Meeting Summary Notes

A quick bullet-point summary of the meeting notes.

Summary Topics from the Morning Session:

• What the "bottom-up" approach to a toolkit would look like.

• The experience of others -- both in economics and related fields (eg. so-called agent-based modeling groups with well-funded coding projects modeling the EU)

• Licenses for code - public domain, private-sector friendly, and related

• Lowering costs for policy-oriented researcher participation (eg. CFPB, OFR, IMF, Federal Reserve)

• Setting up incentives for cooperation -- using and contributing back, particularly models which may be useful to the policy community

• There was discussion of the CFPB experience, and general open-source experience, in producing tools and convincing users to uptake and participate in collaborative software tool development.

A common topic which arose repeatedly throughout the day was the need for a few people with a vision -- a working group, for example.

Some important results of the morning: a toolkit would be very welcome. The "bottom-up" approach would be largely an effort to produce an Application Programmers Interface (API), which is easiest to produce simply by creating a number of example models. [Another topic which arose repeatedly: we need examples to push forward.]

The mantra of "if you build it they will come" doesn't apply to open-source development.

For academia, however, it may instead be true that "if you cite it, they will come" -- discussion of simple, easy ways to set up citable code "vignettes" as IPython or IJulia notebooks occurred at the end of the day.

Summary Topics in Lunch Session:

• Open-source languages and tools which are ideal for collaboration. Languages Python and Julia discussed. Their "Notebook" framework is a robust and easy way for researchers to interact.
• Good practices for reproducible research.

Summary Topics in Afternoon Session:

• The need to "produce 10 models" with a top-down framework
• The need for researchers to be able to build a model in any framework incrementally, in an exploratory fashion
• The importance of researchers being able to access code directly

Summary Topics in the Late Afternoon Session:

• Connections to other organizations: Society of Computational Economics, Society for Economic Dynamics
• The importance of being able to direct funding towards researchers who contribute in particular ways
• Compiling a wishlist of papers
• Incentives for building core models
• Importance of concrete commitment from high-quality people. Eg: performance reviews.
• Importance of possible independent organization
• The idea of an "e-journal"