minutes

[Jacques Durand <JDurand@fsw.fujitsu.com>]

Title: minutes

All:

Minutes from last meeting.

Monica will facilitate the next meeting (Monday 4), as I'll be absent.

Regards,

jacques
<<IIC_July_28_03_Minutes.txt>>

IIC Conference Call Minutes: Monday, July 28, 2003
 
Attendance:


Monica Martin (Sun)
Pete Wenzel (SeeBeyond)
Steve Yung (Sun)
Serm Kulvatunyou (NIST)
Jacques Durand (Fujitsu)

excused: Mike Kass (NIST)

Minutes taker: Jacques Durand

Time: Monday July 28, 11am PT
Host: Fujitsu 
Toll Free - : 1-800-251-6413 
Toll - : 1-913-905-1400 
Participant code: 598136 

Agenda: 

1. Status of MS conformance test suite spec. (Mike Kass editor) 
- review current draft of MS conf test suite, and specific issues:
. Issue #1: the problem of "app-level" test cases: do we have a list?.
. Issue #2: ordering / partitioning based on config/CPA? (how to avoid too many reconfigs)
. Issue #3: conformance profiles or levels (detailed content?)(please read last minutes, we'll decide)
. Issue #4: various editorial issues.
. Issue #5: message correlation details.
. Issue #6: check / fix parameter tables.

2. BPSS test update (Serm K., Monica M.): 
- spreadheet from Serm / Monica, BSI testing definition.
- what cooperation/feedback with/to the BPSS team 
(emphasis on testing operations).
- idea of prototype.

3. Implementers corner, and PR: 
- Drake Certivo test driver status. 
- KwareSoft question on ErrorURLNotify 
- presence at AutoTech: who demoes, IIC visibility... 
- IIC members demo (Test Framework, test suites) at XML 2003, 7-12 December
 
 
-----

1. Status of MS conformance test suite spec. (Mike Kass editor) 

Issue #1: the problem of "app-level" test cases. Mike Kass has already listed
all the test reqs he believes are questionable as whether the  Test Framework
has enough control in order to recreate the conditions for their verification
(see "test coverage" column in req table, 3.2.2, red-marked lines.)

[Action Item]: go through this list and reassess them. Pete W. (and Jacques) are
challenging some, like Req #50: the "manifest" values are actually set by the MSH,
as the app only decides of the payloads. So in that case, the test case is OK.
For others, the ability of the test framework to execute the test case is "contingent",
i.e. the feature to test may manifest depending on the implementation.
(e.g. the option to use straight SOAP without attacht, for messages with a single payload,
cannot be tested if impl decides not to.) 

Issue #2: ordering / partitioning based on config/CPA.
Pete Wenzel asks if using a different URL for each CPA could reduce the
"config change" issue Mike Dillon mentioned. Unlikely... (depend on impl ability)

[Action Item]: go through current list and see where the discontinuities are, for
CPA refs. Jacques suggest each test case header could more explicitly point
to the CPA it uses ("test case" lines of the abstract test suite), instead of
just relying on the CPAid burried inside the message expression of each step.

Issue #3: conformance profiles or levels. Agreement that :
Profile #1 covers "core" features,(security, errors, SOAP env)
Profile #2: #1 + covers mostly reliability,
Profile #3: "full" conformance (all spec, icluding supprot for ping, 
status request, multi-hop). Seems odd as a profile, as it is the entire spec.
(should not be called a profile)
Question on whether SyncReply modes other than mshSignalsOnly should be
excluded from Profile #1 (which is the minimal set of mandatory features),
based on possible evolution of ebMS spec. (Shifting some features from Core to Additional)

[Action Item]: Jacques will ask the ebMS TC to clarify their position.

Issue #4: various editorial issues. Will wait for editor.
List of authors should not be so large... many are contributors.
I would propose that authors/editors are: Mike Kass of course,
and Matt McKenzie as he provided the initial core test req material.
All others are mostly Reviewers/contributors. (Opinion?)

Issue #5: message correlation details. 
Jacques and Mike to get closure on the message correlation issue, by email.

Issue #6: check / fix parameter tables. This is related to the correlation
discussion. We seem to have an agreement on wht to do.


2. BPSS test update (Serm K., Monica M.): 

- Monica has sent questions to BPSS team. ebXML dev list had also some issues
on BSI definition... (rules for bus party interface?)
- Monica will send out a proposal.
- Serm : BSI testing: needs to clarify what support in test suite, test framework 
for transaction state management, req/response activities in BPSS.
- We ned to provide comments (to Serm/Monica) on the BPSS spreadsheet (test suite).

3. Other business

- RegRep TC meeting: vote on 3.0.
- The RR TC is asking whether IIC has any plan for RR conformance. We decide that
as a  first step, the  TC could provide/prioritize a set of requirements.
We'd help formulate these in a "Test Framework"-compatible way,
then expect RR TC to complete the set of test requirements, that we would work
from to create a test suite.
- Monica: will send out ebXML-lite proposal from Covisnt. COncerns that is might not
be just an ebXML profile, but may be non-conforming with the spec, e.g. if ack
mechanism relies solely on sync responses at transport level.