Skip to content

Latest commit

 

History

History
68 lines (65 loc) · 4.53 KB

numpy-community-2023-06-21.md

File metadata and controls

68 lines (65 loc) · 4.53 KB

Snippet from the 2023-06-21 NumPy Community Meeting notes. For the full meeting notes, check https://github.com/numpy/archive/blob/main/community_meetings/community-2023-06-21.md

Present: Charles, Sayed, Tyler, Raghuveer, Mars, Sebastian, Ross, Brigitta, Inessa

New Topics

  • Review #1 stage for GSOD-NumPy 2023
    • Will share on this community call and make Github issue to open to wider community who may not be in call
    • Also shared in NumPy Docs meeting earlier this wek
    • Drew some Concept Art
    • Discussion points
      • Writing out specific parts I would like feedback on
        • Is there an issue that was actually completed during a sprint, that can be used in the comics?
        • Metaphor: Making digital pages into physical media (Github issues as paper scrolls)
        • Characters: over-confident, under-confident and just right
          • Over-confident: "I bet I already know how to do this!"
          • Under-confident: "Can I start working on this issue?""
          • Just right: "Let's review the documentation if this already exists and if I'm still not sure, I'll ask someone."
    • Next steps
      • Putting together review: Github issue
        • 3 main drawings, as shown above
        • Chapters
          • Introduction
            • To characters and how in-person sprints work
          • The sprint itself
            • Collaboration and the importance of different approaches
          • After the sprint
            • Meeting remotely, connecting to mailing list and Slack
          • Goals and future
            • Contributor's goals, how open source processes are open to change too
      • Annoucing that open for review: mailing list and Slack channel
        • 2 weeks period for getting reviews?
        • But I think most comments will come from community meetings
        • For Review stage #2, want to coordinate so people know they can join community calls to give reviews
    • Comments from this call:
      • Different levels of problem, of 'sprintable'
      • Emphasize following-up after the sprint
      • It's normal it won't be done in one sprint
      • De-emphasize the onramp of sprinting becasue fly-bys
      • Focus on user-contributor, most likely to follow through, sustainable commmunity growth
      • There's a specific example Ross has about loading data, NPZ archive, why won't it work like a dictionary?
      • Ross opened issue
      • Ganesh wrote implementation, Sebastian and Ross review
      • NumPy is driven by people who use it
      • Maybe start with a character in classroom (meteorology thesis character)
      • NPZ as 20 year old, not brought up until by users
      • Instead of sprint leader, maybe Professor character guiding students
      • Maybe interview, get links to PRs and issues, get narrative flow
      • Sprints as not primary onboarding mechanism, people who make larger contributions come through other channels, driven by scientific research
      • Mars will follow-up with Ross
      • Similar examples: AstroPy
      • We serve everybody who lives on top of NumPy
      • Can check against user surveys how people use NumPy
      • Maybe add sprint to end, not be center of conversation
      • Can talk to a new contributor to NumPy, a maintainer of SciPy but newcomer to NumPy (talk to Melissa)
    • Previous steps in GSOD-NumPy 2023: