So you are arguing that professors shouldn't test their students (the only testing that should exist is external). I vehemently disagree. You've assumed that testing is solely about evaluation - separate from learning - when in fact it is about both. Students learn better when tested, and teachers teach better when they understand how students are thinking.

In your driving school example - testing and learning are not separate. The instructor is in the car with you - he is constantly testing you and giving you advice on what to improve. Your system would take this feedback mechanism away from college professors. If you want to leave internal testing in place but also have external testing for actually determining student grades, that would an idea worth fleshing out* - but removing internal testing would be a disaster.

But more importantly, I'm not seeing the problem that you're trying to solve. You've presented some theoretical reasons for why this system is bad, but not much in the way of specific problems the system is causing. You've sort of alluded to a general - improving educational standards - idea, but I'm not sure what this means in principle. Is it increasing course loads? Improving the relationships between professors and students? Reducing grade inflation?

*Actually I still dislike this system. In classes based around a constant flow of work - exams, problem sets, projects - your record throughout the class is probably a better indicator of your performance than an external test. Also it doesn't work for ... actually most classes. It works for large lecture classes and that is about it. How would you go about externally testing students on Medieval Chinese poetry? Or Computational linguistics?