Hacker News Reader

Comment by latk | Hacker News Reader

The argument that every function has arity 1 is technically true (this is the whole point of currying) but is not useful when definitions like "let a b c = ..." suggest other semantics. It's possible you've had a difference experience with this, but I tend to get confused when the semantic argument list isn't delimited.

There is nothing wrong with relying on the type checker, except that it tends to add cognitive overhead.

View on HN

In my experience, the more you use currying, the more intuitive it becomes (surprise, surprise). In any case, you very quickly develop an understanding that `let foo bar baz = qux` is just syntactic sugar for `let foo = \bar -> \baz -> qux`. Of course, if you want to simulate higher-arity functions, you could just use tuples. It's perfectly acceptable to write `let foo(bar, baz) = qux`.

Every function having arity 1 is reducing complexity. It's extremely uniform, and let a b c = ... is merely syntactic sugar for let a = (lambda b. (lambda c. ...)).

It's very natural, and as Lisp/Scheme/Racket shows, it's perfectly fine in a dynamically typed context as well.

"let a b c = ..." doesn't suggest other semantics. It's saying: "a applied to b, and then applied to c, equals ...". Note Haskell makes an effort to have the LHS of definitions imitate the exact syntax of function application. Patterns use the same syntax as data constructor applications.

Reply on news.ycombinator.com