Stalling says:
OFB encryption can be expressed as
Cj = Pj⊕E(K, Oj-1)
where
Oj-1 = E(K, Oj-2)
Some thought should convince you that we can rewrite the encryption expression as:
Cj = Pj⊕E(K, [Cj-1 ⊕Pj-1])
Is he implicitly saying that
[Cj-1 ⊕Pj-1] = Oj-1
? If yes, why?