[blog post] Proposal for a staged RFC Process

[illustrious-you]

@Ixrec

So I had a mildly crazy idea the other day. The thing I like most about explicit stages is the clear separation of crystallizing a problem statement from designing a specific solution (we get more than a few RFCs that feel like a “solution in search of a problem”). But having a separate repo for every RFC feels like far more fragmentation than we really need.

This extends the idea I had about RFC refinement quite nicely:

@illustrious-you

I wonder whether the stages would be for specific RFCs or for each repo. Many RFCs are refinements of topics covered by existing RFCs. In a repo-per-RFC world, I imagine the refinements would be added to the repo of the RFC they refine. In that manner, the “OP RFC” may be at the “Evaluating” or “Stabilized” stage, and the refinement effectively begins at the “Draft” stage, with spitballing and proposals taking place organically during implementation.

And would also fit the model envisioned in the Language Warts RFC repo thread.

I also think it would be of great assistance in reconciling the various approaches to error management, especially with regard to development of the interaction of faults and their handlers (e.g. throw/catch).