I'm curious to know whether other packages besides ADMB take the uncertainty in the estimation of the other parameters when calculating the estimated covariance matrix for the posterior distribution of the random effects.