If we assume integer representation for the mantissa of floating-point numbers, we encounter certain scaling problems during multiplication and division. Let the number of bits in the magnitude part of the mantissa be (n - 1). For integer representation: 

a. Show that if a single-precision product is used, (n -1) must be added to the exponent product in the AC. 

b. Show that if a single-precision mantissa dividend is used, (n - 1) must be subtracted from the exponent dividend when Q is cleared.