Valid points, but in this case (and many others where you encounter power-of-2 mult/div), I'd consider that a shift might actually semantically be the more natu

Valid points, but in this case (and many others where you encounter power-of-2 mult/div), I'd consider that a shift might actually semantically be the more natural operation in the first place, instead of an "optimization" of the mult/div operation. (With their equivalence being obvious to any reader, it might not matter.)