2024-04-11 Nix teams gathering meeting minutes


Team Name/handle
Foundation board @thufschmitt @ron
Infra @delroth @Julienmalka
NixOS release
NixCon @Janik
RFC steering committee
Marketing @garbas
Nix team @thufschmitt
Documentation @fricklerhandwerk
Nixpkgs architecture
Nixpkgs @raitobezarius (15 mn late)


Proposal for project maturity levels

  • Problem: Beginners have a hard time to evaluate what they can rely on in the Nix ecosystem, while stability and maturity is one of the major reasons to choose Nix.
  • Goal: Significantly reduce the cognitive overhead for new and existing users when choosing Nix-related tools
  • Prior art:
  • Proposal:
    • Make something like Thoughtworks’ technology radar
    • Proposed degrees of maturity and relevance:
      1. Ecosystem
      • Maturity:
        • Can be run or set up with one command using Nix
        • Project is alive
      • Relevance:
        • Supports a well-known use case
      1. Community
      • Maturity:
        • Has at least one active maintainer
        • Contributions are accepted regularly
        • Can be used without reading the source code
      • Relevance:
        • Supports a common use case
      1. Official
      • Maturity:
        • Is maintained by a community team
      • Relevance:
        • Supports an essential use case
    • Notes:
      • Higher degrees include lower degrees
      • Criteria deliberately leave room to interpretation and refinement
      • Community team leads should set up a decision-making scheme (e.g. majority vote) and regularly revise the list (e.g. every 6 months)


  • @thufschmitt: agree on the big picture; decisionmaking is the crucial point
    • @fricklerhandwerk: this team could decide things; from experience all that’s needed is someone to drive the process, make proposals, implement change requests
    • @garbas: add to that that anyone should be able to make proposals, e.g. in an issue
  • @garbas: what we should add to the metrics is what we’d expect from maintainers. or should there even be obligations?
    • @fricklerhandwerk: the criteria already give an idea of that, and supposedly we should start evaluating based on that and refine as we go
    • @delroth: yes, if the same group of people makes the decisions, they should converge on something consistent. don’t need to be exact up front
    • @garbas: should still be clear what it entails to have an official project, what to expect. is it more work, more visibility?
      • @delroth: and what do you get out of it?
        • @garbas: one thing could be eligibility for funding, etc.
          • also mention on the website or other visible places
          • @thufschmitt: having a project under the NixOS GitHub org should already signal higher trustworthyness than random other things
    • @thufschmitt: linking back to the NixOps4 discussion with @roberth as an example, a community-blessed project can be expected not to operate like an individual’s project, where there will be some sharing of responsibilities. it’s improtant to set expectations in that regard
  • @delroth: taking a step back, having official projects also matters because official projects “block” development of alternatives that address the same use cases
    • @fricklerhandwerk: saying that particular problem areas have no canonical solution would also help, because currently much of that is simply up in the air
    • @garbas: would expect consolidation of alternatives given the proposed curation happens
    • @delroth: expect that setting up a bunch of criteria may produce no official projects in the end
      • @fricklerhandwerk: that’s unlikely but would be fine; the purpose is to be honest about the situation
        • at least we’d know what to fix first
  • @fricklerhandwerk: probably missing from the proposal is allocating a fixed amount of resources. if we do it in this group, we’d need to say how much time we spend regularly to achieve some measurable goal in a pre-determined amount of time
  • @raitobezarius: agreed on having nix-community as a staging area
    • question on moderation issues: there was a recent PR to awesome-nix by a banned user that was reverted
      • should nix-community align with moderation decisions? since we’re tying nix-community to the “official” brand
      • asked @zimbatm about it, but no answer yet
      • it’s important because nix-community is quite big and semi-official already
      • @thufschmitt: would be ideal if we can get alignment
        • @raitobezarius: ideally we’d be able to tell people to go to nix-community first and then graduate to official as their work matures
        • @fricklerhandwerk: no owners here today, should discuss this together
        • @delroth: there will probably be more action items for full alignment
          • e.g. nix-community has their own funding and infra
          • @fricklerhandwerk: the problem is not that they’re independent entities, but whether we clearly say that it’s supposed to be that way or not
  • @janik: this team should probably not make the decisions because it will take too much time from the group
    • it’s also somewhat unrelated to what we’re trying to do here
    • teams can be asked for advice, but not run that process
    • anyone should be able to make proposals though
    • can run a call for action to have someone drive it
      • @fricklerhandwerk: and then this group can delegate the responsibility and back that person up
  • @raitobezarius: we could invite leads of projects that have potential to become official
    • People affected by the change should drive this
    • Home Manager comes to mind
    • Could have proposals that will be accepted after a timeout when there are no objections
    • @thufschmitt: it will be easy for promoting, but hard from demoting
  • decision:
    • we agree on the spirit of the proposal, but need to talk to nix-community owners
    • @fricklerhandwerk will post a call to action to find candidates to take the responsibility

Hydra: propose the repo to be not user-facing

  • @raitobezarius: the infra team has needs towards the repo that are hard to implement
    • there is no maintenance going on
    • maybe we should have internal fork that will serve hydra.nixos.org?
    • @delroth: @zimbatm gave the infra team full access to the repo now
      • we can deal with our needs, but we don’t have visibility into external use cases
  • @raitobezarius: if we demote Hydra, how do we advertise to users that it’s dead?
    • @thufschmitt: Hydra has a slightly bigger problem as opposed to NixOps, because more big-time users likely depend on it
    • @raitobezarius: if there are big internal deployments by companies that did not give back, arguably we don’t have to care about continued support
      • such orgs are free to form a consortium to support Hydra development, but we currently don’t see them participating in the community
  • decision:
    • @raitobezarius will provide communication on the README and on Discourse about the state of maintenance
      • make recommendations for alternatives
      • @delroth will merge the PRs
    • let the infra team do what they need on some other branch

Follow-up on the teams transparency

Making this group more official?

  • @thufschmitt: should document what we’re doing in a more structured fashion so people know. ideas?
  • @delroth: this doesn’t feel team-shaped but rather assembly-shaped
  • @garbas: this is a meta team; having a team would imply being part of this higher-kinded team
  • no conclusion; ran out of time

I am sorry for not joining again for Geospatial. I keep overlooking this in my calendar. I made some adjustments to make sure I’ll be present for the next time.