The way priors are handled is a bit ad-hoc, and inflexible, e.g. no way to specify a birth-death prior if you wanted to in BEAUti, or log-normal prior on theta's.
Also, the gamma distribution is parameterised different (shape, rate) than the Gamma distribution in BEAST (shape, scale), which is terribly confusing for users.
The way priors are handled is a bit ad-hoc, and inflexible, e.g. no way to specify a birth-death prior if you wanted to in BEAUti, or log-normal prior on theta's.