orthogonal polynomials relative to measures with mass point

Hi all,

I'm studying a class of orthogonal polynomials that are orthogonal relative
to a weight function on [-1,+1] complemented with a finite number of
discrete mass points (Dirac impulses). Is there some general theory written
down in a paper (I have Nevai's book and a paper by Nevai and Chihara, but
I'm looking for something more practical, not so theoretical)? What I would
like to know is whether there is an elegant and stable way to evaluate the
polynomials (for low and very high order) in the mass point located that is
furthest away from the origin without knowing the location of this point
exactly (I have the three-term recurrence at my disposition). I can
calculate this point, but only up to machine precision and it is the latter
fact that starts annoying the calculations for higher orders....

Thanks for any pointers,
gert
.