Mailing List Archive

my thoughts on RFC process so far
I concur with with the others have said, it's interesting that these emails are popping up and think it's a good idea. The following is only above and beyond what has been stated.

1. I am very happy and encouraged by it so far
2. there is no need for a pre-pre-RFC, this only causes confusion; rather I think an email asking for participation in the creation of a pre-RFC would suffice and then it could be handled off list
3. for pre-RFCs clear "yes", "no", or "not at this time" I think would be sufficient
4. clarification on who can determine #3 would be helpful
5. also helpful would be some "fairness" in #3, i.e., "the RFC body meets monthly to provide an answer on this month's pre-RFCs; agendaof the pre-RFCs being considered will be set X days before and provided ahead of time"
6. something like #5 would facilitate a single email that listed the pre-RFCs considered in a list, then simply the "yes", "no", "whatever", instructions regarding the now blessed RFCs, + a status on current RFCs, and any other coming deadlines (e.g., the end or state of an ongoing RFC), etc
7. encourage an ordered, fair process of RFCs for official ones - namely, use github PRs as the target for official "comments" - encourage one comment per author, though they should be allowed to revise
8. state that discussions on p5p, while relevant, are not collected or preserved in any meaningful way (this is my preference)

I'd also like to suggest that the primary sponsor of the RFC, if selected, become the defacto "head" of what amounts to an ad hoc advocacy committee for said RFC. They should be aware of comments and be continually refining the proposal - in addition to organizing the effort and research required to implement such RFC. I don't think anyone looks at an RFC as a letter to Santa - this means if the primary RFC sponsor doesn't deem themself to be technically capabile of implementing such RFC or "making it happen", they should have the wherewithall to recruit those who can to see any eventual action through. All of this is said with the idea that all of this, from pre-RFC to the conclusion of a "feature" should be a scalable process.

Cheers,
Brett