A secure protocol for the oblivious transfer. (Extended abstract) (Q1924198)

The article represents the ``official'' publication of a well-known but for years unpublished paper that constitutes the first zero-knowledge proof of knowledge. In fact it deals with an improvement of Rabin's well-known protocol for the oblivious transfer. First, Rabin's protocol is described and conditions under which one of the communicating parties is provably unable to cheat are discussed. It is concluded that under an assumption originally considered by Rabin there it is still possible for one party to cheat and to obtain some extra information. Even if it is unclear whether knowledge of such information is of some advantage, it still represents potential flaw in the protocol. To remedy the flaw, the addition of one step to the original protocol is suggested and its realization by a new (sub)protocol is given. The protocol forms the first known example of zero-knowledge proof of knowledge. However, its formal proof of correctness is ``left to the final paper''.