| [hyades-dev] Meeting minutes for 5/6 Hyades Execution Team Meeting |
|
I've attached the meeting notes for last week's meeting. I apologize for the delay.
Kent Siefkes
|
Execution Team Group Weekly Meeting, 11AM Eastern, Tuesday, May 6, 2003
Attending
---------
Kent Siefkes, IBM Rational Raleigh
Anthony Miguel, Scapa
Kim Coleman, IBM Rational Cupertino
Kris Kobylinski, IBM Toronto
Victor Haven, IBM Rational Cupertino
Richard Duggan, IBM Toronto
Mike Silverstein, Silvermark
Dwayne Dreakford, IBM Rational Raleigh
Proposed Agenda
-------------------------
1. Get status on coding progress from Kris, Dwayne, and Antony (with input on Richard's side as well)
- any unresolved issues on Dwayne's list of questions?
- discuss any blocking issues and how to resolve them.
- what's missing to get simple end-to-end use case ("hello world" running on remote machine)
2. Review remote execution diagram that Richard put together that came out of our additional meetings last week
- confirm ownership of boxes
- confirm process boundaries on remote side
- factory for creating ExecutionComponents
3. Make/confirm decision on immediate vs. deferred remote communication. Need owner for extending use cases
based on that decision.
4. Discussion transition (Kris)
5. Review UC4 [if time, have Kris provide latest UC4 realization by email if not]
Recorded Discussion/Decisons
----------------------------
1. Get status on coding progress from Kris, Dwayne, and Antony (with input on Richard's side as well)
* Dwayne and Richard to meet again at 2pm today to push through until have something working, e.g. simple JUnit
test case
* Generic logging / debugging / error handling - Richard will discuss with Dwayne later today
- 2 logging packages already dropped (loggings.common - Apache common logger, loggin.java14 - JSR 47 based)
* Kris - Has completed remote common Junit/Manual/http test case launcher built on test facade, logging Agent, specialized package for
JUnit test (relies on JUnit jar from workbench). For logging it is still using comptest record since execution events not
defined yet, and still using old component test names - not yet changed to Hyades names. Still needs to
be plugged into Session, and driven from workbench side across network.
2. Review remote execution diagram that Richard put together that came out of our additional meetings last week
* Antony owns Sessions, Dwayne owns stubs and skeletons, Richard's teams owns infrastructure below stubs and skeletons.
3. Make/confirm decision on immediate vs. deferred remote communication.
* Confirmed that immediate communication to remote side is done at the time of each workstation ExecutionComponent action (rather than defer
until launch). This simplifies the implementation and should ease vendor adoption, and is consistent with recent design decisions made
this past week.
* deferred actual update of use case realizations to reflect this until at least next week when Joe returns
4. Discussion transition (Kris)
* Kris returns 5/28
* Component Test resource in Kris's absence would be Marcelo
* JUnit is what is "done", Manual & HTTP needs to extract commonality from JUnit implementation (note that HTTP uses Java executable).
* Kris will send out all his code to group, and instructions, and what's left to do
5. Review UC4 use case realizations [deferred]
* Kris noted there has been no changes to the last UC realization he submitted
New Action Items
----------------
* None, other than Kris and Richard's meetings May 6 afternoon.