milesfrain
commented
4 years ago Here are some notes on ReasonML differences from reading their overview.
- Philosophy
- Less strict about immutability and effect tracking.
- Less FFI ceremony.
- More efficient JS (speed and size).
- Syntax
- Closer to JS than Haskell.
recrequired for recursive functions.- Looks like an alternate solution to CycleInDeclaration errors.
()is shorthand forunit.
- Records
- Record types must be declared before creating values of that record type.
- Records may have mutable fields.
- Nominal typing - Cannot use a record with a superset of fields in place of a record with a subset of fields.
- Record update uses spread syntax.
- Adding fields is complicated by nominal typing. Easier in PS with
disjointUnion.
- Functions
- Named function arguments built into language.
- Different than passing a record argument with named fields.
- Automatically uncurried whenever possible (even if this produces multiple functions).
- Built-in optional arguments.
- Technically uses an
optiontype, which is likeMaybe. - Possible to do with
purescript-option, but certainly a lot more ergonomic when built into the language.
- FFI
- One-liner. PS requires two lines in separate files.
- Type Classes
- Uses Module system (rather than type classes).
- Could mimic haskell-like type class hierarchy, but this is not common among libraries.
- Frameworks
The differences from Haskell doc in this repo is really helpful.
Likewise, this Elm to PS cheatsheet is great too, and we have permission (https://github.com/alpacaaa/elm-to-purescript-cheatsheet/issues/9) to absorb its content.
Proposing we include docs for comparisons to all of these languages:
There's an outdated section in the book's intro that touches on comparisons to other JS options. I was originally considering adding some notes on other languages there, but I think it would be best to have all this info in a common location.