Minutes WS Choreography WG conference call 23 August 2005

See also: IRC log: http://www.w3.org/2005/08/23-ws-chor-irc


1. Attendees:

Martin, Monica, Gary, Greg, Nick, Yves 

2. Scribe

Martin

3. agenda: 

http://lists.w3.org/Archives/Member/member-ws-chor/2005Aug/att-0029/agenda_20050823-0.txt 

Agenda additions:

Martin proposal on modified text:  http://lists.w3.org/Archives/Member/member-ws-chor/2005Aug/0023.html 

More comments from Steve: http://lists.w3.org/Archives/Public/public-ws-chor/2005Aug/0021.html 

Charlton/Monica RFC 2119 mods: http://lists.w3.org/Archives/Member/member-ws-chor/2005Aug/0031.html 

amended agenda is approved 

4. Approve minutes 

16th August:
http://lists.w3.org/Archives/Member/member-ws-chor/2005Aug/att-0022/minutes_20050816_-_0.txt 

18th August: 
http://lists.w3.org/Archives/Member/member-ws-chor/2005Aug/att-0028/minutes_20050818_-_0.txt 


5. Issues

current editors copy at at: 
http://lists.w3.org/Archives/Member/member-ws-chor-editors/2005Aug/0040.html 

Resuming issues from: 
http://lists.w3.org/Archives/Member/member-ws-chor/2005Aug/0020.html 

Gary's comment: 


section 5.3.1 getchoreostatus 

change: 

. If choreoInstanceId is not specified, then there MUST only be a single 
instance of that performed Choreography in the scope of the current Choreography 
that is checking the status. 
to: 

. If choreoInstanceId is not specified, then there MUST only be a single 
instance of that performed Choreography in the scope of the current Choreography 
that is checking the status, otherwise a status with 'ambiguous' MUST be 
returned. 

agreed 

ACTION: Editors to update 5.3.1 getchoreostatus with wording agreed 8/23

monicas Comment #3,

Section 1.6: Timing assuptions seems out of place here. Suggest this be placed later 
around Section 5 where the CDL functions on time are specified. 

Discussed whether to move section.
Agreed leave as is.

monicas comment #5 - suggested adding to primer 
What can you do with roles that do not have interfaces?

ACTION: create an issue for primer to cover roles without interfaces.

monica comment #1 

already agreed to use participant , done as part of last weeks revisions 

monica comment 4 

No solution in place if propogated excetionnot caught 
<nick> A Choreography in an Enabled State completes unsuccessfully when an 
Exception is caused in the Choreography and its Exception Block, if present, is 
enabled. This causes the Choreography to enter the Unsuccessfully Completed 
State. 

change: 
The unsuccessfully completed Choreography enters the Closed State once the 
Exception Block, if present, is completed. If the Exception Block is not 
present, the Choreography implicitly enters the Closed State and the Exception 
occurred is propagated to the enclosing Choreography. 

to: 

The unsuccessfully completed Choreography enters the Closed State once the 
Exception Block, if present, is completed. If the Exception Block is not 
present, the Choreography implicitly enters the Closed State and the Exception 
occurred is propagated to the enclosing Choreography, if an enclosing 
choreography exists. 

Agreed 

ACTION: Editors to update section 5.8 with text agreed 8/23

monica issue 10
section 6.2.2 

Suggest some rewriting to enable understanding: 
 Change from: "The /time?to?complete/ timeout, identifying 
 the timeframe within which an Interaction MUST complete, occurs 
after the 
Interaction was initiated but before it completed" 

 Change to: "The /time-to-complete/ timeout MUST occur after the Interaction 
was initiated but before it completed. The time-to-complete attribute and the 
optional timeout element are defined in Section 6.2.3. " 

suggestion: 
The /time-to-complete/ timeout MUST occur 
after the Interaction was initiated but before it completed and the 
time-to-complete has passed. The 
time-to-complete attribute and the optional timeout element are defined in 
Section 6.2.3. 
<m2> If an Interaction has been initiated but is not completed and the 
time-to-complete has been exceeded, a timeout MUST occur. ... 

change: 
An Interaction completes abnormally when: 
to: 
An Interaction MUST complete abnormally when: 

Agreed

change: 
·The time-to-complete timeout, identifying the timeframe within which an 
Interaction MUST complete, occurs after the Interaction was initiated but before 
it completed 
to: 
 a time-to-complete timeout occurs after the interaction was initiated but 
before it completed 

Agreed 

also change: 
An Interaction completes normally when its message exchange(s) complete 
successfully. 
to: 
An Interaction MUST complete normally when its message exchange(s) complete 
successfully. 

Agreed 

ACTION: Editors to update section 6.2.2. with text agreed 8/23

monica issue 15 

section 8.1 upcase the two MUST 

ACTION: editors to upercase the musts in 8.1 as agreed 8/23

Gary noted that there is some problems with the example in 6.2.3:
http://lists.w3.org/Archives/Member/member-ws-chor/2005Aug/0015.html

Agreed to fix

ACTION: editors to fix example in 6.2.3. as agreed 8/23

Martin proposal on modified text:  http://lists.w3.org/Archives/Member/member-ws-chor/2005Aug/0023.html 

section 2 change 1st bullet to: 
participantType, roleType, and relationshipType - Within a 
Choreography, information is always exchanged between participants 
within or across trust boundaries. All interactions occur between roles 
being exhibited by two participants, and are constrained by a 
relationship. Within WS-CDL, participants are abstractly modelled as 
participantTypes, roles by roleTypes, and relationships by 
relationshipTypes. A participantType groups together those parts of the 
observable behavior that must be implemented by the same logical entity 
or abstract organization. A logical entity or organization may be 
represented by more than one participantType within a Choreography. A 
RoleType enumerates potential observable behavior a participantType can 
exhibit in order to collaborate with other participants. A 
RelationshipType identifies the mutual commitments that must be made 
between two participants for them to collaborate successfully. 

nick proposed:

participantType, roleType, and relationshipType - Within a Choreography, 
information is always exchanged between participants within or across trust 
boundaries. All Interactions occur between roles being exhibited by two 
participants, and are constrained by a relationship. Within WS-CDL, participants 
are abstractly modelled as participantTypes, roles by roleTypes, and 
relationships by relationshipTypes. A participantType groups together those 
parts of the observable behavior that MUST be implemented by the same logical entity 
or abstract organization. A logical entity or organization MAY be represented by 
more than one participantType within a Choreography. A roleType enumerates 
potential observable behavior a participantType MUST exhibit in order to 
collaborate with other participants. A relationshipType identifies the mutual 
commitments that MUST be made between two participants for them to collaborate 
successfully.

Gary doesn't like the restriction that collaboration must be between two distinct participants, why restrict?
Martin says cant see the point why one would model observable behaviour between the same participant as that is an internal thing.
No decision today, people to rethink the wording.

6. AOB

Agreed to have an extra meeting on thursday 25th
2:30pm-3:30pm Eastern / 18:30-19:30 UTC/7:30 UK


7. Summary of Action Items

ACTION: Editors to update 5.3.1 getchoreostatus with wording agreed 8/23

ACTION: create an issue for primer to cover roles without interfaces.

ACTION: Editors to update section 5.8 with text agreed 8/23

ACTION: Editors to update section 6.2.2. with text agreed 8/23

ACTION: editors to upercase the musts in 8.1 as agreed 8/23

ACTION: editors to fix example in 6.2.3. as agreed 8/23