Support for types in the math/big package

[GCrispino]

Hello!

Would there be any interest in having support for the big.* types defined in the math/big package, such as *big.Float?

With the current implementation if we're using these types we can't use the functions in the present package without transforming these values to float64 values, which may not always be possible

Another option would be to make the methods in this package take an interface, such that any user could implement support to any numerical type as long as it implements certain functions necessary for them