I would say it’s actually easier in many cases. Nix has really fantastic packaging tooling. You do have to learn a bit of the nix language, however (not become an expert).

The issue comes when trying to build from source. In most other distros, ou just follow the readme. In nix, you have to package it.

I wasn’t trying to go into typing as much as using structs or objects when working with known data attributes. Sorry that it was a bit misleading.

The original actually went into using trees, sets, heaps, tries, etc., but it felt way too… ranty. After writing all that out, I realized that most of those other cases come up really infrequently, and that my biggest gripe was about not using structs or other pre-defined key container types. I thought it would be better to keep things short and focused.

Maybe I should re-write and publish a data structures edition.

I love the addition of dataclass. Makes refactoring such a breeze. If you need to extract some function, boom, you already have a class that you’re using everywhere.

I’ve used pine64 boards for this. They have a few more options and are always available.

I feel like this is a trend with dynamic languages that have a REPL. I’ve done a lot of Common Lisp in the past, and had the same feeling.

The best way to get over this is to pop open the python REPL and start playing around with the options and functions. It takes very little ceremony to get a nice example rolling.

https://realpython.com/python-repl/ has some nice advice and tips on extra things you can do in the REPL.