Just to piggyback on to what the other answers have suggested, this technique is most commonly used (in browser developement) when you "namespace" some JS code. Namespacing is useful because it helps the developer reduce global variable pollution by adding all their code under one namespace under window
.
This also helps the developer modularise the code.
Here's a pattern you might see from time to time that adds code to a single namespace under window
.
// This first function is what's known as an Immediately-invoked
// function expression (IIFE). `window` is passed in as an argument
// to the function - all other declared variables are local to the scope
// of the function so you don't get anything leaking to global.
// PS, the semi-colon is there before the opening parenthesis to prevent
// any errors appearing if you minimise your code
;(function (window) {
// Here we declare something new to add to the namespace
// It could be another object, a string, an array, or a constructor
// like Model
function Model() {...}
// If namespace `window.app` doesn't exist instantiate it with
// an empty object. If it _does_ exist it probably means that another
// module like this one has already been added to the namespace.
window.app = window.app || {};
// Then assign your new module as a property under that namespace
window.app.Model = Model;
})(window);
You can then use it later something like this:
var model = new app.Model();
var game = new Game(model);
For further reading: Addy Osmani on namespacing.