honestly, where NixOS shines for me is in my homelab. i don’t always have time to fully document what i’m doing, but my NixOS config is code-as-documentation for when work burns all of my memories away and has a git log and conflict management so i can manage multiple systems that share common config.

and once you find out you can have services run on systemd with syntax like services.jellyfin.enabled = true you’ll never want to go back to containers, although it has ways to manage those as well.

it’s overall a great OS for tinkering and deploying small services across small networks. not sure how it scales, but for my use case it’s damn near perfect

Nice write up, I enjoyed reading this! 15-20 years ago I would have loved nix and dumping hours into configuring every little detail.

I'm at the point now where I just want a distro I can install in 10-20 mins, copy over a couple of dot files, install a few apps, and be up and running.

recreational coding

Well, good news, it actually is fun to dick around in the Nix configuration and see those changes manifest on your system.

Interesting read, thank you. Also I recognise this type of procrastination: sit down to do one thing, end up installing a whole different OS :D

To address some negatives:

The dot files could be generated from your nix configuration (given that they are, and usually just symlinked). This way you still keep one source of truth. Only catch is any hardcoded nix store paths from pkg stringification. Also not sure if there's an easy way to do this, besides maybe automating some way of grabbing them from your nix build. (alternatively, keep your dot files and just have nix yoink them)

The "patch for nix" is a serious downside, admittedly. VMs and containers can get you a lot of milage, but properly nixifying things can hurt and be a big time investment.

Dev shells and missing deps is the same. Updating deps is usually okay but "repackage everything" starts to feel bad. But you aren't limited to nixpkgs for deps; you can pull in other managers (conan, vc, etc) for the project. Issues may arise if packages need or expect FHS / system stuff, then you're back to patching and pain.

The benefit is that when it's done, it's done; other than maintenance / upgrades. Reproducibility is like automation: the time and effort you put in, and the tools you use, determine the quality of the result. Similarly, the costs can quickly outpace the benefits.. Requires careful allocation and wise judgement.

Nix is a silver bullet, but silver is expensive and not everyone is hunting werewolves.

You have Arch in NixOS? I guess I don’t understand what NixOS really is, even if I’m interested in its ability to be reproduced from one computer to another.

So.... is it fair to say that it's like Ansible then?

Ok, Ansible has the config files on a separate control node, but for a multi-device setup, would that be comparable?