Wikimedia Apps/Team/Retrospectives/2015-08-14
Appearance
ANDROID
[edit]August 14, 2015
Previous Retrospective Actions
[edit]- Action! Talk with iOS about splitting QA resources [DMITRY] NOT DONE (iOS needed less QA lately, we should talk about strategies moving forward around QA)
- Action! Clarify what devices/configurations are supported in Android/strategies for support, deprecation [Stephen] IN PROGRESS (making good headway, Stephen started successful etherpad)
- Action! Get more android devices [ADAM]
- Get Samsung devices? Figure out which devices are most relevant. High pixel density devices, Galaxy S, etc. Adam to share devices analysis NEARLY DONE (Crossing t's and dotting i's, putting in purchase requests to IT in next week)
- Action! Discuss velocity calculation with team [MAX]
Worked Well
[edit]- Good conversation around clarity of roles: PO, devs, triage, priorities, etc
- Got Needs Triage down to 10!
- Production release! o/
- Excited about the CI improvements Stephen is undertaking
- Users seem to be sharing positive feedback.
- Pretty huge amount of code submitted.
- Strong and productive code reviews.
- ^ continued attention to detail and pursuit of perfection
- Looking forward to digging into the service. I feel this opens up a lot of really great options most strictly app devs don't have.+9999
Worked Poorly
[edit]- Rollout to RESTBase service got delayed, but coming real soon. Using the service from the app is harder that I thought. (Doesn't fit into java-mwapi structure) <-- would like to see longer term plan for networking, RESTful consumption, third party libs
Confuses Us
[edit]- How do we know when QA or Design is working on something on the board?
- Stephen being a whiny Baggins: a little sad that we're talking about Q2 planning right now. I just finished Q1 planning!
- More whininess: Release builds at WMF
- Let's talk about crash reporting
- "Still" versus "tap" water
- "Gaseous"
- Sparkling?
- "Gaseous"
- Magnets: how do they work? Are they magic (pretty sure the answer is yes but seeking detail)?
Top three items, from "Worked Poorly" and "Confuses Us":
[edit]N/A
Action Items
[edit]- Action! Team to handle more of it's own QA, and own decision on whether or not a task needs Elena's eyes
- AB: Team could use budget to leverage specialist's guild
- Action! Let's talk about crash reporting [MAX to start an email thread]
- Action! Weekly meeting with Design/PO to hash out sprint needs [MAX/DMITRY]
iOS
[edit]August 14, 2015
Previous Retrospective Actions
[edit]- Action! Try some new manual QA methods? Try a "testing party" where we try to produce crashes that users are encountering [BG] NOT DONE (Focusing on 5.0, could address in 2 weeks or table)
- Action! Figure out testing strategy going into final month of 5.0 [BG/iOS] NOT DONE (related to above, rough ideas)
- Action! Talk with Android about QA's availability/resources. [iOS/ANDROID TEAMS] NOT DONE (QA not needed for iOS lately)
Worked Well
[edit]- GH - continues to foster better discussions around patches
- Assuming QA work will go smoothly for 5.0 (big assumption), velocity is way up from sprints prior to Kanban
- ~ 6 weeks of data
- BG: Visual aids in code review (IMO) +1+1
- Code review continues to work well
- Communication has been good
- BG: Asking for more beta testers!!
- 70 OTRS tickets of people asking to be added as beta testers
- Contacted by volunteers to improve app accessibility, yay leveraging working in OSS!
- App has been "coming together"
- The community thanked us! Better than the alternative...
- Design iteration/feedback
- Developer feedback on design
Worked Poorly
[edit]- QA has not had much in the Needs QA column, but 5.0 Beta Staging has un-QA'd tasks.
- BG: Not poorly, but could be improved: Travis had some downtime recently and build times are a tad long (even though we're not running many tests)
- BG: "Rogue" transition work, coping with unease of app stability
- Data layer continues to bite us
- Lack of repro steps or examples in Phab tickets/bugs++
- Maybe in-app bug reporting w/ screenshots, logs, etc.?
- Schedule has fallen behind - or rather we have fallen behind - admittedly the scheule is agressive
- Design has been slow
Confuses Us
[edit]- How do we know when QA or Design is working on something on the board? ++
- Process for raising, tracking, and resolving design questions/concerns. How we work with design ++++
- MH: Blurring division between native/webview article presentation (also search results showing large images speed tradeoff)
- WIP
- Kanban is supposed to help us work on one thing at a time, but WIP limits based on points messes with that concept.
- BG: WIP Limits are supposed to help *enforce* working on one thing at a time. wondering how you think they work against it
Top three items, from "Worked Poorly" and "Confuses Us":
[edit]- How we work with design ++++
- How do we know when QA or Design is working on something on the board? ++
- Lack of repro steps or examples in Phab tickets/bugs++
Additional Actions
[edit]- Action! Calculate velocity now that we have the equivalent of three 2-week sprints of data [MAX]
- Clarify with team what is actually done-done, and what still needs to be addressed.
- Action! Brian and Monte can come to Design Reviews [MAX]
- Carve out some time for interaction issues for this meeting.
- Action! Break up Design review for iOS/Android [VIBHA/MAX]
- Action! Meeting notes from Design Review: why decisions were made around design [VIBHA/COREY]
- When we do things, how long they take.
- Action! Create a new column for "In QA"? [MAX]
- Action! Email "Lack of repro steps or examples in Phab tickets/bugs" issue for email convo [MAX]