Since it was not rejected, courts should do as the Third Circuit said (if not as it did) and exclude any testimony claiming that the error rate of latent print identification (or, for that matter, anything) is zero because of the extreme danger that fact-finders will give it credence.