I tried for some time to work out a natural implementation of Mux using my brain, but it wasn't working. So i found the canonical representation for it. Being a three input gate this generated 4 sections each with three parts.
To implement this I could write a large (prob > 16 line) HDL description using intermediate variables to maintain the state of my partially evaluated terms. However i can only assume there is a better way not alluded to in the book (that I saw).
I've done very little maths so i'm not sure how to go about reducing my CR to a more simplified form.