OSPR Brainstorm Meeting notes 04/27/2021
Goal, What does success look like?
“Small” backlog
Few stale PRs
Increase quarterly merged contributions
Increase (or just measure for now?) edx-used contributions
Problems:
Lead Time
OSPR tickets take long, don’t want to pull it into a sprint.
Closing tickets after 2 weeks of no response after edX took weeks to not respond, it was very disappointing tickets.
Pulling 5 tickets into a sprint and not being able to close them, waiting for author to respond is frustrating.
Metrics
Also long OSPR resolution times make squad’s metrics look bad.
Burn down rate looks bad if sprints have tickets that keep lingering around.
Authors
Imposing rules on authors can be detrimental and seem imposing.
Squads
Squads don’t see OSPRs as priority, however, 60 OSPR per month for all of edX doesn’t seem that much.
Squads feel they have more important things to do than to work on OSPRs
Tools
Is Jira too far away from where the community engages with our code for them to know to look there to find out what’s on our backlogs? (problem w/r/t alignment on priorities)
Ideas: ⚡️✨
⚡️⚡️⚡️⚡️(edX Buy-in++) How might we connect and prioritize squads with Open edX goals and community?
Open edX “buddy system” to establish stronger connections.
Can each squad have an engineer who is the OSPR manager for the squad? (Engineer who can do a decent job of representing other members of the product delivery team (product, ux) and make informed choices about what sort of review is necessary)
✨✨✨Do a consumer review
Explain the goals
“Why we care and why you should care”
What are the metrics
✨✨Implement means for accountability
Standard dashboard to incorporate Eng Ops Review
✨✨Seek more input from other team’s on their impediments, etc.
⚡️(Process++) How might we optimize OSPR reviews?
Better distribution of OSPRs across teams (not all on T&L).
✨Author should have as much time to respond as edX took.
Self-serve OSPR process
(OSPR API++) How might we clarify expectations of the OSPR process?
Could we implement a two sided contract where both edX and author commits to certain respond times.
Yelp for contributors - reliable authors with short times get five stars, we review them without fear.
✨✨✨Clarify the different categories of openness for repos
(Roadmap++) How might we externalize edX’s needs to the community?
How can we help the community understand what kinds of contributions we consider valuable? (and are thus more likely to review and merge)
I keep thinking about this line in our OGSP: “Align with community on valued contributions”
How might we use INCR tickets (as Github Issues?) to bring them closer to the code/community, where people might be more likely to see them?
Can we get product managers on other teams to use INCR as a mechanism to help their velocity? If they receive OSPRs for things they asked for then they may be more likely to take them seriously.
✨⚡️How might we create a shared roadmap?
technical/architectural to start - start with the domains w/i our control
How do we measure OSPR outcomes?
Better metric would be how much time a ticket was blocked by edX.
Prioritization Techniques:
See which things we can control ourselves versus requiring things of others.
Get more data.
OSPRs reviewed by core committers?
Segmenting stats by committer type?