Comments on the Bayesian Method for Estimating Reactor Core Melt Frequency