I'm not sure total lack of knowledge can ever be encoded into a prior distribution, and if it can, I think Jeffreys prior makes a more compelling case than a un

I'm not sure total lack of knowledge can ever be encoded into a prior distribution, and if it can, I think Jeffreys prior makes a more compelling case than a uniform prior. After all, to believe that all possible values of the prior parameter are equally likely, is to believe that all possible values of its square are not.