In Haskell, it is considered good practice to explicitly declare the type signature of your开发者_运维问答 functions, even though it can (usually) be inferred. It seems like this isn't even possible in OCaml, e.g.
val add : int -> int -> int ;;
gives me an error. (Although I can make type
modules which give only signatures.)
- Am I correct in that this isn't possible to do in OCaml?
- If so, why? The type system of OCaml doesn't seem that incredibly different from Haskell.
OCaml has two ways of specifying types, they can be done inline:
let intEq (x : int) (y : int) : bool = ...
or they can be placed in an interface file, as you have done:
val intEq : int -> int -> bool
I believe the latter is preferred, since it more cleanly separates the specification (type) from the implementation (code).
References: OCaml for Haskellers
In general, the syntax to let
-bind a value with a constrained type is:
let identifier_or_pattern : constraint = e ...
Applied to a function, you can specify the signature as follows:
let add : int -> int -> int = fun x y -> ...
This is analogous to the syntax required to constrain a module to a signature:
module Mod
: sig ... end
= struct ... end
精彩评论