The problem with that, is again, spam. How do you differentiate fake votes from real ones? You can do per-IP limiting, and probably other things I'm not aware of (please comment about them), but it's still pretty open to abuse I'd say.
Sure, it could be exploited, and I'm guessing that's why they haven't implemented it, but there's got to be a solution that would make it work.
I think something like this would be a cool thing for Google to test for a year (a voting system might take a while to set up properly)