In theory, the key length could be log2(2n)! bits. For example, assign each mapping a number, from 1 through (2n)! and maintain a table that shows the mapping for each such number. Then, the key would only require log2(2n)! bits, but we would also require this huge table. A more straightforward way to define the key is to have the key consist of the ciphertext value for each plaintext block, listed in sequence for plaintext blocks 0 through 2n – 1. This is what is suggested by Table 3.1. In this case the key size is n 2n and the huge table is not required.