This chapter contains two views of the future. Since the validity of these views cannot be completely tested until the time period covered by the forecast has passed (so that predictions can be matched against actual events), how can we ever hope to establish in advance which view is better? What criteria might be proposed for evaluating predictions?