[colinsane]

i've spent two years of my life getting NixOS to work well on my mobile phone because i value things like privacy. i assure you, actually ditching Android/iOS on principle is in fact very much not convenient. i do not see the link.

[steve1977]

You are using a mobile phone based on a RISC CPU (presumably) and the Internet. Those are the conveniences I was speaking of.

[colinsane]

closed source CPUs are also a thing i have invested inconvenient amounts of time in trying to overcome [1]. i expect by the latter decades of my life it will actually be feasible.

i'll be honest i don't understand what point you're trying to make. if i owe some loyalty to the military for the conveniences of their products, then would not Anduril owe me loyalty for the convenience of my products? the actual request to Anduril/military is way less than that, btw -- it's less "don't use our product" as "don't advertise in our spaces".

1: https://git.uninsane.org/colin/fdtd-coremem

[steve1977]

> it's less "don't use our product" as "don't advertise in our spaces"

I guess the problem here is the definition of "our". If Anduril sponsors development of nix, it's as much "their" product and "their" space as it is yours.

Or in other words, they are included in "ours". You might not like that, but that doesn't change the facts.

[colinsane]

i don't disagree with this. it takes care to figure out how to actually share the commons. despite my very obvious (at this point) anti-state motivations for being involved in open source, i've actually kept that discourse away from the NixOS spaces for two years. it's only after Anduril is reshaping those spaces to be pro-military that i bring my own world views into this space overtly.

i'm not certain of the ideal approach. i would be content with an agreement that our shared spaces be neither overtly pro nor anti military -- to the degree which this is possible or enforceable. but it is extremely difficult to actually establish consensual agreement on that, and attempting to force it (in any direction) leads to the type of escalatory situation you're watching unfold now.

my most honest takeaway is that NixOS doesn't deserve to be some monolithic thing. communities grow and reshape into loosely connected smaller communities as a pretty natural effect of success every day, and do so peacefully. there are plenty of spaces occupied by people i don't get along with (say, hyprland, or lemmy.ml), and i simply keep my distance. but nixpkgs is a monorepo, with intense infrastructure needs that require a foundation/governing body to meet. for as long as all of our work is so closely linked to that governing body and brand, there's little way for us to arrange ourselves into the more socially intuitive structures which allow for that type of "live and let live" approach wherein we all flourish even without finding consensus.

[steve1977]

I think these are very fundamental issues with open source in general. Or even with "open anything" probably, including societies.

If you want something to be open, you have to accept it will also be open to entities and ideas you don't like. If you don't want that, you don't actually want "open".

One way out of this specific situation would probably be a fork, which I suspect is what will happen. But forking is interesting, because in one way it's only possible with open system and on the other hand it's kind of an admission that the openness has failed.

> i would be content with an agreement that our shared spaces be neither overtly pro nor anti military -- to the degree which this is possible or enforceable.

I live in Switzerland. In some ways, this is our official stance in regards to neutrality and I don't think it works very well. Some things are just binary and you can't really be neither pro or anti, unless you're lying to yourself and/or others.

[colinsane]

> If you want something to be open, you have to accept it will also be open to entities and ideas you don't like. If you don't want that, you don't actually want "open".

sure, but i worry people see a single label like "open source" and derive a book-long prescriptivist explanation for the thing instead of seeing that it's just an incomplete, fuzzy short-hand. i don't care about "open source". i care (among other things) that when my environment isn't the way i like it, i have the power to improve it -- and in the digital space being able to access/modify the code is how i do that. if one were to bring that interpretation to a physical space, people would apply a much different label to it, even if it meant the same thing to everyone involved. "open source" is just a label: not a map.

> One way out of this specific situation would probably be a fork, which I suspect is what will happen.

one final thing from me: i believe hyprland (mentioned earlier) is a fork of sway, and despite the lead devs of the two being rather at odds with each other... the projects in fact still collaborate, just at a distance, on a common upstream (wlroots). alternately, `rofi` is very adamant to remain X11-only, and there's a sustained fork of it for wayland. but because it's plugin-based, a very large amount of contributions happen in areas completely unaffected by this split.

"forking", in the colloquial duplicate-the-codebase sense is just the most callous manner by which to provide separate spaces for those who can't share a space. plenty of projects out there are full of the same type of disagreements we see in NixOS, and have effectively forked (i.e. provided separate spaces or brands where there otherwise would be only one) but either did so very early on (by adopting something like a plugin model) or did so in a behind-the-scenes manner (by splitting components into separate spaces). i hope that's not too wishy-washy: i believe the technical structures and the social structure of NixOS will eventually settle into some convergent equilibrium, and i believe the result will be better if we're intentional in that process.