That could be done but it doesn't seem useful enough to bother with. The `im` constant is really just for notational convenience and doesn't make anything faster.
On Tuesday, January 12, 2016, Jeffrey Sarnoff <[email protected]> wrote: > complex.jl defines im as Complex{Bool}(false,true). Is there a reason > that re is not defined as Complex{Bool}(true,false}? > > If I were to build a parameterized collection of Quaternion types > (different from Quaternions.jl), including Quaternion{Bool} and > Quaternion{FloatNN}, > is there an advantage obtained by defining the Quaternion `imaginary` > units i,j,k as > qi=Quaternion{Bool}(false,true,false,false}; > qj=Quaternion{Bool}(false,false,true,false}; qk > =Quaternion{Bool}(false,false,false.true}? > Would some of the op(Quaternion,Quaternion) math run faster (or jump > higher)? > > > > > >
