PackagingWG/2020-07-22-pip-teamwidemeeting¶
Legacy Wiki Page
This page was migrated from the old MoinMoin-based wiki. Information may be outdated or no longer applicable. For current documentation, see python.org.
Team Meeting: Wednesday 22 July
Attendees
Nicole
Bernard
Georgia
Sumana
Ernest
Pradyun
Agenda
Status/blockers
Nicole - research plan for pip functionality. now done with docs research plan. Need review of pip search survey from Georgia and Bernard. Standardizing further followup call to action
Bernard - working on pip in package manager ecosystem research plan. Looking at #8594 (#8546?) which is a blocker for Pradyun at the moment.
Georgia - will look at the pip search survey, working w/ b&n to put the research plans on a calendar
Pradyun - Circling back to 20.2. Looks like a few PRs that need to be reviewed and merged. Need UX input from #8594, and need a changelog/release notes support from Sumana.
Many minutes after this call - Pradyun and Sumana to talk
Ernest - nothing to add.
Sumana - No blockers at the moment. Want to support the July release and ready to do the reviews that people need of me.
July release blockers
https://github.com/pypa/pip/pull/8594 (PR) (https://github.com/pypa/pip/issues/8546) – need UX inputs/go-ahead? Bernard is on it! Will ping Nicole if needed.
Need changelog/release notes entry for the new resolver. (ideally, needs input from Sumana)
2 PRs needed – updating vendored deps + removing a deprecated thing… Pradyun is on it.
date: hope for release tomorrow. If not tomorrow, probably Monday to avoid Friday release
UX Team Planning Updates - Draft Schedule (https://www.notion.so/simplysecure/Draft-PIP-UX-Schedule-July-October-2020-a2d55ed05fa340728da75d139b24ab0b )
Docs Feedback ideas
Sumana: Many of us are friends with the ReadTheDocs team — maybe there should be a readthedocs feature that would help with this!
Maybe there’s work that they could do and then we could inhereit
Nicole can follow up with Eric/ReadTheDocs Team
Maybe we should move the recruitment plan into July?
Questions for Sumana/Team
We’re realizing that this scope is quite large for the time, and wondering if we should prioritize, and how to plan for other ad hoc support that may be needed
Are we still aiming for September/October wrap?
Technically CZI goes through December 2020, so we can re-work the flow of hours/time - Focus on End of November
TODO: Georgia check contract end dates
What other timing should we be aware of?
Release timing - October release will have the resolver by default
Testing pre & support for feedback on the new resolver as it switches to default in October
Planning around vacation/travel etc
Ernest out the first week of August
Telemetry research proposal idea — do we want to pursue it? Under pip-ecosystem or is it out of scope?
Bernard: A concrete thing that we might be able to do is focusing on how telemetry is done in other package managers, e.g. pros/cons, their communities reactions, ideas gathered from other projects (e.g. groups like Tor Metrics)
GA: Simply Secure is actually working with Tor Metrics right now as well, so we have some insights from that data and community needs
Nicole: Agree that it’d be nice if we could bundle it into the pip-ecosystem epic as well. Would the end output be a recommendations? Implementation ideas? etc?
Bernard: We wouldn’t be able to recommend an implementation plan without input from the developers, so maybe instead it’s more recommendations of the telemetry could look like
Nicole: also, we could share the community openness to telemetry
Sumana: Helpful to understand what the bits of work are — what is possible, ethical, desired, outcomes we’re actually seeking. This feels like an extremely important thing that I want to have happen, but maybe isn’t in scope / possible in the time we have available
A good balance would be to avoid putting any specific deliverables for the telemetry, sometime at the end *after* the deliverable things – get a “quick” list to then make it possible to add funding on a follow up.
Pradyun: if we had more time, YES, this would be awesome – but I feel like we don’t have enough time?
Georgia: Two ideas:
Georgia reach out to an academic who works in this area to see if someone is working on this already
SH: Tidelift or PyUp (and more)! Some of them work across ecosystems.
Georgia: CHAOSS and CleanInsights work from Guardian Project
UX Team ask some questions about this as part of the interviews, but not focus on an in-depth telemetry research question
“get community feelings about this”
SH: Right now, a desktop client sending info back to a server about stuff, eg: Crash Reporting on FF, OK because they don’t care when they [note taker could not follow along]. If you are able to ask about telemetry, one of the ways is “ongoing” vs “on-crash/significant-error” telemetry.
Yay expertise! \o/
Privacy preserving telemetry!
Georgia: Maybe we can align with other efforts that are going on in open source
If this leads to a proposal in the future, Simply Secure can help work on this

If we wanted to explore the idea of creating a pip package for user research, how could that potentially happen?
Sumana: debian bug report CLI tool - reportbug
Bernard: debian popcon — lightweight consent around collecting metrics
Sumana: It would need to make an API call to some service, what API does it call?
Georgia: The idea would be something that could be a tool in the toolkit for user feedback in an ongoing way
Sumana: this is something we could get volunteers to build if we have the bare architecture and policy details worked out. Leverage conference sprints.
What API does it call?
Where does the data go?
Who has access to it?
Add this idea as something to explore during August for the UX team
Georgia: could be a CLI tool to interact with user research surveys