Ok. How about a markdown doc on the repo so that the comments and discussion are kept with the final product. It would make it easier to go through line by line. It really needs it. All kinds of stuff isn’t really true.
Example:
If Stan-provided diagnostics do not fail then Markov chain Monte Carlo estimators of square integrable functions will satisfy a central limit theorem for sufficiently long chains.
The Stan diagnostics are not nearly powerful enough to support that sort of statement.