As PyMC continues to mature and expand its functionality to accomodate more domains of application, we increasingly see cutting-edge methodologies, highly specialized statistical distributions, and complex models appear. While this adds to the functinoality and relevance of the project, it can also introduce instability and impose a burden on testing and quality control. To help address this, a pymc-experimental respository could act as a home for new additions to PyMC, which may include unusual probability distribitions, advanced model fitting algorithms, or any code that may be inappropriate to include in the pymc repository, but may want to be made available to users.

If implemented thoughtfully, a pymc-experimental repository could act as the first step in the PyMC development pipeline, where all novel code is introduced until it is obvious that it belongs in the main repository. This would improve the stability and streamline the testing overhead of the pymc respository.

pymc-experimental would be designed to mirror the namespaces in pymc to make usage and migration as easy as possible. For example, a ParabolicFractal distribution could be used analogously to those in pymc:

import pymc as pm
import pymc_experimental as pmx

with pm.Model():

    alpha = pmx.ParabolicFractal('alpha', b=1, c=1)
    
    ...

• newly-implemented statistical methodologies
• distributions that are tricky to sample from or test
• infrequently-used fitting methods or distributions
• any code that requires additional optimization before it can be used in practice

Since there is a lot of code that we may not want in the main repository, does it make sense to have more than one additional repository? For exmaple, pymc-experimental may just include methods that are not fully developed, tested and trusted, while code that is known to work well and has adequate test coverage could reside in a pymc-extras (or similar) repository.