[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Minutes, and more.
All:
1. Minutes attached, also contain info for the f-2-f at New Orleans.
2. Jeff Eck, our secretary, can no longer participate in IIC due to new assignments.
Thanks to Jeff for his support of several years (Jeff has been of all initial POCs).
We need a new secretary:
Secretarial job is not much load (help with meeting logistics, sporadic Web site maintenance and updates)
Nominations are open, please let me know if you are interested.
At this time we have Mike Kass interested.
We'll close the nominations next week , Wed 28 night, and start an email vote.
Regards,
Jacques
<<IIC_April_19_04_minutes.txt>>
Time: Monday April 19, 2004, 2pm PT Host: Fujitsu Toll - : 1-512-225-3050 pass:89772 Attendees: Michael Kass (NIST) Monica Martin (Sun) Jacques Durand (Fujitsu) Steve Yung (Sun) Pete Wenzel (SeeBeyond) Agenda: 1. Status of Test Framework, review of last comments and Mike response. Submission for vote if no major issue remains. 2. Prioritizing Next work: - BPSS testing, overall approach: conformance? use of a "test process"? - Registry: status and approach. 3. New Orleans update. -------------------------- minutes --------------------------------- 1.a Status of Test Framework, review of last comments - Jacques : After review, the main remaining technical issue is the unclear semantics of test case outcome: definitions of fail/pass/undetermined are unnatural for test case writers. - Monica compared the scripting capability of the test framework 1.1 with what was done in other workflow languages (and with BPSS requirements) and finds that we have unnecessary complexity, would not be easily unerstood by wkflow users/develpers. - Pete is also concerned that the TFk1.1 will be perceived as complex. How much more do we need to add, compared to the "abstract tst cases"? - We noted that Tim (Drake C.) extended scripting to support BPSS, without much additions. - Jacques: the reason could be that we started with an initial clear mapping of test case execution to boolean semantics (for fail / pass), and this cascaded into giving boolean semantics to operators and concepts that initially were purely operational (workflow). This overloading is not intuitive to workflow community. 1.b Decisions: - Test Framework 1.0 was successful: it has been implemented by at least three companies/orgs. We fixed several things in TFk 1.1, but the control flow extensions are now where the complexity is perceived, and if we release it as is, we run the risk of (1) implementors leaving out many features, (2) test case writers dislike of scripting language. - we should not release TFk 1.1, if we have doubts on its acceptance by users. - we will simplify the scripting: keep what is sufficient to script the most common wflow cases, follows the current practice in wflow scripting. If we lack some advanced feature, we can always add them in a future release. - objectives are: workflow users should easily understand the scripting. - we will be driven by a small set of representative use cases, that we will try to script in the most natural way. - Monica will send out use cases from IV&I and also Electric Utility test cases. - Jacques will select some PIPs use cases. - We will script these, and refine the script design based on these examples. - these use cases, and their final scripting (after simplification) will be added as examples in the final doc. - we'll make sure that potential users like the test cases. 2. New Orleans update: - We have a meeting scheduled Thursday 3pm-5pm. We may take only 1 hour of this window, to be decided depending on constraints for attendees (who also are members of other groups.) - No call-in. - We'll have informal follow-up meeting Thursday night. Agenda: - Edit / spruce up Use cases for choreography testing. - Discuss scripting options, consider updates in script language. - More general discussion on how can we better involve users/implementors in our design work (are "use cases" a good support for this?) 3. Others. - no meeting Monday 26.
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]