John,
> A variation of this, which would be useful in some elliptic curve
> calculations, would be a function
> RR(x).nearby_rational_whose_denominator_is_a_perfect_square().
>
> For either problem, is there a better solution than going through the
> continued fraction convergents until o
A variation of this, which would be useful in some elliptic curve
calculations, would be a function
RR(x).nearby_rational_whose_denominator_is_a_perfect_square().
For either problem, is there a better solution than going through the
continued fraction convergents until one is found which has the
