On 21/01/2008, at 10:00 PM, Roger Levy wrote: > Has anyone gotten WinBUGS to work on OS X through R2WinBUGS? I have > managed to get WinBUGS 1.4 to work on my Intel Mac (I run Tiger) > through > Darwine, but cannot get the bugs() command in R2WinBUGS to work. I > follow the directions in Andrew Gelman's example here: > > http://www.stat.columbia.edu/~gelman/bugsR/runningbugs.html
I did some post some messages about this a few weeks back on both BUGS and CrossoverMac discussion lists. (look for messages about winepath) As you discovered Winbugs will run happily in using WINE or the CrossoverMac packaged version of it. However, the WINE support in R2WinBUGS is thoroughly broken. It depends on undocumented and unsupported features of particular versions of WINE. These have been changed in recent versions of WINE and now that bit of R2WinBUGS needs a complete rewrite. I can see what needs to be done, I just need to find the time You can run BUGS models in JAGS which now has a working R interface package - rjags. I am hoping to get a fat MacOS binary done for this in the next few days and Martyn Plummer says he will put it on the JAGS site at http://www-fis.iarc.fr/~martyn/software/jags/. It is even possible to use the standard 32 bit R distribution and a 32 bit build of rjags to run a model on 64bit JAGS on a Core 2 Duo Mac. It would still be nice to have Winbugs available to generate a comparison with results from JAGS. Bill Northcott _______________________________________________ R-SIG-Mac mailing list [email protected] https://stat.ethz.ch/mailman/listinfo/r-sig-mac
