> > it's all better analyzed in terms of the cost of knowing and proving knowledge