3 ms·

Isn't crates.io owned by The Rust Foundation?

by brunoqc 3y ago

SkyMarshal 3y ago
"crates.io-like"
- brunoqc 3y ago
  Would it make sense if crates.io was owned by a 3rd party? I don't think I would use it. So shouldn't flakehub be owned by The NixOS Foundation instead of being a 3rd party and proprietary? It seems to be a critical feature for flakes.
  - tazjin 3y ago
    Considering that flakes aren't a critical feature for Nix, it balances out.
    - brunoqc 3y ago
      It's not? I thought that the NixOS creator calling flakes "the future of Nix", implied that it kinda was.
      - tazjin 3y ago
        A thing with a clunky UI that technically only restricts the capabilities of its parent tool is not the future. It's a tarpit for staying in the local optimum that most Nix use is in right now.
        Macha 3y ago
        I don't see Flakes as a decision that's impossible to work past more or less than traditional nix, given that you can use flakes from traditional nix (not as easily as overlays, I will give you that), and frankly refusing to take a real improvement for end users over a hypothetical better one that no one has managed to produce yet seems counter productive.
        tazjin 3y ago
        You're talking about a different thing. I'm not interested in Nix for end-users who maintain office suite installations and Steam on their desktop computers. By all means, if somebody wants to make a great UX improvement for that - go ahead, but it doesn't need to be in the core of Nix.
        pxc 3y ago
        > It's a tarpit for staying in the local optimum that most Nix use is in right now. I know that pre-flakes conventions for reusing Nix code are pretty open-ended, and naturally 'pure evaluation mode' means turning a bunch of things off. I can also see how if you are interested in integrations with stateful systems or maybe relaxing certain constraints or allowing/tracking some 'impurities', that might also seem like a step in the wrong direction. Do you have any possibilities (or examples) in mind for how Nix might be used beyond the 'local optimum' of most current usage?
    - rmorey 3y ago
      I don't think Nix really has a viable future without flakes
  - SkyMarshal 3y ago
    Oh I see what you mean, I misunderstood your comment. Yeah I'm not sure what would be best in this regard. Flakes are on track to becoming a critical feature for Nix, though I'm not sure Flakehub will be a critical feature for Flakes. It seems there are some issues with standardization of pinning and versioning that need to be fixed in upstream, but the Flakehub folks seem committed to transitioning to those when fixed: https://discourse.nixos.org/t/introducing-flakehub/32044/14 https://discourse.nixos.org/t/introducing-flakehub/32044/14 As long as those standards get worked out upstream and Flakehub and similar projects follow them, and they're technically interchangeable, then they may be ok. It's if there are a bunch of different proprietary solutions with different proprietary standards where it becomes problematic.

Isn't crates.io owned by The Rust Foundation?

SkyMarshal 3y ago

"crates.io-like"

brunoqc 3y ago
Would it make sense if crates.io was owned by a 3rd party? I don't think I would use it. So shouldn't flakehub be owned by The NixOS Foundation instead of being a 3rd party and proprietary? It seems to be a critical feature for flakes.
- tazjin 3y ago
  Considering that flakes aren't a critical feature for Nix, it balances out.
  - brunoqc 3y ago
    It's not? I thought that the NixOS creator calling flakes "the future of Nix", implied that it kinda was.
    - tazjin 3y ago
      A thing with a clunky UI that technically only restricts the capabilities of its parent tool is not the future. It's a tarpit for staying in the local optimum that most Nix use is in right now.
      - Macha 3y ago
        I don't see Flakes as a decision that's impossible to work past more or less than traditional nix, given that you can use flakes from traditional nix (not as easily as overlays, I will give you that), and frankly refusing to take a real improvement for end users over a hypothetical better one that no one has managed to produce yet seems counter productive.
        tazjin 3y ago
        You're talking about a different thing. I'm not interested in Nix for end-users who maintain office suite installations and Steam on their desktop computers. By all means, if somebody wants to make a great UX improvement for that - go ahead, but it doesn't need to be in the core of Nix.
      - pxc 3y ago
        > It's a tarpit for staying in the local optimum that most Nix use is in right now. I know that pre-flakes conventions for reusing Nix code are pretty open-ended, and naturally 'pure evaluation mode' means turning a bunch of things off. I can also see how if you are interested in integrations with stateful systems or maybe relaxing certain constraints or allowing/tracking some 'impurities', that might also seem like a step in the wrong direction. Do you have any possibilities (or examples) in mind for how Nix might be used beyond the 'local optimum' of most current usage?
  - rmorey 3y ago
    I don't think Nix really has a viable future without flakes
- SkyMarshal 3y ago
  Oh I see what you mean, I misunderstood your comment. Yeah I'm not sure what would be best in this regard. Flakes are on track to becoming a critical feature for Nix, though I'm not sure Flakehub will be a critical feature for Flakes. It seems there are some issues with standardization of pinning and versioning that need to be fixed in upstream, but the Flakehub folks seem committed to transitioning to those when fixed: https://discourse.nixos.org/t/introducing-flakehub/32044/14 https://discourse.nixos.org/t/introducing-flakehub/32044/14 As long as those standards get worked out upstream and Flakehub and similar projects follow them, and they're technically interchangeable, then they may be ok. It's if there are a bunch of different proprietary solutions with different proprietary standards where it becomes problematic.