I don't even understand how the frequentist view is a valid alternative. It always seemed to me like either you are honest about your priors, and use Bayesian logic to take them into account, or you sweep it under the rug. Lying to yourself always produces bad results, is my overriding heuristic. But I'm not good at math.