Fast and reliable programs with OCaml
Static typing, type checking, and low memory usage make this easy-to-learn language a strong candidate for application development.
OCaml is a statically typed programming language with a focus on correctness and speed. First released in 1996 as "Objective Caml," OCaml extended the older Caml language with object-oriented features. OCaml  is now mature, stable, and well documented.
It supports both functional and imperative programming styles, making it a good choice for Python, C, or Java programmers interested in writing more reliable software.
OCaml's syntax can look a little alien at first. For example, here's a complete OCaml program:
let x = 3 let y = 4 let () = Printf.printf "%d + %d = %d\n" x y (x + y)
let declarations are evaluated in order (the compiler infers the types automatically). Expressions that don't return any meaningful value, such as
printf, return the empty tuple
(), which is called
unit, similar to
void in C or Java.
Assigning the final result to
() causes the compiler to check that the expression really does return
unit and you're not ignoring something important.
Printf.printf means the
printf function inside the
Printf module. Module names always start with a capital letter in OCaml.
Functions are also defined with the
let add x y = x + y
Unlike C, Python, and Java, OCaml does not use parentheses around arguments and does not have a
return keyword; it simply evaluates the function's body as a single expression. In OCaml, even an
then construct is an expression and produces a value, as in this factorial definition:
let rec fact n = if n = 1 then 1 else n * fact (n - 1)
When defining a recursive function, such as factorial, you must use the
rec keyword to allow the function to call itself. Without this, the name
fact wouldn't be in scope within the body.
OCaml functions really only take a single argument. A two-argument function conceptually takes its first argument and returns a new function to handle the second. This process is known as currying. These expressions are equivalent:
add x y (add x) y
add 5 is a function that takes an integer and adds 5 to it:
let add_five = add 5 let six = add_five 1
You'll often see partial application of a curried function to make code more concise. These are equivalent:
let hi x = printf "Hi, %s!\n" x let hi = printf "Hi, %s!\n"
In both cases,
hi is a function with type
string -> unit (it takes a string and returns nothing). Partial application is particularly useful with loops. For example, to print each item in a list, you can do:
List.iter (Printf.printf "- %s\n") ["foo"; "bar"]
The first argument to
List.iter is a function to be invoked on each item in the list. Partial application saves having to define a new function.
Curried functions allow some useful tricks. Consider the pipe operator, which can be defined as:
let (|>) x f = f x
This takes an argument and a function and calls the function with the argument. It works a bit like a pipe in the shell:
["foo"; "bar"] |> List.map String.uppercase |> List.iter (Printf.printf "- %s\n")
List.map produces a new list by using the given function to transform each item.
This code converts each string to upper case and then prints it. Can you see how it works? Hint:
x |> f |> g is interpreted as
(x |> f) |> g. The pipe operator can make function chains easier to read, and it's also a built-in operator in later versions of OCaml.
Buy this article as PDF
3ROS attack tool lowers the technical bar so anyone can be an intruder.
Mozilla's latest browser offers powerful new privacy feature
If attackers are on your system, saving your passwords in a password vault is no protection.
Faulty hash algorithm persists, despite efforts by experts to raise awareness.
Powerful man-in-the-middle attack is now targeting online shopping.
Another high-profile coder says the kernel team needs a kinder, gentler culture.
Bug database has a bug of its own that could allow an intruder to create an unauthorized account.
Report focuses federal resources on achieving universal Internet access.
Leading browser makers say “no” to porous encryption algorithm