Rotation3 could use some clarification

[milibopp]

There are lots of conventions as to what constitutes Euler angles. I think it would be good to make the convention used in cgmath explicit in the documentation of the Rotation3 trait (maybe adding some neat diagrams with axes and angles to visually convey the sense of rotation?). Another idea that goes along these lines: the convention should be required by any Rotation3 implementation and asserting it sounds like a great property to test using quickcheck.

All this is probably low priority atm but I thought I'd open this issue after spending a bit of time empirically determining the convention.