This kind of thing has been on my fantasy wishlist for Nix for a long time. I am so glad to see experiments like this being done at a company that is dedicating resources to the Nix ecosystem with the potential to rapidly follow through if they see something valuable there.

With so many fruitful experiments in the Nix ecosystem being carried out by people who can dedicate more time to Nix than ever, I increasingly feel like I'm wasting my life by not seriously working on or with the Nix ecosystem as my main daily occupation.

I know there are still unresolved difficulties and growing pains, but later seems like such a stupid time to really dive in, for people who already know Nix and are already excited about it.

I'm kind of surprised this didn't link to the Nix RFC that considers changing the default builder from Bash to Oil: https://github.com/NixOS/rfcs/pull/99

Of course it does have a different goal (of replacing Bash with OSH instead of only offering an alternative), but it has a lot of good discussion.

Personally I think having alternative builders available is a great idea to improve the developer experience here.

I'm very happy alternative builders such as nushell and Nickel are being worked on.

Even happier that concurrently the other side of Nix DX that is documentation is very actively being worked on.

Nice to see an alternative to dealing with bash, though was wondering why a shell based on a proper language like python wasn't considered

default links bash with well know linter and one may make bash or cp other bash into nix and tune. also with nix it is easy to add alternative cli tools like sad, rg, etc.

if i want to be heavy and code real stuff, in nix it is easy to build and run any language. many things people do in bash to glue, i glue using nix declaratively.

so imho until other alt shell wins without nix, will use ugly weird non scalable bash.