Minutes WS Choreography WG conference call 8 February 2005
Issue categories: (Editorial, Closed Won't Fix and Requires Tech Discussion)
Agenda: http://lists.w3.org/Archives/Public/public-ws-chor/2005Feb/att-0001/Agenda.txt
1. Role Call
Roll call: Martin C, SRT, YL, Abbie, Anders, Gary, Monica, Nick, Charlton, Jeff
2. Appointment of scribe
Charlton volunteered to scribe`
Group expresses thanks to Greg for providing the issues list for our triage
3. Agenda Changes
None
4. Approve minutes from 25th Jan
http://lists.w3.org/Archives/Member/member-ws-chor/2005Jan/att-0010/2005-01-25_WS-Chor_Notes.txt
APPROVED
5. Action item review
1. ACTION: Chairs to make exit and implementation criteria proposal for the group.
IN PROGRESS.
ACTION: Chairs to present implementation criteria proposal to the group in
the next conf call. Supercedes ACTION ITEM
2. ACTION: Martin to do UML diagram from scratch for CDL
TO BE DONE.
3. ACTION: Chairs to ensure proper documentation and tracing of decisions for resolved issues,
and to inform the submitters.
IN PROGRESS.
4: ACTION: Martin to add sub categories to last call comments in Bugzilla, such as editorial
IN PROGRESS
5: ACTION: Steve to collect up the issues requiring discussion.
DONE
6. F2F Planning
Yves: all information on admin page, including dates (Feb 28 & Mar 01)
F2F Planing: 28 Feb 2005 - 01 Mar 2005
Face to face meeting, Boston, MA, USA - W3C Technical Plenary
Registration page and registration for hotel
http://www.w3.org/2002/09/wbs/35125/TP2005/ (registration page)
http://www.w3.org/2004/12/allgroupoverview.html#Venue (hotel registration)
Yves: there is a limit for meeting announcement, not for agenda announcement
Yves: even if it is better to know in advance
MC: Only two meetings between now and the F2F
MC: Need draft agenda for F2F to circulate on next week's call
NEW ACTION: Chairs to draft agenda for F2F and circulate for next weeks call
7. Issues triage
Issue 1062
"The current text in the choreography lifeline section 2.4.7 does not describe
whether the 'complete' could cause the premature completion of a finalizer."
Nick: Intention - completion condition for normal activities
Gary: Sounds reasonable - thus an EDITORIAL
Issue 1062 EDITORIAL
Issue 1063 EDITORIAL
Issue 1064 Not fully resolved the WSDL dilemma about outbound operations
monica: not an editorial
charlton: i think it is an editorial
Nick: Spec is now relaxed to allow inbound and outbound
charlton: WSDL 1.1 - we still have restrictions; we relaxed restrictions on WSDL 2.0
Gary: JJ implying outbound op on one roles i/f c/b mapped to inbound op on another roles i/f
Gary: CDL supports any MEP operation (WSDL 2.0)
SRT: Mark this as editorial
Yves: Let's ask JJ for clarification - what is he looking for in the spec
NEW ACTION: Charlton to discuss with JJ what clarification he is seeking with this issue?
Issue 1064 left open subject to response from JJ
Issue 1066: Section 2.2.5 Semantics, working towards solidifying two items
monica: Need more clarity in the text
SRT: No objections to proposal from m2
EDITORIAL - fixed with text in proposal
Issue 1067: Section 2.4.2 Be specific about when the information is available
monica: Has to do with timeliness and availability
EDITORIAL - fixed with text in proposal
Issue 1066: RESOLVED FIXED based on enclosed proposal
Issue 1067: RESOLVED FIXED based on enclosed proposal
Issue 1068: How are the blocking guard condition semantics would be projected to BPEL?
Nick: std BPEL cannot support this, so it cannot be projected from CDL to BPEL; Microsoft, for example, has extensions for this
SRT: There is not standardized BPEL
Nick: If what we have in 2.0 is close to what w/b finalized, we don't have support for blocking guard semantics
SRT: may or may not be an issue - our stated goal is t/b amenable to endpoint projection through BPEL
SRT: m/b a primer issue - if doing endpoint projections to BPEL, there are certain things one should not use in CDL
Nick: In real products, one should expect to see support for this sort of feature
Gary: Good idea of BPEL members of CDL group raise this as an issue in BPEL
Nick: Raised a year back - response - control links w/n be extended to handle such constructs; also, extension of event handlers approach to handle this was killed as well
Nick: Attempted to be short on reply - authors of BPEL do not want to pursue this
SRT: This is an issue - close + resolved when fixed: we're not going to change CDL to fit BPEL; reassign to Primer
ASSIGN 1068 TO THE PRIMER
Issue 1074 Language Extensibility and Binding - suggested change to an Extensions rule
Anders: If do mapping, may want to add semantics to be more restrictive that is more restrictive than the specification
Anders: May want to restrict certain parts
Nick: Example?
Anders: It may be the case with variable bindings - we may want to add restrictive semantics regarding when bindings can occur
Nick: We do not want to change CDL semantics
Nick: I'd like to see a simple concrete example to illustrate what you are talking about
SRT: I don't see it as a critical value add
SRT: An example would help clarify this
Anders: Will try to find an example
Yves: Keep this is "pending use case"
OPEN pending use case (REQUIRES TECHNICAL DISCUSSION)
Issue 1075 Channel Types - attribute "usage" should be made an Element
Anders: Could be possible to have communication resources used once and unlimited, by making this an element, can provide more detailed semantics
SRT: usage once - linear typing - unlimited - any number
Anders: Similar to time to perform - resource c/b available as long as variables not out of scope...
SRT: Propose marking this issue as requiring discussion
REQUIRES TECHNICAL DISCUSSION
Issue 1076 Information Exchange Capturing Variables Definition
Anders: Most likely editorial - no semantic/defn changes
Anders: Recording part of interaction s/b more clear on event semantics
Anders: S/b more clear as to when variable is changed or made available
monica: This does not seem to be editorial.
no seconding for this as editorial
Nick: how is this not captured in current definition/verbage?
Nick: "Used to populate the content of a message to be sent, before or at the occurence of message-has-been-sent-event" is already captured in the spec - adding language like this will only clarify the language
monica: We should discuss as this does not seem to be editorial
Nick: This is captured in the existing definition, so it seems to be editorial
SRT: Link back for something already described
SRT: Propose - mark this simply as editorial
RESOLVED FIXED based on enclosed proposal
Issue 1077 Add enforcement level to the silent action specification
Anders: MUST may make it difficult semantically to use
Nick: Assume someone created endpoint projection - if there is a hole, this language indicates that you must fill the hole; otherwise, there is a hole, and no one fills it
SRT: If have given it a name in pi-c, you must have something for it
SRT: makes no sense for me to say anything other than MUST here
SRT: only becomes observable when you project the endpoint and fill in that definition
Anders: Is interpretation in form of business semantics or pi-c; if business semantics, the language will present a problem
SRT: Never in the CDL contract will say how this will be done - silent serves as an inline construct to stub out endpoint projection
Anders: Maybe being more explicit will make this clear and end any confusion
SRT: Assign this to the primer - only by illustrating fully via example does this become clear
Anders: Some language in the indicated section (2.5.5) would help clarify this
SRT: Assign this as editorial needing an example and assign to the primer.
ASSIGN TO PRIMER with example
Issue 1078 Marking the Absence of Actions - reword section to better relate to what actions are not performed
SRT: Assign this as editorial needing an example and assign to the primer.
Issue 1079 Interaction Syntax - successul-sending and properly-received are not clearly defined
Anders: Specify - is it before sending, after sending, before recv, after recv?
SRT: Legitimate issue in need of further discussion
SRT: Not editorial
Nick: Let's revisit the definition and discuss after review
REQUIRES TECHNICAL DISCUSSION
Issue 1084 State the relationship between WS-CDL and the UN/CEFACT Modeling Methodology
http://www.unece.org/cefact/umm/umm_index.htm
SRT: Include this in section 1.5 to CDL - editorial
SRT: Important to cite this work
EDITORIAL
8. AOB
Presentation of late:
SRT: Gartner - CDL presentation in LA (20th Apr)
SRT: New York presentation/Java SIG
SRT: Distribution to ensure no single point of failure
SRT: Particular interest to wholesale banking
Nick: In Microsoft, it is rumoured that BPEL is passe, and that they are now working on "C-Omega" (C# derivative + some join-c + some xml handling)
SRT: Sounds like an interesting endpoint language
http://www.xml.com/pub/a/2005/01/12/comega.html
BPEL and Java:
SRT: Is BPEL competing with Java?
SRT: Great tools, good ids with graphical grammars - why put XML in between?
SRT: Does BPEL have business value?
Examples:
Nick: Do we have any status on the example Mayilraj was working on?
Nick: I have something I can donate but it has been a long time since we heard last w.r.t. examples
SRT: I am working on it - and I did ask people to submit examples some weeks ago
9. Meeting adjourned
Summary of Actions
ACTION: Chairs to present implementation criteria proposal to the group in the next conf call
ACTION: Martin to do UML diagram from scratch for CDL
ACTION: Chairs to ensure proper documentation and tracing of decisions for resolved issues,
and to inform the submitters.
ACTION: Martin to add sub categories to last call comments in Bugzilla, such as editorial
NEW ACTION: Chairs to draft agenda for F2F and circulate for next weeks call
NEW ACTION: Charlton to discuss with JJ what clarification he is seeking with this issue?