featurep(e, real)



There are definitely other cases of polarform which could be improved:

    polarform(abs(a)) => is a pz?
    polarform(a^2) => is a pnz? n => a^2 * exp(2*%i*%pi)

The second case of course has nothing to do with using conditional
expressions or not (though simplification would be messier with them...).

The first case I suppose could simplify

   (if a=0 then 0 else a)
to
   a

          -s