Priors are pointed to by those critical of the Bayesian approach
as *the* major weakness of the theory.
Instead, Bayesians consider them
a crucial and powerful key point of the method.
Priors are logically crucial because they are necessary to make
probability inversions via Bayes' theorem. This point remains valid
even in the case in which
they are vague and apparently disappear in the Bayes' formula.
Priors are powerful because they allow to deal
with realistic situations in which informative prior knowledge
can be taken into account
and properly balanced with the experimental information.

Indeed, we think that one of the advantages of Bayesian analysis is that it explicitly admits the existence of prior information, which naturally leads to the expectation that the prior will be specified in any honest account of a Bayesian analysis. This crucial point is often obscured in other types of analyses, in large part because the analysts maintain their method is `objective.' Therefore, it is not easy, in those analyses, to recognize what are the specific assumptions made by the analyst -- in practice the analyst's priors -- and the assumptions included in the method (the latter assumptions are often unknown to the average practitioner).