On Wed, Apr 11, 2012 at 12:22 PM, Holger Schulz <qdl at gmx.net> wrote:
> Okay. The cases I'm thinking of will have no variables in the solutions.
> Usually roots of polynomials and stuff like that. I think it should be
> possible to decide if these are real. My question wasN't precise enough in
> that point.
>
Consider the casus irreducilis <
https://en.wikipedia.org/wiki/Casus_irreducibilis>; case cubic polynomials.
The roots are real, but the solution from solve will contain cube roots of
complex numbers.
If you want numerical answers, then you can use allroots for bfallroots.
Then it's pretty easy to remove the complex solutions, assuming that the
complex roots don't have spurious non-zero imaginary parts due to roundoff
in the algorithm.
Ray