Brian J Smith

@binarybana, nice job on the SAMC julia package.

@goedman: Yes, `DGS` is the discrete sampler describe in my earlier post (10/22). A quick tip for implementing the Jags example... the stochastic nodes are k (the response), theta, and...

**Long-Term Update:** In addition to the [Discrete Gibbs Sampler](http://mambajl.readthedocs.org/en/latest/samplers/dgs.html), several binary samplers have been implemented recently in Mamba by @bdeonovic, including: - [Binary Hamiltonian Monte Carlo (BHMC)](http://mambajl.readthedocs.org/en/latest/samplers/bhmc.html) - [Binary Metropolised...

The pmap errors we are getting now in 0.5 do not occur in 0.4. Below is the 0.5 error, for what its worth. When I get some free time, I'll...

Great! Thanks for coming up with a reduced case. I'll keep an eye on your other thread and possible developments in julia to address this. A work-around on our package...

I'll look into it tomorrow.

Thank you for letting me know about the error. I've pushed a commit up to the master branch that should fix it. If you would like to clone the package...

In the toy example, `insupport(D, a, b)` would not be a valid test of `a` and `b`, because their supports are defined by their prior distributions rather than by `D`....

Compatibility with the Distributions package must be maintained. `insupport` is imported from that package, where it is defined as a test of whether a value is in the distributional support,...

The distributional constructors are in the user-defined Stochastic nodes. A node is the smallest model component that the package could place a try/catch block around, and nodes can include more...