| Who | Notes |
|---|
| Jim | - Nothing new to report that's not already entered into Jira.
- Jim is working with Perry on getting the slots mechanism and tables into working in the new measurement framework.
|
| K-T | - Middleware items are now in Jira and they have started a sprint.
- Greg is at CondorWeek so he's making some progress in his task - he is learning more about Condor and talking to the people on the team.
- Steve is delayed due to unexpected family leave.
- K-T raised a question which he'd hoped to have Mario answer at this meeting: what should be done with stories which are pure refactoring and don't alter the external demonstrable results other than things still work, maybe they are more efficient, and maybe they're easier to maintain in the future. How should those be dealt with in the process?
Josh Hoblitt responded in later email: "Refactoring is a risk mitigation activity. If the accounting for the "value" of a story is only in terms of "positive" value (as in, adding features or functionality) it's difficult to justify any sort of risk reduction strategy. However, if you consider an additional risk dimension for every story the picture [should] change. A refactoring story may not have much direct feature/business value but it should contain calculable risk mitigation value. I would also account for future cost savings due to reduced effort as a risk "value" because cost & effort are business resources as opposed to direct business value. On that basis, it should be fairly easy to rank refactoring or other risk reduction tasks as being more valuable than feature stories in the back log."
- K-T is currently at QSERV Developer Week and going to New York most of next week for XLDB Healthcare.
- Rebooting the SAT and the new Butler are his primary work tasks (and are in Jira).
|
| Simon | - Simon worked with the group in France to fix a bug in obs_cfht with how it represented ISO dates. They are back up and running. Simon is not convinced that the defects provided in the obs_cfht package are apparently not sufficient to do the processing they want to do. The collaborator said he would look into that.
- K-T: But they should understand that a lot better than we do, right?
- Simon: I am hoping that is the case.
- K-T: My question is - do they understand how to put the right defects in the defects registry, right?
- Simon: Yes, we represent defects in a different way than CFHT did. They need to be added to the defect registry with date ranges, etc. So there's a little bit of a learning curve there but it should be pretty shallow.
- Simon looked a little bit in how LSST DM might use git-fat to store binary files in git repositories. He got some feedback from K-T, Jim, Mario, Josh. It looks like it's totally doable. The only question is how the anonymous poll is implemented.
- K-T: You didn't like any of the things I mentioned?
- Simon: They're great. It just which one of those we choose. I think using rsync daemon is cleaner since we won't have to distribute a public key. The only thing I don't know is how hard it is to setup rsync to get both read-write and read-only to the same module.
- K-T suggested an alternative. On the ssh side, a public key wouldn't need to be distributed if that username was locked down to a single command and have no public key, no authentication whatsoever. Simon said that was another option.
- Simon will look at the rsync daemon option first.
- Jim Chiang brought up a problem he's having reducing deccam data. Simon thinks it may be related to using the bracket notation. Simon wondered if it would be possible for JimC to cherry pick daf_butlerUtils back to version 7_2 since there's so much overhead to upreving to v8_0. K-T said he would look into the issue that afternoon.
|
| Dick | - Dick finished a tutorial on processing Stripe 82 data on the users' own machines. He's already responded to Yusra's excellent review comments and is ready to move the Task into 'Done'.
- Dick is now starting on the next tutorial which is on PhoSim processing. He hopes to have it in good shape fairly soon.
|
| | |