This bin-pack problem is extensively used as an example for approximation algorithms in NP-complete problems. You should be able to find the explainations and proofs (for both FF and FFD) in a reasonable college computer algorithm textbook, under NP-complete section, but it might not be covered in an undergraduate algorithm course.

Briefly put, given an optimal solution of value O to an NP problem, and an approximation algorithm (in polynomial time) that outputs a solution that is k*O on average, then this approximation algorithm is called a k-approximation.

If I remember correctly, FFD is an 3/2 approximation algorithm.