It was suggested that meeting minutes be taken for the weekly Stan meetings and that they be shared. These meetings are open to the community, 11am New York City time, held on Google hangouts. The link for the next meeting is: https://meet.google.com/gzm-wmum-pfm?authuser=1. I don’t know if that link will remain valid for subsequent meetings.
August 9, 2018
Decision: Somebody needs to take minutes of Stan meetings to share with the broader community. Breck voluntered.
Individual reports: Individual reports were already done by the time the decision was made, so no notes.
Decision: The Thursday meeting is not a good venue for detailed technical discussions.
Decision: The meeting is good for introducing ideas, getting buy-in, and so on. It sets the stage for finding people to help or those that want input that will happen independent of the meeting.
Decision: The meeting is a good forum for expressing ideas of general applicability. Some concern that not all parties are present, reinforcing the idea that minutes might be a good thing.
Meeting moved to a techincal discussion:
Ben Bales: ODE senstivities, adjoint sensitivity analysis, should talk to Sebastian. ODE’s with a bunch of states is slow. Arya volunteered to help.
Andrew: Requested control over how verbose output is in RStanArm. Resolution is to supply null writer to Stan from RStan???
Andrew: Raised the point of how to handle business category ideas? This includes consulting, help on commercial basis. Resolution is to send to Stan Governing Body once it is up.
Ben Bales: Raised the issue of the complex numbers pull request that got stalled by inadequate testing. Resolution: Cool thing to do, we will press on with writing the unit tests to hopefully get the work merged.
The complex numbers case raised issues with policies with the math lib. Resolution: We need policies that are clear that we can be upfront with people with.