> If the point of the exercise is to allow public scrutiny of models then IMHO the code should not be changed "for public release". Some code I've used from sc

> If the point of the exercise is to allow public scrutiny of models then IMHO the code should not be changed "for public release". Some code I've used from scientific papers in the past made a git tag at the time of publication, and said "Use version XXXX and the provided sample data to get the results in the paper", or something of the sort. This could allow review and improvements post-publication.