Webpage validation is, in my mind, a complex matter. On the one hand, you have the W3C recommendation - that is just that: a recommendation - that might or might not (probably not) render equally in all browsers. On the other hand, you have your CSS/HTML tweaks and hacks that make each page render pixel-perfectly, but most probably doesn't validate with he W3C validator.
In real life, it's a world of compromises. I, personally, try to do both - have as few validation errors, but having the main emphasis on it actually looking good with widely used browsers.
But, in an academic life, I think it's entirely fair for the professor to require 100% W3C compliance. It is, however, the closest that we have of a spec for HTML/XHTML, and that's what the academic people are ultimately interested in - the theory.