The whole thing is here, but this is the good part:
One of the most amusing things Langer and others in his cohort claim is that polls should not be judged by their accuracy- i.e. PPP, Survey USA, and Rasmussen are bad no matter how good their polls are at nailing election outcomes because he doesn’t like the methodology. That attitude may still hold water in the small self appointed fraternity of elitist pollsters, but in our results oriented society most consumers of polls will continue to turn more and more toward organizations that have proven track records of accuracy.