Consider a trivial example where two semi-honest parties want to flip a biased coin. Since the sender is semi-honest, the sender can locally flip a biased coin and send it to the receiver. A natural question is: What does the sender's random tape look like in the above trivial real protocol?
I know that, in the 2-party computation literature, the random tape is assumed to be uniformly random. I am curious about the convention of such an assumption. Is it a mandatory requirement? Many thanks for your comments.