A FAIR Case for a Live Computational Commons
summary
Proposal for a planetary-scale live computational commons called Fairground that stores thousands of executable artifacts, runs them with a distributed scheduler, and enforces FAIR defaults. Implements a purely functional subset called Fair Python plus a strict FFI and a foreign data interface so large remote datasets stay in place while typed interfaces and reproducible build environments (Nyx mentioned) let the runtime control side effects, support incremental parallel typechecking, and execute only necessary recomputations.
AI-written summary · about
chapters (8)
- 0:00FAIR and scientific workflows
- 3:52Problems with software tools
- 7:22Fairground computational commons
- 12:22Orchestration language design
- 16:32Foreign function and data interfaces
- 20:20Liveliness at scale in Fairground
- 23:10AI in the commons
- 24:13Q&A
