Also, a way of creating inner functions that won’t keep reference to the inner values is to use the Function constructor instead of the function keyword. Again, it is up to developers to choose when to use which.

Example:
function foo(e){
var i=0
return new Function(“alert(” + i + ” ” + e + “)”);
}

This statement is incorrect. Many languages have support for closures. Many have had it before javascript was created.

But still, I take your suggestion and will think about a good way to say it.

function createTimer(doSomethingLater, duration) {
// You also don’t need “window” – window just defaults
// to the global namespace. top, parent, content, and
// a few others do this too.
var id = setTimeout(doSomethingLater, duration);
return function removeTimer() {
clearTimeout(id);
}
}

This example is a bit contrived since removing a timeout is easy, but you can see how the logic behind the function is hidden, so you can just call the function instead of capturing the id and calling clearTimeout yourself.

The removal-example becomes alot more relevant when you’re dealing with stuff that’s complicated to remove since you abstract the removal, and if you don’t have any need to remove it, just ignore the return value, and it will be happily gc’d.

You can usually tell when you should be using closures if you’re finding yourself having a object pool somewhere to keep track of data you’d need to do something. When you do find this, closures may very well be your cleaner solution.

Another nifty thing that you showed, but didnt mention, is that variables are scoped by function, instead of the usual way of by code block. That is to say this code is valid:

function foo() {
var arr = [1,2,3];
console.log(i);
for(var i = 0; i