Community Tech/Retrospectives/2015-10-26

Action items from last time

• Should have email convo about most work being in English so far. (Frances)
  • email sent
  • Wasn't as bad as we thought, but should keep it in mind
• Maybe could be a standard practice to circulate a couple paragraphs about each completed investigation task to see if that triggers discussion? (Kaldari)
  • Informal norm
• Frances will move the github workflows email thread forward
  • Should be written up on the wiki, probably, maybe after another few rounds of trying out the process
  • Successfully got feedback/etc. by following the proposed process, excellent
  • (done)
• Danny: Should we have a standard story grooming meeting, mid-sprint? +++
  • -- scheduled
• Should we have a central place for people to report broken gadgets/templates etc.? Not allowed on Phabricator right now. Perhaps a page on Meta? Or a dedicated Phab board.+++
  • -- Email sent (Danny will follow up more)
• Kaldari + Danny need to work together on wishlist survey
  • -- Done, more coming up later today
• Kaldari will involve Johan in getting Template:Cite doi RfC closed
  • Done, but still not closed
• Danny will set up a meeting with team and Kunal (and others?) about Gadgets 2.0
  • -- Done, happening tomorrow

What has gone well?

• Dev work has gone well. RevisionSlider prototype and Special:GadgetUsage page are both nearly done
• Impressive results on the satisfaction survey, far more responses than we thought +2
• Story grooming debut meeting was very useful
  • We decided on a set of tasks for next sprint, currently In Analysis
• Reviewing things on GitHub seems to be going OK and is less confusing than it used to be +1
• We have a PM and it's *fantastic*, basically, I can more easily focus on my work without worrying about larger-scale stuff; Ryan does more engineering management +1

What could have gone better?

• Danny's mostly focused on tech + wishlist surveys, getting up to speed on Phab tasks
• Unclear on our sprint commitment (is it really a commitment?) ++++ (this is probably similar to the process discussion)
  • Should we try doing some themed sprints? (Gadgets, Page stats, i18n?)+
• I feel blocked on plagiabot but am not sure why, some of it is probably that I don't know how to set up a dev environment for that +++
• Haven't been having agile workshops

What else is on your mind?

• Who is the WMF DBA now that Springle is gone? i.e. who do we ping for database-related code review (Jaime Crespo?)
• Possible designer resource
• Part-time Outreachy mentoring tends to take up serious amount of time (as volunteer, doesn't sound sustainable to me...)
• Creating subtasks from sprint tasks and keeping them out of sprint causes unnecessary context-switching
• Scrum/kanban/how are our processes going/what works best? ++++

Deeper discussions

Sprint commitments, kanban, etc.

• Unclear on our sprint commitment (is it really a commitment?) ++++
• Scrum/kanban/how are our processes going/what works best? ++++
• How do you deal with things currently coming in?
  • Should still go through review/estimation/prioritization before being worked on
  • But new items can jump straight to the top of the queue
• Not much different from today's process
  • Maybe break down into smaller subtasks?
• As a dev, most important is to have sensible work queued up--minimize contex switching
  • Stream of tasks seems like a better fit than a multi-week commitment
• Find a balance between large and small tasks
  • Splitting tasks creates overhead, but value might outweigh costs
• Would splitting tasks reduce chances of being blocked?
  • Some devs prefer granular tasks
  • When shouldn't we split up tasks?
    • Distinction between breaking at PM level (stories) v.s dev level (eng tasks)
    • Shaving vs. exploding to split stories/tasks bit-by-bit or all at once
      • This is reminding me of learning styles differences, global vs. sequential (http://www4.ncsu.edu/unity/lockers/users/f/felder/public/ILSdir/styles.htm)
    • Devs should feel free to shave off additional tasks when helpful (involve PM for prioritization if [and only if] appropriate)
• For tomorrow's sprint, consider whether it's a commitment or just a goal...have a shared understanding as a team as to the intent
  • Smaller tasks might make commitment easier.
• Single sprint goal is harder for this team due to diverse tasks
• Phab can set WIP limits, but not task count limits if you use story points
  • Can use the story point limits as motivation to get more granular (i.e. break those 15-point stories down to 5 points each)

Action Items

• For this week, aim for smaller tasks
• Kevin: Agile workshop next week (Kanban/sprint planning)
• Follow up on plagiabot offline (on mailing list?) (Frances)
• Write up github workflow on wiki somewhere (Meta Development subpage) (Frances) (Done)
• Talk to Quim about GSoc commitments - Niharika will initiate email
• Kaldari: Post redacted retro notes
• Danny to follow up on template/broken gadgets page/board
• Follow up on DBA question (Kaldari will investigate)

Retro of retro

• Still really like the longer meeting, feels much less rushed +1
• Helpful, this format works well ++
• Good facilitation+
• Do we feel like we get to have enough in-depth discussion?
• Would be nicer if everyone fills up the buckets before the meeting starts (not always possible, I agree)
• 2 weeks is good I think +