At a university level, that was the original idea of the University of California Santa Cruz, which gave students written evaluations of what they had done well and poorly on in each course, rather than numerical or letter grades (though you could still fail a course). As a result, there were also no GPAs or rankings. My understanding is they eventually abandoned it precisely because external people wanted to see the grades and rank students: employers and grad schools like looking at GPAs. So even if pedagogically it worked great (I would need to read more to determine if that were true), external factors made it difficult to continue with.
Neat. I like the idea. As a prof a school with grades, I'm often disappointed by how grade-oriented, rather than learning-oriented students are, being much too concerned with what precisely is going to be on the test, how many points each thing is worth, etc. But it's hard to blame them personally for it, when the entire incentive structure of grades, GPAs, and class rankings is set up to encourage them to focus on it.
