Should you represent the `i` as an int or should you represent it as a double?

Depends on what you do with it!

Say that the body of the loop is just `foo(i)` and you don't know what `foo` is at compile time. Whether it's better for `i` to be an int or a double depends on what `foo` does:

- If `foo` uses `i` as an array index, then you absolutely want `i` to be an int.

- If `foo` uses `i` for float math (say, multiplies it by 1.5), then you absolutely want `i` to be a double.

- If `foo` does a mix of those two things, then who knows. It depends on what happens more.

So, even in your simple example, it's not obvious what types to use when compiling JavaScript.