Abstract
In contrast to previous analyses, we demonstrate a Bayesian approach to the estimation of the CKM phase $\alpha$ that is invariant to parameterization. We also show that in addition to {\em computing} the marginal posterior in a Bayesian manner, the distribution must also be {\em interpreted} from a subjective Bayesian viewpoint. Doing so gives a very natural interpretation to the distribution. We also comment on the effect of removing information about $\mathcal{B}^{00}$.