Packaging/2020-11-23-pip-team-meeting¶
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 Information may be outdated or no longer applicable.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.
Tuesday 23 November Participants
Nicole
Bernard
Sumana
Ernest
Georgia
Agenda
status & blockers
Pradyun: can join tomorrow’s meeting in its entirety
only 1 blocker, #9011. Not been able to wrap his head around why resolvelib is misbehaving. Stuck. Need to tag and work with TP.
also, would be nice to have if we could do it in time: Mac OS Big Sur changed things, including how Mac OS tags work. That then requires an update in
packaging{.backtick} + pip. But it’s ok to go without.this will go in 20.3.1
Bernard: spoke with Sumana about expectations for UX training. Now: continuing with pip research analysis, reports.
would like the mental model review (spreadsheet). Would be really helpful if some people looked at it.
TODO: Pradyun will look today
wants to get training stuff ready, ideally to do work in December.
TODO: Bernard need to pick up again the interviews with people who run package managers
TODO: Sumana to nudge people
Nicole: interviews re doc: done 3-4, have more scheduled this week. Starting to get a picture on what we should do to improve pip’s docs. Will be ale to make strong recommendations based on that!
continuing to summarise research finds we have
no blockers right now
Q from Pradyun: has nicole seen the Matomo PR/discussion? https://github.com/pypa/pip/issues/8517#issuecomment-730373194
? what is blocking it being merged?
consensus
is there a policy on PyPA/PSF stuff?
Ernest: no. We use Google Analytics… respects Do Not Track cookie. If we need a policy, we should contact Legal and start that process
Georgia: we don’t want to do something out of line with what PSF is doing
Ernest: comes down to where data is stored. per GDPR, we do not store really anything on any PSF property that is not critical to behavior of service. [This means a lot of GDPR does not apply] Here, user behavior, need to be more careful, since not critical to behavior of pip.
Google Analytics - storing user behavior [more detailed question about storage]
Ernest: for purposes of docs, we have used GA for other similar site properties
Pradyun: would be using the hosted version of matomo.org
Georgia: wondering whether to align with, well, PSF is the org here, so, use the analytics
Ernest: we’ve had minimal pushback on python.org and pypa.org having GA
Ernest: hosted/free service would be rad
by precedent, GA would not be a problem at all. a free service that is an alternative that PyPA wants to use, that is fine as a reasonable alternative
TODO: Sumana to take a fresh look after call at Matomo discussion
Bernard wanted to ask: what questions are we trying to answer by adding analytics to docs? to minimise extra data being unnecessarily gathered.
Georgia: what sections of docs are being used.
Georgia: it would be a PSF conversation to possibly move over to Matomo from Google Analytics
it can be hard if different sub orgs do different things
discussion of 20.3 release - deferred till tomorrow
discussion of prepping for UX training - deferred till tomorrow