Coefficient functions manipulate the fraction without affecting the exponent and sign, to support decimal-floating point conversion library functions.