Thanks. I think it’s most fair to count what a language has without extensions but thanks for the correction. To that end, Haskell basically has Dependent Types now too if you pile 10 extensions together (singletons, linear types, and others) and squint a little. It’ll easily be the first production-grade language to do so.
That’s true and they’re great for prototyping but I’ve fallen in love with Purescript for that domain. So I’m a bit biased.
Typescript doesn’t have linear types. Not even Haskell has linear types (correction GHC has had LT by extension since 9.0). Give Rust some credit for having linear types (AKA borrow checker).
Ps. Typescript doesn’t have dependent types and their type system is a joke even compared to Purescript. Here’s a little writeup about it: https://lemmy.dbzer0.com/post/66729809
The features Haskell has been “trying to copy” from TypeScript are, without exception, features TypeScript copied from type theory research that predates TypeScript by decades – row polymorphism from Didier Rémy’s 1989 work, untagged unions from the intersection type literature of the 1980s, type-level computation from Martin-Löf’s 1975 intuitionistic type theory – and what you are observing is not Haskell enviously watching TypeScript and taking notes, but rather two languages drinking from the same well of ideas, one of which is doing so with a formal semantics and a proof of soundness, and the other of which is doing so while standing in a JavaScript runtime and hoping no one looks down.
deleted by creator
Thanks for the comment. Thanks for the intellectual honesty.
I think the only reason I’d have chosen Typescript over Purescript would be if my employer would have forced me.
deleted by creator
deleted by creator
Neural Amp Modeler. Free, open source, and objectively superior to anything out there from any closed source company.
deleted by creator
deleted by creator
deleted by creator
deleted by creator
deleted by creator
deleted by creator
deleted by creator
deleted by creator
Of course, they have side effects. Side effects are what the computation does. Otherwise, they would simply make the box hot.
The difference between Haskell and most languages is that Haskell forces one to keep that impurity at the outer edges. It enforces that constantly, making programs easier to reason about.
I heard it explained succinctly lately: types are all of the possible inputs and inputs of a program while functions encapsulate the mutation. Excel, for example, is a purely functional language.