http://www.w3.org/2003/07/15-ws-chor-irc

 

Appointment of scribe:

 

Jim Hendler, MIND Lab, kindly volunteered to scribe for the meeting.

 

The following is a list of recent scribes (in order): Kevin Liu, Tony Fletcher, Jon Dart, David Burdett, Ed Peters, Greg Ritzinger, Monica Martin, Len Greski, Jean-Jacques Dubray, Monica Martin, Mayilraj Krishnan, Francis McCabe, Michael Champion, Abbie Barbir, David Burdett, Jon Dart, Carol McDonald, Yaron Goland, Leonard Greski, Ed Peters, Greg Ritzinger, Daniel Austin, Peter Furniss, Jim Hendler

 

 

 

Minutes Approval



 srt: need to approve 3 sets of minutes

 

June  face to face minutes: http://lists.w3.org/Archives/Member/member-ws-chor/2003Jul/att-0038/June_F2F_2003-0.htm

 

correction on f2f minutes - amended to reflect a change on one of the comments. Greg to send in correction

corrected minutes at : http://lists.w3.org/Archives/Member/member-ws-chor/2003Jul/att-0052/June_F2F_2003-0.htm

 

RESOLVED: approve minutes of f2f as true and correct record (as amended)

 

1st of July telecon: http://lists.w3.org/Archives/Member/member-ws-chor/2003Jul/att-0038/Minutes_1July2003.htm

 

RESOLVED: approve minutes of July 1 telecon

 

8th July Telecon: http://lists.w3.org/Archives/Member/member-ws-chor/2003Jul/att-0038/Minutes_8July2003.html

RESOLVED: approve minutes of July 8 telecon

 

 

 

Action Item Review

 

 

ACTION: DA will do the wordsmith, summarize and put together revised mission statement. Ongoing. See agenda item below

 

ACTION: Chairs, Arrange hosting for September 2003 face-to-face in Seattle.

 

ACTION: Chairs, Oversee the organisation of the task forces. Ongoing. Not all TFs kicked off yet

 

ACTION: Martin, Track through past minutes for issues and put them into Bugzilla. Ongoing.

 

ACTION: Chairs, Have an agenda item on the requirement to have some level of understanding

of the meaning of messages and what that level should be, when Jim Hendler can be on the call –

perhaps put on the agenda for two weeks' time. Ongoing

 

 

 

 

ACTION: Chairs, Include ‘Burdett’ ML on the agenda when David is back (15th July, 2003 teleconference?) and

 decide what to do with it. Ongoing put on agenda for 22 July.

 

ACTION: TF owners, The owners of each task force to put together a task list for that task force, agree with task force members, and mail to the main group. Ongoing. 

 

ACTION: AlB will develop a use case / scenario for this broadcast use case. Ongoing

 

NEW ACTION: SRT to contact Alistair on the progress of his action.

 

 

ACTION: DB - Need design collaboration use case.  Ongoing

 

 

ACTION: Yves, determine what format is required for the documents we produce.  Done. email sent [ed I see no email in the archive]

 

 

ACTION: DB to check with Commerce One legal, issues around IP and report back. Ongoing.

 

ACTION: SRT/MC to find venue in Seattle. Done (see next action)

 

ACTION: Steve ask Attachmate to make a formal offer to the group. Done

 

 

discussion of Attachmate offer.

 

Formal invite from Attachemate sent to list [ed I see no email in the archive]

 

no objections to Attachmate offer, thus the chairs will accept their kind offer.

 

NEW ACTION: Chairs to accept attachmate offer, work out logistics

 

ACTION: YL to give GR and DA editing access to Bugzilla. Done

 

ACTION: MC/SRT to summarize issues process. Done

See http://lists.w3.org/Archives/Member/member-ws-chor/2003Jul/0040.html

 

ACTION: Greg will go through the June F2F minutes and log those issues. Done

 

 

ACTION: Martin propose text for call-back use case. Ongoing

 

ACTION: Martin to post a straw poll via email about having f2f in Australia in Dec. Done.

Poll at: http://www.w3.org/2002/09/wbs/34340/20030709/ and continues another day or two

 

Standing tracking issues:

 

Mission statement work continues, resubmission of requirements continues

use cases are being written up as separate documents that can be edited in

 

            Requirements

spreadsheet sent to WG - for those with problems w/HTML

yellow titles in reqs sheet - in "category" is a list of numbers --

 

Sari Shein is really Steve Ross-Talbot (don't ask)

 

Steve: attempts being made by Sari/Steve to link the requirements to the emails where they were raised

Steve: try setting your user in the general preferences

aim is to edit requirements text which will make requirements doc and requirements text correspond (and the uri will take you to the email it was raised in)

final version expected Thurs or Fri

comma separated numbers at end are an attempt to link reqs to each other

Daniel says we expect a reasonable draft for discussion by next conf call

 

currently around 68 reqs (we started from 0)

please send comments to SRT copying Daniel and other editors - please don't edt the sheet - send sep. comments

 

Issues Tracking

Greg has entered issues from f2f. Martin has action to harvest from other minutes.

Martin sent out issues process – comments please.

 

 

Task force reports

 

Monica: a task list is being sent

Input task force - Mayilraj has joined the group. See mail tagged by [InputTF]

Requirements taks force is moving along well (even if they didn'tknow they were a task force)

 

 

Composition

 

srt: Monica sent an email, Tony sent some followup, lots of discussion

discussion of "web service composition" v. "choreography" composition

DavidB: composing choreographies between sets of choreographies makes sense

DavidB: but if we focus on composing a single service, then it seems we are in BPEL space

martin: web service comp results in a new web service -- chor composition results in something diff

DB: but you might expose this as a composition

Discussion of exposing the composition as a web service

mc: don't think a choreography is exposed as a web service

DB: discusses composition of a choreography -- quote followed by order example

mc: but we need to be careful to distinguish WS comp vs. Chor comp - so we need a good definition of WS comp

MC: think we build choreographies on top of WSs

DB: more like building a program out of modules - the thing created could be a program itself

DB: but if there is no single entity in control - then maybe this is a chor

srt: is a chor executable? (discussion of whether this is true)

tony: it's not executable in the sense of being executable on a single processor/process etc

tony: I think it requires at least two process cooperating to be a chor

tony: from these we could build more complex, multiparty choreographies

Tony: discussion of roles and how they're used in choreographies

tony: summary - 1 - two systems and a single message between them is elemental chor, 2 - adding messages before or after v. adding a new role - latter is core comp

jimh-scribe, you wanted to ask where recursion grounds

jimh: tony's model of how to ground a chor makes sense to me -- if the recursion (composition) ends at a WS instead of a message, then WS Comp == Chor Comp

david: sending messages may be where this grounds (lots of background stuff could be happening)

davidB: proposes these things are in sort of a hierarchy -- the business chor is backed by the messaging chor which is backed by ...

This means that we ground at a minimal system that has two service that run in parallel and have some message exchange, n'est pas?

db: so it is really abstraction that is important

db: this ties into composition as well -- the higher level can refer to "states" of the underlying protocol - and this means same chor could ground in multiple ways

yaron: makes sense to me -- and we also want to surface things that could be "messages" even if they're not a message per se

yaron: so we would need a couple of ways to somehow make things look like messages

yaron: and second issue would be what do we expose in the chor

db: and this sort of state-based thing could be useful

MC: are we getting off topic? DB: this is sort of a different view of what a composition it

MC: important - this ground differently -- i.e. a chor doesn't ground to a WS, but to something else (message or whatever)

srt: so how would I use a chor if it isn't realizable as a WS -- how do I hook into one as an "observer"

mc: would there be entry points to a chor -- (answer own question) yes, and these entry points would be based on messages. Observability is more like "sniffing" or something

mc: so observability would have to be something where you watch the participants

srt: but you are talking about entry points for composing w/another chor -- but how do we synchronize, link? Must be an input/output - and some way to say "these now cooperate -- and like this"

mc: so challenge would be to come up w/this mechanism

srt: (and Jeff M said something like this) -- if you have two chors, and you dont' need to change them, then they're really one chor, not two...

srt/mc: If you have two choreographies, A and B, and there is no observable commuication between them -- they are clearly two separate chors

srt: but if there is communication between them (some glue) then it becomes one chor (i.e. so composition is adding the glue)

mc: suppose we have B and C and no observable between them -- I now create a chor called A, which links to B and C, then I've created one Chor of A,B,C (and this is composition)

DB: and Tony's example is exactly this

mc: but we have to be careful about instances v. definitions of choreographies

jimh: we need to make sure to get the "recursion" correct -- that is if we ground at a WS doesn't that make us WS comp again?

mc: we must be careful

srt: I think it is back to what Tony said earlier -- a minimum chor is two WSs and an observable message betw them

srt: if we ground this in WSDL, then we would need two WSDL based WS's atthe bottom

mc: but what about below that - a WS and a client interacting w/that -- is that a chor?

srt: could be - but if we ground at WSDL, they must both be WSDL

jimh: summarizes where we are -- we have an observable message, and that is the important thing

mc/srt: so what we're talking about is whether we're grounding at WSs per se (WSDL) or would a "dumb" client count

ISSUE TO BE CAPTURED: where do these ground - we seem to have captured the idea of message in the middle

(logged in bugzilla as http://www.w3.org/Bugs/Public/show_bug.cgi?id=276)

mike: <slippery-trout> do chor's ground on messages or on "message exchange patterns"

mc: that is something we would need to work out

(some discussion of this continues)

frankM: brings up "rendezvous" issues (i.e. even one message has issues) - so wherever we ground we'll have issues of what the minimum entity is

fm: so whether a MEP or message, we'll need to figure out where our "atomicity" ends

srt: when we talk chor - we have to be careful to use the term "Web Service" Choreography -- MEPs may or may not be WS choreographies... WSDL may be required to make it Web Service per se, and we need to own the term "WEB SERVICE CHOREOGRAPHY" and not just arbitrary choreography

david: this seems to depend on what a WS is defined as (scribe misses detail - concerns WSDL v. messaging)

david: but agree w/you that the issue is important

 

 

 

Any other Business

 

srt: suggests we suspend telecons for August

mc: TFs could use the time

mc: we should have them or not

others: informal didn't work

suggested that Hugo would point out that taking a month off when we're already months behind is a problem

MC: suggests that we've done a lot - we could use the time to step back, get some stuff moving more strongly, and then get in shape for f2f.

mailing list to continue - just not telecons (Mc)

srt: makes various observations that are not recorded -- take the time off suggested

PROPOSED: to have no calls during August *dates to be announced to WG*

RESOLVED: to have no telecons during August

 

New ACTION: SRT to inform WG of schedule for dates we will not be meeting

 

The chairs brought the meeting to a close.

 

 

Summary of New Action Items

 

ACTION: SRT to contact Alistair on the progress of his action.

recorded in http://www.w3.org/2003/07/15-ws-chor-irc#T20-19-26

 

ACTION: Chairs to accept attachmate offer, work out logistics [10]

recorded in http://www.w3.org/2003/07/15-ws-chor-irc#T20-27-13

 

ACTION: SRT to inform WG of schedule for dates we will not be meeting [15]

 recorded in http://www.w3.org/2003/07/15-ws-chor-irc#T21-31-24