I'm aware of it, in the example given I used float32. Anyway Max/Min was just example I used other functions from math as well.

Anyway my point was that with generics, they wouldn't need to copy anything, the math package would work with both float32 and float64 and many functions likely would also work on all integers.