All,
It might seem odd but mission statements really matter. A mission statement line like “Stan fits the hardest Bayesian models” helps clarify decision making in the SGB and TWG (Technical Working Group), e.g. we should spend money on extensions to the math library, support inference research and focus on scaling. Beginner and intermediate users are less of a concern, we are serving advanced users.
A statement like “Stan is THE language for Bayesian modeling” conveys that we should be concerned about beginning and intermediate users and how many people use Stan vs PyMC3, etc…
Once the SGB has adopted a mission statement it then can help craft the TWG goals which then influences the technical roadmap. Further impacts are how we spend NumFOCUS funds and so on.
As exec dir of the SGB I am forming a committee open to everyone in our community to draft some mission statements. The end result hopefully be a few options that will be presented to the SGB for consideration. I expect this to be a short, intense committee that lives for two weeks managed by a Google group. You can join/leave at any time until the committee is terminated, which ends April 4 unless we need to extend it. Email breck@mc-stan.org if you wish to join.
Just to get the juices flowing I’ll throw a mission statement out there:
“Stan is a Bayesian modeling language with state of the art inference, validation and math infrastructure in service of science, particularly the hard parts. Our greatest asset is our community. We believe that Bayesian modeling will become the dominant language of science and Stan’s role is to hasten that inevitability because society needs it. Our user base spans beginners to experts and we support them all.”
Hate it? Have a better idea? Join the committee, please do not comment on it here.
Breck