Modularity for making Stan more Pythonic and Rthonic

[seantalts]

@Bob_Carpenter

My biggest complaint about the RStan and PyStann APIs is that they work through these monolithic Stan fit objects (that’s also in the roadmap going forward, so at most I’m lobbying for new functions here)

Wait, can you put these complaints intothe roadmap thread? You once asked a question there about whether fit objects were mutable and the reply was that they weren’t, but I didn’t know you had some further distaste for them. Would love to hear more about your views.