Given the problems for the principle of indifference, a lot of bayesians favor something more "subjective" with respect to the rules governing appropriate priors (especially in light of Aumann-style agreement theorems).

I'm not endorsing this manuever, merely mentioning it.