Shannon put forward the concept of perfect secrecy and proved that some kinds of cryptosystems are perfectly secure. The paper analyzes Shannon's proof that some kinds of cryptosystems were of perfect secrecy and points out that Bayes' theorem was used mistakenly in his proof because of his mixing up the probabilities under different conditions. An example is given to show that one-time system is not perfectly secure and this leads to a foundation for further study of cryptosystem's secrecy.
View on arXiv