> would it be difficult to decouple the assessment of an instructor from the test results of his/her students?

In fact, I think it would be critical to do so. Students may perform well on the exam for any number of reasons other than having a good teacher, and they may perform poorly for reasons other than having a bad teacher.

> If the assessment of a student's learning comes from somewhere else, what other mechanisms could we use to evaluate the effectiveness and (or) competence of their instructors?

Well, it's not like we can use exam scores now to evaluate the college-level teachers. The sad truth is that to a large extent, most people who teach at the college level are evaluated primarily on things other than teaching. I know I'm a pretty good teacher not because of good test scores, or good evaluations, but because students come back to me months or years later and talk about things they learned in my class(es) that set them up well to find their way through a later class or a job. But that kind of feedback is very difficult to quantify.

I saw a suggestion elsewhere on hn that moving to a single, cumulative exam after 4 years can be very effective because it's almost impossible to pass by just cramming the week before and it allows students and professors to focus on learning for most of the students' time in undergrad instead of on passing a test or teaching to a test.

This is similar to how the system at Oxford works (I think this is how the rest of the British universities do it as well.). Your degree class (essentially your GPA) is solely based on tests you take at the end of every year of your degree (three for a BA, four for a MComp). There are practicals (lab sessions) and tutorials (conversations with tutors about problem sets), but those don't affect your degree class, unless you do terribly.

With all the testing in American universities, the British (?) system is quite refreshing.

I believe this is the case for most British universities, most "traditional" universities anyway (not sure about former polytechnics, etc). Sometimes there is also coursework that contributes to your final degree (my CS course had a group project in the second year and an a dissertation based on an individual project in the final year), but all the exams take place at the end of the year.

My uni also weighted the exams so that the final year contributed much more to the final "grade" than the first year, effectively putting most of the pressure on the final term of the last year.