Discussion items


Jim provided the following notes in advance of the meeting.

  • I'll miss today's standup due to a conflict with our department orientation.
  • We have no blockers.  Still waiting on a review of some of last sprint's code, but it's a big one, so I understand why it's taking a while.
  • I moved an issue out of the current sprint and a couple of new ones in to replace it (not as many SP, though), as the new ones were higher priority and the old one is now blocked by a need for a design review (ended up being more controversial than expected).
  • Mario was out earlier this wekk at a Data Science Workshop retreat.
    • His sense is that in the last 2-3 years python use has grown and changed.
    • He is worried that we are developing a Not-Implemented-Here syndrome and are ignoring interesting developments going on elsewhere.
    • He's going to think on this over the next few weeks and come up with solution.
  • Mario is trying to clear out tickets from way back when. E.g. rebasing was enabled.
  • In response to Mario's mention that he attended a CCB meeting today, the assembled masses gave him an 'attaboy'.
  • Jacek was at ADASS this week presenting a talk on QSERV which went well.  There was a lot of interest in QSERV.
  • Regarding the recent qserv review, all the documentation related to using eups for qserv was updated to reflect what Mario suggested. Jacek said he considered that task completed.
    • Mario said he hadn't looked at it yet but that was one of his highest priority items.
    • K-T mentioned that Fabrice was able to cut a Release and no one has been screaming about eups getting in the way of their development work.
    • Jacek said he also tested the new Release that Fabrice built.
  • Qserv team is improving Zookeeper-based metadata; also looking at whether clients are 'good enough'.
  • Jacek is also starting to think about the Database Archive, Global Archive, Image Services, Image Catalog Services.
  • SLAC/LSST is extending an offer to one candidate; they expect him to accept it.
  • Fabrice will be at SLAC for 6 months; he will start Oct 16.
  • K-T thought he didn't have much to report but ...
  • K-T was also in clean-up mode - getting rid of old issues.
    • He needs to look at the last of Dick's documents which he hadn't gotten to yet.
    • He did some stuff for TOWG
    • He needs to work more on the design documents: Ingest QA and the Release Process.
  • Mario inquired about the Butler.
    • K-T said it was published with a prototype. However bringing it to its final state will take some time; He anticipates towards the end of the month he'll be able to work on it.
      • K-T said that no one has given him any feedback on it yet.
    • Mario asked: Is it now up to Apps folks to try it out or modify the codes to use the new API or ?  
  • Russell has been able to build everything on Mac Os X C++11.
    • There was conversation on hipchat that he was not able to read all of it. He asked if there an executive summary of everything we're planning on doing with compilers?
      • K-T: Basically we're enabling C++11 in all compilers which will alllow us to use a certain minimum number of C++11 syntax and features but not the full thing.
      • Simon: We're not deprecating any compilers?
  • Simon has a task ( DM-1112 - Getting issue details... STATUS ) this sprint to work on a utility which will build a camera from fits files. He had a conversation with JimC yesterday where they chatted about JimC's requirements. Simon wrote a brief document (Design page for DM-1112) on those requirements so now he can start looking at what they're doing. K-T mentioned he'd added a link in the Design Page of additional info of which Simon might be interested.
  • Not much to report. Has been looking at the older JIRA tickets to prune out deadwood from the backlog.
  • Robyn noted that much how-to information on the rapidly evolving developer environment is going into hipchat and/or maillists. That info needs to be harvested to update the various Guides.
    • K-T suggests that we wait until after the package reorg or the proposed move to GitHub occur so the info settles down into the final facts on the ground.

K-T provided some feedback on Developer development

  • K-T's been providing some pointed commentary on commit comments for SLAC personnel. Reviewers should comment on comments, but leads need to keep watch as well until everyone is well-trained.
    • Mario thinks it a better way to learn - by example - rather than reading a thousand pages of documentation.
    • K-T: We should all be looking at opportunities to educate.
  • Everybody should be aware that 'master' should never be broken. If a developer knows that not everything in an update is working, even if buildbot doesn't fail, that update should not be committed to 'master'.
    • This situation occurred because the sole integration test in use (demo2012) did not trigger a failure when the results didn't compare with the benchmark.
    • Robyn said that comparison test is being tightened up..
  • Nobody should wait for any significant amount of time thinking they don't know what to do. If teams are not having daily, or maybe every other day, standups so people can ask questions, make sure there is some avenue to ask questions. Make sure all developers know that they can escalate to K-T or Mario with their questions if they're not getting a response. And tell them not to let up in letting it be known that they need a response.


Action items