Jump to content

Mobile web/Etherpad/Retrospective-2013-11-12

From mediawiki.org

Mobile Web retrospective, 2013-11-12 Etherpad: http://etherpad.wikimedia.org/p/MobileWeb-Retrospective-20131112

What's been working/should we continue doing?

[edit]
  • Being on the deployment train
  • Alpha/beta promotion/demotion+
  • visual editor spikes - rob moen
  • good learning about editors
  • Good iteration names+1 and the fun part of naming them+
  • team communication (including remote folks)
  • being able to merge quite large changes in reasonable time in gerrit
  • progressively cleaning up the codebase

What hasn't been working/should we stop doing/should we do better?

[edit]
  • QA (esp. Nearby tests)++++
    • making it easier to run tests on ANY instance?
    • Juliusz was supposed to do a TDD tutorial, hasn't gotten around to it
    • ANYBODY should be able to write tests, it shouldn't just be Michelle
    • Pool collective knowledge into documentation around testing
    • Have a day-long workshop around testing?
      • Can we make this non-remote e.g. everyone in the office to make it easier
    • Book on Cucumber and related tools: http://pragprog.com/book/achbd/the-rspec-book
    • Documentation + links to documentation for tools we use

Action:

   Chat with QA, Juliusz, and Tomasz about coordinating a long workshop for the mobile team. Make documentation a condition for this [Arthur]
  • Being on the deployment train++++
    • Long waits until things change/get fixed++
    • Confusion around what is currently deployed and where

Action:

   Get back to posting git logs pre-deployment [Michelle]
  • IRC attendance especially around LD times++++
    • sometimes code not merged by LD and no one is around to merge!
    • who should actually be around during LD? patch author, product owner, subject area specialist, deployer
      • deployer responsible for creating calendar invite and alerting necessary participants
    • LD for emergencies - we have been using them for non-emergency things

Action:

   Whole team polices making sure we're actually using LDs for emergencies
   Deployer responsible for:
       * creating calendar invite for deployment
       * making sure people who need to be present are added to the invite and aware that they need to be available during LD
  • Good practice to be on IRC all the time, or at least as often as possible

Action:

   Add this to team norms [Arthur]
  • Velocity seems slower than usual+
  • The whole diff spike
  • A desire for everything to be perfect first time round
  • Designer confusion - product / designers not coordinating as well as could
  • meeting meetings meetings
  • Code review backlog
  • Broken patches being sometimes submitted to gerrit
  • can do a better job of making learning actionable++

Any outstanding quesitons/concerns?

[edit]
  • how do iterations fit into roadmap?
    • Kenan will tie things back to the roadmap during itertation planning meetings
  • how things like BetaFeatures and MediaWiki UI should take my time, and Limn graphs, and AB tests analysis
  • can we get/do we need more pepple able to review Limn graphs - bottle neck
    • do a brownbag!


Action items from last time

[edit]
  • visibility of alpha/beta features - some we just forgot about - Alpha as a place for features to bitrot+++++++ [Jon]
  • bingle could do more for me - resolving duplicates, backlogging enhancements, transitions +++ [Arthur]
  • Didn't QA editing across projects - generally QA'ing across many projects that are not enwiki +++ [Michelle]
  • communicating with ops / cross team working +++ [Arthur/Erik M]
  • Documentation is hard to find [Kaldari]