At the heart of all programming is the idea of [decomposition], where a complex problem is broken down into a series of simple steps. [Method chaining]—where a series of functions operate on a single object that gets transformed step-by-step as it passes through the function chain—is a common way of achieving this. The jQuery library uses this technique. Currying is a similarly-powerful way of decomposing a problem through chaining, which comes from purely functional languages, where—unlike in JavaScript—objects are not used to maintain state.
A curried function is a function with two or more arguments that has been transformed into a function to which each of the original arguments can be individualy applied successively in a chain to get the same result.
f(a, b, c) -> f'(a)(b)(c)
For this to work, the new function f' must take a single argument and return a new function that also takes a single argument.
Crucially, the returned function will need to have access to the argument applied to the original function, in order to make use of it in the returned function.
Currying is a special case of the partial application of a function, where some of the arguments of the original function have been applied and a new function has been returned that can take the remaining arguments and compute the original function with all of the required arguments.
In ES6:
function curry(f, ...args) {
[ to do]
}
In ES5:
function curry(f) {
[ to do]
}
[Decomposition] | [Closures] | [Chaining]
- What is JavaScript Function Currying? by John Lindquist (video)
- John Resig on Partial Application in JavaScript *
- currying in JavaScript
- A Beginner’s Guide to Currying in Functional JavaScript
- Currying on Wikipedia
- Partial Function Application is not Currying
- What's the difference between Currying and Partial Application?
- ECMAScript 6 and Currying