TPTP-PMC-20091028

Logistics

Attending: Kathy, Oliver, Chris, Paul, Eugene

• Ernest on vacation in Orlando

Any objections to last weeks summary

• No

4.5.2.1 Discussion

• Kathy notes that M4B is posted to consumers and we will get feedback soon

4.6.2 Discussion

We are toward end of a development iteration and are entering a test pass next week.

• The current schedule has a week of smoke and two weeks of full test pass
• Kathy suggests that instead of doing the big test pass now, just do lighter test now
  • Motivation because the number of defects fixed is not big enough to justify the larger pass
• Kathy has updated the schedule to reflect this.

Iteration 1 recap

• 31 bugs fixed.
• A number of these were backported to 4.5 where they have been tested by IBM already
• Oliver asks if 31 bugs in 5 weeks is expected thruput
  • Kathy thinks that thruput is as expected
  • Kathy said that on her side people were splitting time during this part of the project. Expecting more over time.
  • Paul noted on test he is doing most of his part at end of cycle. Jerome was working on one big defect
  • Paul thinks thruput is as good as expected

Oliver asks what investment was over this period

• Test project has ~.75 resource
• Platform project IBM has ~2.5 (including build/etc)
• Platform project Intel has ~1.5 engineering plus part of person to assist with build

Paul asks if we should start putting crude estimate of time worked on defects.

• Test project does this pretty regularly
• Kathy notes that there is time spent going into reproducing defects as well as fixing them.
• Chris asked about some non-bugfix related items (knowledge transfer underway was one example)
  • Team noted that knowledge transfer does have a bugzilla associated with it.
  • Chengrui has a wiki page under development for sharing between IBM and Intel
  • He can fill in in the time field (updating the previous value)

The part that might not be accounted quite right are test, build/release engineering

• Mailing/new list traffic is low right now so that should not be a big invest.

Kathy asks if there are intelligent queries to turn some of this into trendlines

• Paul notes that there can be a running tally from week to week.

Oliver says that this metric could be useful as we ask for resources next year

• He does not want to add unnecessary tracking tax so it should be lightweight
• Paul would like to see us using the resources we have efficiently
• Kathy asks if this should be forward or should we go back to "all" 4.6.2 defects.
  • General agreement to have it going forward
  • Oliver says don't go backwards but to start at iteration 2.

Team discussed some project that are even more rigorous and have a model of "if it is not tracked in bugzilla, you did not do it."

Chris notes that field wont really work as a personal timesheet because different people work on the same bugzillas (reviewing, etc).

• We reiterated that we're really tracking the project forward progress rather than individual contribution.
• We will never see 100% of engineers time tracked
• It might however be useful for helping get gut feel for whether a 10% or 20% resource is effective or not.

Misc

Do we need to meet next week.

• Yes

Itanium bulid system update

• IBM got their unpacking authorization and are getting it set up.
• Expect about a month to get them all set up and allow decommissioning Intel side builds.

We discussed vacation schedule for the upcoming holiday seasons.