COSMOSCommunityMeeting20081119

Att: Jason Saurabh David Ruth

• figure out how we can handle support requests coming through
• schedule v1.1 around SAS' needs
• think June/July would be a good target for 1.1. In terms of what we would need for SDD. Probably more late June. June 26 2009? Align with Galileo? Ruth to come up with an iteration schedule and post it on the project page.
• one milestone in the middle? four iterations, with M1 the end of the second iteration. Milestone early March?
• Major deliverables:
  • SML bug fixes
  • Reconciliation taxonomy
  • Finish the SDD tooling for the BTG
  • Have our SDD runtime driver for conformance level 1
• weekly candidate driver - did QA help us test that? They used to do the testing for it.
• Keep the weekly candidate driver. Disciplined that we have regular checkpoints, confirm things compile, even if not heavy testing.
• Ruth to attend an Architecture call, do a meeting that reviews that describe the process, from Legal to Development
• Jason to speak with ME (tentatively Arch call on Dec 3). Ruth to coordinate the IBMers. Brad Beck of CA also needs to attend, anyone developing in COSMOS to attend.

• suspend RM, etc. Combine RM and ME? Majority of ME people don't attend the Arch call. Could ask more folks from the ME team to attend the Arch call. Let's keep the ME call as-is. David, Saurabh to cancel notices for their meetings and to attend the Arch call instead.

• Ruth to send Jason a list of the SDD CQs, mark the obsolete IBM one as obsolete, review the Legal (iplog+) process
• another issue that is concerning, one of the comments in the release review was that it was a small community. For 1.1, unless participation changes, it's going to be even smaller. Will that be a problem? It's a wildcard, a risk. Hope that it won't be an issue. Hopefully by then CA or IBM would have increased participation again. How much evangelism Jason has to take for COSMOS 1.1 and implied will IBM and CA be coming back? And if not, does Jason need to pursue that more aggressively than he has?
• Jason to talk about some of the cool stuff that SDD is doing on the blog might help bring in some interest.
• support requests, up to 3 different people now. Person on the newsgroup. Couple of emails yesterday morning, David asked some questions, they haven't responded yet. Got a direct email that Hubert responded to (Qualcomm). And there was a new one this morning. Saurabh is looking at one of the ones from yesterday morning, the issue with the toolkit, run the example and the developer guide. Both Sheldon and Hubert were helping a little with that as well. Turned out to be more involved than we thought. We have SDD excluded because they have a pretty good roster. Rest of the project, we have more leaders than technical people. As part of the evangelism, we need more people to do the work. Certainly if we get a new person, get them started on test and support.
• Mark McCraw helped with RM before. Would he or anyone else want to resume that or not possible? That's a possibility, Jason to talk to Mark. Pleased with Mark's enthusiasm and the work that he did.
• Mark has more job experience than John Arwe, and John is trying to make some fixes.
• Put a hold on the CMDBf drivers for now. Spin SML drivers to pick up John's fixes, but due to resources don't change CMDBf. For CMDBf, allow for small fixes to be checked in with a gang approach (check in but don't build, or build with a "buyer beware"). Saurabh to run the 30 minute end-to-end test at the end of the build.
• bug fixes different than feature development. Think it depends on the scope of the bug fix. Fix a typo different than a big change.
• Compromise: CMDBf is in shut-down mode. Requires approval for check-in. Saurabh to run the 30 minute end to end on each weekly and the testing at the end of the iteration limited to just the changes that were checked in. On the iteration builds, don't put them in there, only available from the integration build pages. For 1.1 we need to run the full tests: one for milestone 1 and one for GA.
• SML: shut-down mode. May want to GA it with 1.1, or may still be "In Research", but at least build John's fixes and test them ... regular JUnits. Then there's the manual JUnits that Srinivas was running every iteration. Running the JUnits will be sufficient. Doubtful that we'll change the UI, which is what's covered by the manual tests. Approval, JUnits run weekly, JUnits run iteration.
• For those components in shut-down mode, prepare a list of defects for the Arch meeting for the defects that we want to commit. Any of the team leads who have any objections, speak up otherwise they will be approved. Any changes to Legal, haul in Ruth.
• Check and balance, when the list is presented, the other team leads ask, "any 3rd party changes?" if so that tells them to notify Ruth. Rely on peer accountability.
• List to be at the end of the Architecture call.
• question remains: who do we have available to look at issues coming from the outside? may just mean that we have a slower response time. Is there some sort of minimal service level that we would want to informally agree to? Puts people off if there's no response at all. What we should at least attempt to do is ask questions back. Keep track of the questions that we get, and if any are asked over & over, add them to the FAQ.