| 29 May 2023 |
 infinisil | Frankly, the RFC process having some flaws is minor compared to actually resolving controversy in the content of RFC's | 16:00:44 |
 Robert Hensing (roberth) | I think it makes sense for a different group to decide FCP. It's an essential part of the process that gives legitimacy to the process and therefore the RFC. We shouldn't be able to push an RFC through by declaring FCP before resolving questions. Whether questions have been resolved reasonably can therefore only be decided by a separate group, which is the RFC committee. | 16:04:18 |
 @piegames:matrix.org | Robert Hensing (roberth) I think your statement strongly depends on the perspective and role in which you view an RFC's shepherds | 16:05:44 |
| Robert Hensing (roberth) | Changing that would speed up the RFC process by perhaps two weeks on average, at the cost of the legitimacy of RFCs in general. | 16:05:52 |
| @piegames:matrix.org | Because one could also argue that this precisely the shepherds' job. | 16:06:26 |
| Robert Hensing (roberth) | piegames yeah, the shepherd could take such an "independent" role, but I don't know if that happens | 16:06:36 |
| Robert Hensing (roberth) | clearly in our case there's some doubt about that | 16:06:43 |
| @piegames:matrix.org | Not that I fundamentally disagree with you, it's just that I think that the process should be clear about it either way | 16:06:53 |
| infinisil | Yeah I think it's the intention for shepherds to be this. The steering committe should pick shepherds of as different views as possible. | 16:07:00 |
| Robert Hensing (roberth) | I suppose the same thing could be achieved by having the steering committee approve FCP async, as FCP by itself has very little lasting effect | 16:08:25 |
| Robert Hensing (roberth) | It's RFC acceptance that matters | 16:08:36 |
| infinisil | I think effectively FCP means acceptance nowadays, because there's no reason to FCP if it wouldn't be accepted. So FCP = 10 days to acceptance | 16:09:51 |
| infinisil | * I think effectively FCP means acceptance, because there's no reason to FCP if it wouldn't be accepted. So FCP = 10 days to acceptance | 16:10:10 |
| infinisil | Oh and there's generally no major new feedback during FCP which would cancel the FCP | 16:11:06 |
| infinisil | (or has this ever happened?) | 16:11:21 |
| @piegames:matrix.org | RFC 127 🙃 | 16:11:37 |
| @piegames:matrix.org | The one that was pretty much silent until FCP | 16:11:50 |
| infinisil | Oh yeah was just thinking of that 😅 | 16:11:55 |
| infinisil | To be fair, FCP wasn't announced on discourse for that one | 16:12:03 |
| @piegames:matrix.org | That's kind of on the steering committee | 16:12:44 |
| @piegames:matrix.org | I would like to see more FCPs with disposition close though | 16:12:57 |
| @piegames:matrix.org | Most of the time it either stalls, or the author gives up and closes | 16:13:45 |
| infinisil | I feel like there should be an #rfc-process:nixos.org channel, also to ask the steering committee for actions | 16:14:59 |
| infinisil | * I feel like there should be an #rfc-process:nixos.org channel to discuss this, also to ask the steering committee for actions | 16:15:19 |
| @piegames:matrix.org | But this would require shepherds being more neutral towards the RFC. If we have this selection bias towards people in favor of it, this will continue to be pretty rare. Compare this to for example Rust and Matrix RFCs, where different people are responsible for the acceptance | 16:15:43 |
| @piegames:matrix.org | In reply to @infinisil:matrix.org
I feel like there should be an #rfc-process:nixos.org channel to discuss this, also to ask the steering committee for actions That's a good one, I'll put it into my forum post | 16:15:56 |
| @piegames:matrix.org | https://discourse.nixos.org/t/improving-our-rfc-process/28552 | 16:29:42 |
| infinisil | piegames: Thanks a lot! I'm watching the thread :) | 16:33:07 |
| @piegames:matrix.org | Good :) Feel free to add your own experiences to this | 16:33:32 |
| @piegames:matrix.org | While writing it a lot of the raised questions have been kind of resolved, but at this point I don't want answers, I want them answered by the RFC process itself. | 16:35:21 |
