A2.9

Coding and Cryptography
Part II, 2003

(i) Answer the following questions briefly but clearly.

(a) How does coding theory apply when the error rate p>1/2p>1 / 2 ?

(b) Give an example of a code which is not a linear code.

(c) Give an example of a linear code which is not a cyclic code.

(d) Give an example of a general feedback register with output kjk_{j}, and initial fill (k0,k1,,kN)\left(k_{0}, k_{1}, \ldots, k_{N}\right), such that

(kn,kn+1,,kn+N)(k0,k1,,kN)\left(k_{n}, k_{n+1}, \ldots, k_{n+N}\right) \neq\left(k_{0}, k_{1}, \ldots, k_{N}\right)

for all n1n \geqslant 1.

(e) Explain why the original Hamming code can not always correct two errors.

(ii) Describe the Rabin-Williams scheme for coding a message xx as x2x^{2} modulo a certain NN. Show that, if NN is chosen appropriately, breaking this code is equivalent to factorising the product of two primes.