Sorry about this but I thought the answer from radcan() was wrong for all x
> 0, but it isn't. I double checked, it is wrong only for 0<x<1. Now that
I understand it better, maybe I agree that the best simplification comes
from ratsimp(%) or even better scanmap('factor, %). I guess radcan() is
okay the way it is, but I never saw this issue come up before so I know now
to be careful with radcan().
I don't know if there are any issues like this with ratsimp() or
scanmap('factor, %).
Sorry about the noise. That is all.
Rich