Wikimedia Apps/imported/Retrospective-June 04 2015
Previous Retrospective Actions[edit]
Action! Clarify what devices/configurations are supported in Android/strategies for support, deprecation [Stephen] NOT DONE
Action! Write user stories to clarify requirements for improving CI & dev workflow ONGOING
- some motion on this @hackathon for Android (there are some Phab cards) and commitment from RelEng to support
- not so clear for iOS, but BG is planning on pushing on this front.
- Action! Next step: reach out to Greg to communicate requirements [Brian]
Previous previous actions:
Action! draft suggestions for reforming Quarterly planning and share with team [KL] DONE
Worked Well[edit]
- Actual Ideas from Hackathon that look like we can turn into features
- Prototypes are moving quickly
- Some much needed changes like full width images got in before we dropped ios 6 support despite hackathon (go team!!!!)
- Alpha/ beta channels working nicely on Android
- Productivity of apps at Hackathon (WOW!)
- Using Crash reporting to fix issues
- Prioritzing, Fixing, and Responding to bugs and UI tweaks
- Prioritizing performance and user experience
- Dev<->Design tight feedback loops
- Formal Design & QA sign-off steps
Worked Poorly[edit]
- BG: (Lack of) continuous integration
- BG: Unit tests have broken multiple times, further exacerbating the lack of incentive to write more tests
- BG: Regression testing on old OS's
- BG: Had one nasty release blocker on iOS 6 that should've been caught *much* earlier in the process
- BG: Testing & responsiveness to issues with various languages
- BG: we've been crashing on startup in Norway for... let's just say too long +
- BG: App Store reviews as a feedback signal (as one review noted: we are "hedgehogs")
- BG: need to look into something that can provide notifications (i.e. a loop for the feedback) like App Annie
- BG: Process for updating regression test cases+
- BG: We need a way to make sure all the important stuff we do has a corresponding update to regression testing docs
- Crash reporting accidentally getting turned off...
- BG: beat me to it, but I kind of covered this in "Process for updating regression test cases"
- BG: not only was crash reporting broken in production, but it wasn't even working in Alpha channel :-(
- BG: OTRS++(+BG)
- Engaging Research for User Testing+
- We dont understand how are users are or where they are - need analysis at this point +1 (+100000BG) -99999BSj/k
- BG: or what app/device versions they're on...
- Code review meetings (too busy)
- Communication with mobile web (readership as a whole)+++++
- Double meetings (iOS and Android work closely possibly at the expense of excluding mobile web)
- Fighting with the Org - refighting old battles
Confuses Us[edit]
- CI (see previous point about fighting with org) (+BG)+
- Criteria for task acceptance (might just confuse me); where to pick up "accepted" tasks?
- What does velocity actually measure/what do we want it to? (Why so many 0 tasks?) +
Top three items[edit]
- Communication with mobile web (readership as a whole)+++++
- Would be nice to have better defined process for upstreaming changes that would benefit all platforms
- Would be nice to have awareness/process around features and wins in the app that would be beneficial across platforms
- Way to collate our services needs to get more leverage as stakeholders
Action! Mega readership standup on eg Weds; scrum of scrums for readership (+ design and product); types of topics = channel/feature matrix or rolling up needs [Kristen]
Action! coordinate needs and roll up to Adam so he can communicate them out
- BG: OTRS++(+BG)
- We dont understand how are users are or where they are - need analysis at this point +1 (+100000BG) -99999BSj/k
- BG: or what app/device versions they're on...
- CI (see previous point about fighting with org) (+BG)+