TPTP-PMC-20080130

Logistics

Attending: Paul, Oliver, AlexA, Harm, AlexN, Joanna, Chris

Does anyone need to leave early? No

Any updates to minutes from last time? No

4.4.1 Release

Paul has been working with Joel to automate test reports (he has manually generated todays). Should be more automatically updated soon. Paul will need to confirm if this is just for 4.4.1 or also for 4.5.

Oliver asks: Are finishing it[testpass] up? Test

• completed lot of testing. plan to finish by end of week
• it is on schedule

Monitoring

• Targeting finishing by end of day Friday

Platform

• Dont think will finish by Friday but next week (lots of tests)

Trace

• Mail threads w/ IBM team responsible for BTM.NET.
• All but BTM.NET plan to finish by next friday
• They had not planned to test BTM.NET since will be providing same binary as previous
  • AlexA notes it is separate tech preview standalone binary
  • Paul indicates that there needs to be at least some packaging testing
  • Someone aske if it is actually rebuilt or is it the same identical binary.
    • Paul does note on download page that this package is not rebuilt. It is separately built.
    • AlexN notes it is built regularly for the 4.5 stream but is not built for 4.4.1 stream
    • Joanna confirms that what is on site is actually from 4.4.0
  • Harm agrees that it doesn't need testing with visual studio however it does produce data that needs to be consumed by the rest of TPTP.
    • At least some smoke tests to make sure the results can still be consumed by 4.4.1.
    • Harm indicated that AlexA asked right questions to the team but didn't get a good answer.  :-)
    • AI to Oliver to touch base with Matt Mings on "BTM.Net tech preview for 4.4.1" to make sure they smoke test at some level.

Profiling in 4.4.1

Request for status update about profiler and JVM JVM issue that Eugene, AlexA, Asaf were working on.

• AlexA indicates for most cases, issue resolve with patch.
  • There are a few use cases that did not get covered.
  • Eugene told AlexA that he was okay with this solution
  • AI: Oliver will follow up with Eugene to confirm that everyone is pleased with the outcome of the emergency fix.

Joanna notes that 2 POG defects opened against 4.4.1 are still listed as open AI: Oliver will discuss this when he follows up with Eugene and copy AlexA

Oliver thanks Alex for being good and concise in his report

4.5 Release

Oliver asks: Are we on track for meeting our freeze date for I5

Test

• Large chunk that did not get completed but does not indiate serious concern
• Attaching some details to assorted bugzillas
• Nothing to prevent them from moving forward from releasing a successful I5 even in the presense of some not completed defects

Platform

• Working on I5 defects in parallel with 4.4.1 test pass
• There are a few more I5 defects trying to squeeze in

Monitoring

• Same situation as Platform.
• A few I5 defects trying to finish up.
  • A few not containable will be moved to I6.
  • But a few will be finished in next day or so.

Trace

• Still have critical defects that are blocked on some stuff that AlexN/Joannas folks are finishing
  • Editorial note... I was a bit unclear if it is blocked on AlexN or Joanna. It was however being actively worked on.

Oliver asks if need a new freeze date

• Ganymede I5 will be on the Eclipsecon memory stick
• Absolute Final Deadline Feb 18th (Monday)
• We plan to finish Feb 15th. If we are lucky and don't have regressions
  • Working plan: Whatever we have on the 15th we ship.
  • That is ~2weeks from today.
• When do we freeze and seriously go into test pass?

Oliver asks: can everyone get I5 stuff done by this Friday (containing what we can and move rest to I6?)

• Concept: Freeze Friday COB and go into PMC approval process at that point.
  • Candidate driver on 4th with all expected stuff.
  • Each lead needs to decide if this is acceptable.
    • Leads did not greatly disagree with this plan
    • Trace noted that 2 critical issues will need to move to I6 but otherwise agrees that the freeze proposal is acceptable.

Paul suggest that with all the stuff going in at last minute (Friday) the build for the 4th may have a bunch of stuff not yet used together which increases risk.

• We discussed this
  • Test only has a bit more stuff
  • Monitoring only a bit more stuff
  • Trace only has a bit more stuff but has to wait on AlexN/Joanna stuff before he can finish
  • Joanna is in triage mode moving critical items only
• So thought was that there will not be THAT much stuff hitting on Friday to cause too much concern.

The plan is to finish all testing by the 13th (Wed) PMC for 15th release.

• This gives us only 2 day leeway

Blocker and Critical bugs to move I5 to I6

Joanna asked for quick query/discussion on blockers/major

• Test
  • Opened critical defect last night.
• Monitoring
  • Has 1 blocking & 1 major.
  • Feels they could be moved to I6 without severe problems
• Trace
  • 2 critical issues will need to move to I6
• Platform
  • 3 blockers in model (moratoriam in place of some sort (translation deadline?))
  • 2 in AC
  • 24 criticals (joanna thinks some can be deferred)

Harm moved us off to discussion deferral process for moving defects to I6.

• If anything is tagged P1 process demands that leads request a vote to move to I6.
  • Joanna asked if this had been done previously.
  • Paul asks about some logistics.
  • AI to leads: if defect is P1 and targeted to I5 and one wants to move it to I6 should vote.
• Oliver notes that by doing it on the list it avoids firefight later when someone complains that their item was silently moved.

The team had a side discussion [AGAIN] about the difference between the plan keyword and the priority setting. Rather than recapturing the details yet again, I provide a pointer to previous minutes where we outlined the process.

• See TPTP-PMC-20071031#4.5_Bugzilla_cleanup
• See TPTP-PMC-20071121#Bugzilla_and_plan_queries_for_4.5

Misc

Oliver got internal review on his document that he mentioned last week

• Will get out to PMC attendees later today to comment/review

Oliver spoke a bit about netbeans/eclipse comparison.

• Oliver wants to report out to Ecilpse board/leadership but
  • Wants to avoid large companies getting nervous about such a comparison
  • Needs to make sure that discussion properly framed for Eclipse discussion
  • Oliver will be touching base with Mike Milenkovich about best way to frame this.
  • Harm suspects mike will agree that presenting to board is good and then letting them discuss what to do next.

• Oliver's engineer is making progress
  • Working with AlexA to get good solid comparison scenarios for OC systems compare to tptp/netbeansprofiler
  • Harm suggests it will be good if we can relate this to our POG scenarios if we can demonstrate/state ways that TPTP profiler actually does things that netbeans cannot do.
  • That list by itself would be of value (i.e., lead with positives rather than concentrating totally on areas we may have some gaps
    • Harm gave a little speach... :-) ... People keep focusing on the framework for eclipse and framework for netbeans and not necessarily the value add that Eclipse brings (e.g., value plugins). Test for example last time harm notices, netbeans had hardly anything in that space. If board likes the profiler comparison maybe looking at some other areas (like test) may be of value. Eclipse may find a few folks to go and do some more serious comparisons.

Any other agenda items?

Paul raised small request about changing ownership of 5 components from JeromeG france to JeromeB france.

• Jerome moved off, Jerome moved on.
• Nobody raised any objections

Weekly Milestone Tracking Process

Joanna raised the question about the weekly milestone tracking process

• Joanna wonders how much value there is in the process.
  • e.g., not finding too much value in weekly schedules.
  • no real consequence if something slips.
  • Harm notes that he looks at this weekly.

Harm noted that in the fall everyone agreed to try weekly schedules for a few iterations. We are now starting to get some feedback at cost/benefit. This feedback is good.

• We had a side discussion about the use of the "slipped" keyword used in weekly schedule.
  • If an item slips in weekly schedule, leave it in original slot in weekly table and add word slipped. also add the item to the new target date. In this way can monitor how far items slipped (as opposed to simply moving the line item).

Oliver notes that if anything is not useful, raising the quesiton is valuable

• Notes that Joanna should be the prime consumer of the weekly table [for platform]
• Allows lead to see overall shape of what is slipping and how far
• Shouldn't be taking much time to keep up to date (~20 minutes perhaps)
  • If truely not helping Joanna she and Oliver should talk about possible ways to reformulate
  • Oliver is open to alternative mechanisms

Oliver gave a little speach on the motivation for weekly schedules.

• All items intended to be done in appear in weekly schedule.
• Leads being honest about what is and isn't on track.
• The weekly milestone page is central repository to show for all projects sidebyside.

Oliver asks if there are any question about the rules/format?

• The page track wed to wed (e.g., pmc to pmc)
• paul notes a concern is that we are putting effort into tracking our progress but doesn't seem to be fallout from missing that target. What is there to do.
  • Harm noted a side discussion he had with Paul about slipping test bugzillas that he noted from the weekly shedule.

To make the weekly milestone more central to PMC dissusions:

• For future pmc meetings.
• AI: to each lead to speak to their weekly schedule column. (current, slip, next steps).

Oliver asks how much time does it take people to update weekly schedule.

• Joanna close to an hour between this an meeting minutes
• AlexN ~same

Oliver asks if this is the same time as previous scheduling technique or additional time.

• Joanna sugguests she may have to change how she approaches it (merge her process w/ this one) as there is some redundancy.
  • Joanna may point her folks at weekly milestones and use it as part of her process

Oliver notes that leads can delegate to appropriate team members do core updates to weekly schedule and then review it before meetings to reduce overhead on lead.