Asked by: SIMON CORFIELD, BY EMAIL
Yes and no. The only provably unbreakable encryption method is the so-called one-time system, in which each letter of a message is turned into a number (say A = 1, B = 2 and so on). The result is added to a stream of random numbers known only to the sender and recipient, and only used once. Yet its absolute theoretical security can be undermined by blunders such as accidentally re-using a set of random numbers.
But even simple codes can be made unbreakable: the trick lies in only using them for a few very short messages. This makes it impossible to tell what the true decryption is. For example, XCFG could stand for ‘HELP’, ‘STOP’ or ‘PLUM’ – who’s to know?