(3) Consider a discrete memoryless source {X } with alphabet A = {a1, a2, a3, a4, as, a6} and probability distribution given by p1 = 10, p2 = To and p3 = PA = P5 = P6 = 10 10 Where Pi := P(X = ai), i = 1, . .. .6. (a) Determine the source entropy rate H(A ) in bits. (b) Design a binary first-order Huffman code for this source and compute its average codeword length and codeword length variance. (c) Design another (different ) binary first-order Huffman code for this source and com- pute its average codeword length and codeword length variance. (d) Which of the two codes would be preferable in practice? (Justify your answer.)