Passphrase with Semantic Noises and a Proof on Its Higher Information Rate

Abstract

Key size becomes very important to a cryptographic algorithm according to Kerckhoff's law where a civilian cryptosystem shall depend fully on key secrecy. Currently, there are four passphrase generation methods: sentence, acronym, diceware, and coinware. Unicity distance is the minimum size of ciphertext for unique decipherability of ciphertext when number of spurious keys is zero. A key with size less than unicity distance is good where there are spurious keys which allow a protection method using limited unsuccessful logins. Here, stronger forms of passphrases using textual semantic noises like punctuation marks, mnemonic substitution, misspelling, and associative morphing, which improve the key entropy, are proposed. An ASCII mutual substitution table is presented together with its proof on information rate increment. Higher information rate has lower redundancy, and hence bigger unicity distance ensures encrypted keys the short cryptogram in a key vault, like password safe, cannot be cryptanalyzed within certain limited login attempts.