回复 第2楼 的 dingpeng：Yeah, I agree with you that one could (and probably should) tailor the code for each problem. But part of the reasons why Bayesian methods become more popular is the availability of WinBUGS, i.e., a software progam for general use instead for any specific problem. So, it is still desirable to have such a program. The question is then, among such general programs, do you like SAS Proc MCMC or not, and why.
