Defer/separate R code?

[Protonk]

Looking over the codebase I'm further convinced that the right course of action is to treat R and Python just like we would treat plumbing and porcelain. I don't know how rpy2 handles environments or scoping and if we're doing real modeling we need to know those things.

I'll take a close look this weekend at where this can best be implemented without totally breaking the flow of things but I think you and (eventually) the SL folks will be happy we've done so.