Premature optimization is a common sin in software development. One of the reasons is that in the beginning of the process you don't even understand the problem fully. Heck, often you don't understand it even when you actually encounter it.
The solution is to have an iterative process, to work on problem at hand.
DHH's first distro was based on Ubuntu. It was the right decision at the time until it wasn't. Omarchy needed to be based on something more dynamic, more maleable. Arch was the solution. It worked well for some time, but with more time, more experience was gained about pros and cons. Maybe Nix is part of the solution, maybe staying on Arch and fixing problems there, maybe something else. Omarchy team will figure it out in time.
A year ago nobody could have predicted the level of success Omarchy would get. Heck, even a month ago nobody would.
But within a month the situation (popularity, money, support, ambition, plans) changed dramatically. Now with the changed environment, fixing the underlying Arch issues became relevant and timely. Without the resources, Omarchy would still be happily on Arch, not going anywhere. This once again illustrates why one should not do "premature optimization", worry about things that may (or likely won't) become relevant only in the future.
If we admit that iteration is part of quality, then you've basically just described why Omarchy keeps stubbing it's toe on embarrassing issues like default container misconfiguration.
Maybe Nix is part of the solution, maybe staying on Arch and fixing problems there, maybe something else. Omarchy team will figure it out in time.
This is the right attitude, and much better than entertaining a new base system on NixOS. The features of Nix and nixpkgs are modular and compatible with all the stuff they do on Arch. It's much easier to put Nix on Arch than it is to port your Arch plumbing to NixOS. My original "AI sycophancy" comment was total bewilderment that an LLM would rubberstamp an idea like switching the base system to NixOS while rejecting off-the-shelf solutions like home-manager. Nothing that Omarchy offers is worth keeping over the Nix equivalents if you're fully committed to a total distro migration.
Seems like some Nix users reached out to clarify some stuff, so we'll see if he avoids this particular iceberg. A lot of these misconceptions could have been avoided by skipping the AI-generated outline and talking to Nix users, but I guess hindsight is 20/20.
Comments
Premature optimization is a common sin in software development. One of the reasons is that in the beginning of the process you don't even understand the problem fully. Heck, often you don't understand it even when you actually encounter it.
The solution is to have an iterative process, to work on problem at hand.
DHH's first distro was based on Ubuntu. It was the right decision at the time until it wasn't. Omarchy needed to be based on something more dynamic, more maleable. Arch was the solution. It worked well for some time, but with more time, more experience was gained about pros and cons. Maybe Nix is part of the solution, maybe staying on Arch and fixing problems there, maybe something else. Omarchy team will figure it out in time.
A year ago nobody could have predicted the level of success Omarchy would get. Heck, even a month ago nobody would.
But within a month the situation (popularity, money, support, ambition, plans) changed dramatically. Now with the changed environment, fixing the underlying Arch issues became relevant and timely. Without the resources, Omarchy would still be happily on Arch, not going anywhere. This once again illustrates why one should not do "premature optimization", worry about things that may (or likely won't) become relevant only in the future.
If we admit that iteration is part of quality, then you've basically just described why Omarchy keeps stubbing it's toe on embarrassing issues like default container misconfiguration.
This is the right attitude, and much better than entertaining a new base system on NixOS. The features of Nix and nixpkgs are modular and compatible with all the stuff they do on Arch. It's much easier to put Nix on Arch than it is to port your Arch plumbing to NixOS. My original "AI sycophancy" comment was total bewilderment that an LLM would rubberstamp an idea like switching the base system to NixOS while rejecting off-the-shelf solutions like home-manager. Nothing that Omarchy offers is worth keeping over the Nix equivalents if you're fully committed to a total distro migration.
Seems like some Nix users reached out to clarify some stuff, so we'll see if he avoids this particular iceberg. A lot of these misconceptions could have been avoided by skipping the AI-generated outline and talking to Nix users, but I guess hindsight is 20/20.