Agree, the paper seems fairly balanced. Of course an informative priors requires justification and sensitivity analysis. But I think that principle should apply in the other direction too, such that a frequentist analysis is required to justify the relevant data that it ignores.
