It'd be nice if there was a `bootstrap` method (or `pbootstrap` function) that supported distributing the `for`-loop over multiple cores.