rOpenSci Bayesian Modelling Software

Peer review of Bayesian software

With due apologies for cross-posting, rOpenSci has embarked upon a project to develop a system to peer-review statistical software. We are currently developing a few prototype categories, including Bayesian analyses, as part of which we are seeking feedback on our own discussion forum on how software for Bayesian analyses might best be assessed. As members of the stan community, you are all likely well-placed to provide constructive suggestions and advice to help guide our development.

How would this help stan development?

We are well aware of your excellent work developing your wiki as an aid to further community-led development of stan. Our system would be complementary to this, and will ultimately result in the development of procedures to systematically assess the quality of Bayesian software, including software built on top of stan. Your contributions to our discussion forum will help develop standards and assessment procedures for stan software, as well as help embed such procedures within broader ecosystems of software for Bayesian analyses and statistical software in general. rOpenSci is primarily about openness, and the system we ultimately be developed would of course be immediately able to be adapted and used by the stan community to provide analytic insights into, and standards for, stan-based software.

What can we do?

At this point, we encourage and welcome any contributions to our initial discussion on proposed means for assessing Bayesian software. We will then proceed from that discussion towards more concrete formalization of standards and assessment procedures, all documented in our living document – to which we also welcome any and all forms of contribution! We aim to begin accepting submissions for software peer review by the end of 2020, and are hoping for the contributions of the stan community to ensure our success. Please help!


Thanks for posting here. I answered in your discussion forum telling about the posteriordb we are building.