I get the idea. However, I would be severely challenged to actually measure any of that. To me that means that "only a good programmer knows another good programmer", which is a hard situation to start from when you are trying to reward good programmers.