Difference between revisions of "TPTP-PMC-20081105"

Latest revision as of 18:45, 6 November 2008

Logistics

Attending: Oliver, Paul, Chris, Eugene, "Ben"

Any issues with last week's summary?

• no

4.5.2 Discussion

Oliver asks for a 4.5.2 status update

• Paul notes that the test pass shows really good results
  • >99% of results checked in
  • Wow! Kudos to team for follow thru on the test pass.
• Test project has a few tests that want to rerun
  • Unfortunately we currently have some build issues so do not have a build
  • Team is actively working on the build issue now

Oliver asks if everyone understands what they need to do for I2 and has plenty of defects to work on.

• No disagreement

Note to everyone: The next test pass starts November 26

We tried to discussing integration of the China team. Without Joanna present we could not get full feedback. Chris indicated that the team has been operating on a weekly status mode for several weeks and it appears to be working well.

Profiler Outreach

We have yet to get feedback on our plea to the community. We are expanding our net.

• Eugene will post updated message to committer list
  • Oliver will follow up afterwards to underscore TPTP commitment

New Contributor Opportunities

Last week at EclipseWorld, some JProbe representatives were talking a bit to Oliver about Eclipse and TPTP possibilities.

• Some years ago there was discussion with JProbe about coordinating with TPTP.

Oliver plans to follow up with them and wanted to get input from PMC about some items

• What would benefits of jprobe being part of TPTP open source.

• Is TPTP the right place in Eclipse for integrating parts of JProbe?
• What are some of the JProbe competitors
  • Paul mentioned yourkit in this space.

The PMC discussed some of the opportunities and complexities that would be involved by mixing bits of JProbe and TPTP profiler technologies in Eclipse.

• It is hard to estimate the overhead that would be involved.
• There are few resources in TPTP today to facilitate some of that collaboration
• "light" co-existance? Unified data model, unified data collectors?
• There are range of possible efforts. Need to understand a bit more about what proposal would be.

Oliver asks if we want to devote some PMC time to it in discussion

• Probably a good idea to invite them to come and discuss more about the idea

Mac

There is a community member and Mac fan that has a few months to invest. Believe could do basic port of AC and data collector to enable TPTP profiler on Mac.

• This type of thing was suggested in the past by different folks and turned down for lack of ongoing commit for support
• With recent IBM platform interest discussion (TPTP-PMC-20081015) we should revisit this.

• Paul reiterates Harm's earlier suggestion that we should consider it.
  • There is a small but steady stream of requests for this platform
• Oliver has concerns about how much committer time it would take to support the community member as they port
  • We noted that we could set some expectations and see how it goes.
• Paul wants to make sure that all correspondence is in open source

Eugene asks if we need a dedicated committer to be point of contact

• Johnathan was suggested who is doing some porting of AC to other platform already
  • Intel team can also help with advice on TI side.
• Suggest a commit of ~8 hrs of advice over ~1 month period
  • If insufficient, checkpoint and reconsider
  • "Ben" will talk to Johnathan about it

Oliver mentions that we are already spreading ourselves thin with maintenance obligations w/ POG effort (narrow focus, more support)

• Are we spreading ourselves thinner with these other platforms
• Paul notes that community is expressing interest in this platform so it is right thing to do
• If this work is done by external contributor with some support by us, this is a win for TPTP
• Might bring more Mac users over to interest in TPTP which has long term value

We also discussed a bit about distribution models for the result if it is successful

• Similar to discussion we had about other platforms at (TPTP-PMC-20081015)
• Several basic options
  • No binary distribution from project (maybe a link to some external build)
  • Provisional binary support
  • Eventually an as-is component

Chris suggested that without an ongoing support plan, the only real option is "no binary distribution".

• It is possible that if it works initially interest from community will grow to provide basic support to make it provisional
• Paul was pretty sure that some folks would appear to contribute to ongoing build/test of mac port if it is there
  • There were previous volunteers for this.
• In any event, the contribution and distribution/support question can be separated initially.

Eclipsecon 2009

The team discussed current plans about Eclipescon talks

Intel is currently planning on

• Updated tutorial
  • Since maintenance only, very little updates needed
  • Will invite Eugene to participate if desired
• One long talk on a topic that Chris is working to get permission to discuss in more detail
• Some discussion about working with Eugene on a "supporting community" talk like Oliver suggested
  • Team wants to be a bit conservative and concentrate on 2. Still trying to decide how much to do on this one.

IBM is planning

• Short talk or two on POG effort
  • This might supercede cover Olivers goal of a "supporting community" talk above.
• If time allows, a possible long talk on Test project
  • Paul mentioned that he'd like to support this

Eugene noted that there is no concept of "track" this year.

• There will be some categories of talk

Oliver noted that he will be there at EclipseCon

Misc

Eugene noted that he will be on parental leave in January

• Congratulations in advance to Eugene