Sprint Wrap Up
Attendees
General
- Indicates who took minutes -
- Google-hangout at:
Agenda
- Meta analysis of sprint process
- Summarize final state
- Walk through in-process and unstarted tickets
- Teeing up for next sprint
Minutes
- Osman Din finished up a couple tickets and learned a lot about ModeShape storage. This is ongoing work.
- Greg Jansen was getting through PR's for AuthZ work. Added a war file with authorization. Expect PR to be in today.
- Had a general question about branches based on previous PR branches. We should squash commit and make those PRs complete. Is it possible to retain commits that address comments? Chris Beer thinks that it's better squash all those revisions for the sake of future developers. We can reduce the lag time to get things into master by reviewing and merging early and often. Chris Beer suggested that we all merge each others and then have Andrew Woods to accept it.
Should multiple tickets be addressed by a single PR? Probably not a good idea. - Wiring and configuration is becoming an issue.
- A. Soroka: the indexer is functional, but hasn't been fully tested and isn't very exciting yet. Now that the merge has been accomplished (thanks Chris Beer) someone should do the testing and produce examples. There's plenty of room to expand, like mapping RDF instead of named fields, etc. Consider how to map into indexes of particular formats (BlackLight, VuFind, Islandora, etc.) There are code quality improvements.
- Concerned that we've lost touch with former contributors and keeping in synch.
- Unknown User (escowles@ucsd.edu) thinks sprint went reasonably well. Having less of Andrew's time is a big loss. Is wrapping up testing stuff and as much documentation as he can.
- Eric James learned a lot on this sprint, liked the focus and needing to keep up with IRC and such. Thinks documentation and wiring and modularity is a good place.
- Chris Beer really dislikes performance testing and worries he did it unsuccessfully due to the number of variables. Not sure whether any of this reflects real-world usage. Is skeptical that we can't get performance of ingest of F4 to F3 levels. Unsatisfied in our performance testing regimen and thinks it took away from our feature development. A. Soroka added that it doesn't help that we had no reasonable goal (beating f3 isn't a real goal). A. Soroka suggested when people ask for "performance improvements" we should ask what they want to do and how fast. Chris Beer completed a number of small features and those commits didn't get reviewed very quickly, in part because of east coast workers. A. Soroka added that it's not just the time difference is a problem but that we're funneling too much through Andrew. Greg Jansen suggested that we should assign a reviewer at ticket creation.
- Michael Durbin is wrapping up documentation
The team went through all Current tickets and moved any that wouldn't get done.
The team went through the backlog and sorted by which should be completed before the release.
The team went through the list of icebox tickets to delete and deleted all of those that the team agreed could be deleted and promoted those about which there was any uncertainty.
Actions