Choreography component meeting agenda (31st Jan):

1. Discuss any objections or queries about the design of enhancement 50076 in so far as it relates to the choreography component

2. Discuss current BPEL specification progress and any issues known to affect TPTP
	(BPEL Issues List) http://www.choreology.com/external/WS_BPEL_issues_list.html

	- Issue 4 - Dynamic Parallel Processing
	- Issue 147 - Serial and Parallel Foreach

	- Issue 63 - Support of Array

	- Issue 103 - Standardizing $varName syntax for XPath to refer to a BPEL variable
	- Issue 29 - Simplification of XPATH expressions

3. Discuss plans for component refactoring and package structure (see engine architecture document emailed out on TPTP platform dev mailing list)

	

Attending:
	Antony Miguel (Scapa)
	Koustubh Parwar (CA)
	George Din (Fokus)
	Diana Vega (Fokus)
	Joe Toomey (IBM)
	

Absent:	
	Harm Sluiman (IBM)
	Serge Lucio (IBM)
	Jim Saliba (CA)
	Kent Siefkes (IBM)


Point 1 discussion:

Based on the discussion with Harm at the Model component call last week it seems clear that it is important to support the equivalent of WSDL bindings in the test model.  Performing this mapping is a fairly sizeable job and will probably require more time than available in the choreography component call each week.  Joe and Antony will try to make progress on this issue outside the call.


Point 2 discussion:

The BPEL specification is being developed by member companies of Oasis.org.  TPTP has a few requirements on this specification which we need to be present in the next ratified BPEL specification.  According to one of the members of the BPEL Technical Committee, the public mailing lists associated with the BPEL Technical Committee are mostly ignored because of spam problems.  Therefore, in order to get any feedback on the issues important to TPTP we need a representative on the BPEL Technical Committee.

	Koustubh will check with Jim as to whether it will be possible for him to join
	Joe is able to join but has serious doubts whether he will have the time to follow the committee


Point 3 discussion:

The component code should be checked into TPTP CVS by the end of today (monday) or tomorrow at the latest.  The code is unlikely to work properly and have the correct dependencies etc until the end of the week.


Other discussion:

If the engine is used for stress tests and is potentially simulating large numbers of users (threads), how will the engine deal with issues of endpoint availability?  The engine currently only has two working bindings, the engine internal binding and the WSIF Java binding.  Neither of these bindings suffer from availability issues as a new port instance is considered always available (given that it is limited only by the resources available to the JVM).  If a new port instance is needed for either of these bindings then it is simply created.  In the case of the upcoming SOAP binding, it is not clear whether this will be an issue.

Will switching from using a Java binding to using a SOAP binding require changes to the source BPEL?  If a given port type has two bindings available and a set of invocations in a BPEL source file use that port type then the bindings can be used interchangably without changes to the source BPEL.  If a particular binding (e.g. SOAP) is being used in say a stress test to exercise a web service and the test developer wishes to use a different binding (e.g. Java) then it may be the case that it is not possible to use the other binding without changes to the interface, in which case the BPEL source would have to change to use the new interface.




Actions:

- Antony and Joe will schedule some time to go over WSDL bindings and discuss enhancement 50076

- Koustubh will check with Jim about joining the BPEL technical committee