The source tar ball and Windows installer for JAGS 3.3.0 are now available from Sourceforge. Binary packages for other platforms should be available shortly: see the JAGS homepage for details of how to get hold of a binary version for your platform.
- Comment
- Reblog
-
Subscribe
Subscribed
Already have a WordPress.com account? Log in now.
JAGS News 
Thanks! What Mac OS will it work on? OS X 10.7, 10.8…? I currently have 10.5 and am thinking how far up I can upgrade and still use JAGS through R2jags/rjags. Thanks!!
The Mac OS X binaries built by Bill Northcott will run on 10.7 and 10.8.
I installed it through Macports. The current version of rjags (3-7) seems to be incompatible with it — R crashes when calling jags.model.
The rjags 3-7 package is compatible with all 3.x.y releases of JAGS. This must be a problem specific to Macports
Yes, you are right. It turns out the crash is due to incompatible versions of compilers: Macports compiles jags using gcc-4.5 by default, while R compiles rjags using gcc-4.2. I solve the issue by forcing Macports to use gcc-4.2:
port install jags configure.compiler=llvm-gcc-4.2 configure.f77=/opt/local/bin/llvm-gfortran-4.2
If you are using the Macports version of JAGS then you may be better off using Macports version of R, rather than the CRAN binary. I don’t know anything about Mac but it makes sense to stay within the same ecosystem.
Have the Macports maintainer add -fPIC to the build flags.
For problems with Macports, you should write directly to Dan Meliza. When you do so, please explain what your problem is, in addition to your proposed solution. I also suggest that you try to make your request NOT sound like you are giving orders to a servant.
Edit: OK I can now see this comment in the context of the thread, where it seems slightly less abrupt. But you should still contact Dan directly. This is my blog, not his.
Thanks for the benefit of the doubt. I already tried to submit a bug report to the macports bugtracker (where it really should go, for the benefit of the Macports community), year(s) ago. This is a key combo (Macports R with Macports JAGS) and has never worked for me on a dozens of Macports installs since the port has been there.
(And now, please a little deeper for one more condescending, schoolmaster-like comment that properly belittles my frustrated efforts and that of the other users on the web who haven’t gotten rjags to work via Macports.)
In the meantime, please see my comment at John Kruschke’s website:
http://doingbayesiandataanalysis.blogspot.com/2012/01/complete-steps-for-installing-software.html
Thanks for writing JAGS. It it truly amazing and I’m not worthy of summing up it’s value to the statistics community.
Dear Martyn,
What has changed in this new version ?
This is entirely a bug fix release. If you want to see details, look in the NEWS file.
May I know if truncated multivariate normal distribution can be specified in this new version? Like we can use dtmvnorm in R ( in library(tmvtnorm) ) to get its density.
No, it is just a bug fix release. There are no new features.
Dear Martin,
When downloading JAGS (3.3.0) from SourceForge, my virus-scanner (AVG) reports that the windows executable has been infected by a trojan. There are no reported problems with the JAGS version 3.2.0 file – so what should I do? Disable the virus-scanner during installation?
Yes. In my experience AVG is hyper sensitive and so has a tendency to throw up false positives. I built this file myself and Trend Micro Office Scan finds nothing wrong with it. Follow the advice from AVG on their forums.
Is there any automated/silent install method for managed deployment of JAGS for Windows to a classroom environment?
Thanks!
You can call the installer from the command line with the “/S” flag to do a silent install. The optional “/D” flag can be used to specify the installation directory if you don’t want to use the default c:\Program Files\JAGS\JAGS-3.3.0. Note that the flags are case sensitive.
On my Windows 7 system, I find that when launching the installer from a windows share the “/S” flag is ignored. Hopefully you have better luck.