Why is it not available an optimized version of the Lambert W function (or Wright Omega function)?

I suppose that an efficient and accurate approximation has been proposed in one of the last DAFx conference, it would be nice to see it available in the dsp module