The tooling is surprisingly good for such a "young" language. I've had experience with other budding languages, like Nim, Zig, ReasonML throughout the years(not comparing these languages, just their tooling, they have tradeoffs), and Unison is slick. But expectations should probably be managed, it's not IntelliJ or VS Studio.

I'd describe the tooling as "zen like". When I sit down on weekends to write some Unison, I'll pop up my editor in a left tab, and on the right there's ucm. Since the LSP is included in the codebase manager, I didn't have to setup much. The LSP works good, could use some more code actions to control ucm though.

I'll write some code, throw in a `> ` eval expression to quickly mimic a REPL. I'll see the evaluation on the right. Eventually I'll switch to the right side and add my pieces. or write a `test>` and expression, which I can add as well to my codebase. Push my changes, create a branch, or switch project also happens. Maybe I'm not sure what the name was of a piece of code? Or I'd like to be able to search based on function signature. Both can be done in the right tab(ucm) with the command `find`.

Documentation is first class, I can browse the doc comments attached to variables/functions/types, I can link other terms(pieces of code) in those docs and they are discoverable as well, as soon as you add it to your codebase.

I believe a big part of why this is all possible, is that when your code is stored like this, it makes tooling around it a lot less complex. No need make ad-hoc parsers, compilation caches, syntax checkers. I'm no tooling expert, but for example making a visual editor based on the terms in your codebase sounds a lot easier than making a visual editor for TypeScript, the mental gymnastics is a lot less.