META: clarify when an RFCs issue is appropriate

[nrc]

Or rather when to use the following options for ideas:

• a discuss post
• an RFCs issue
• an RFC (i.e., a PR to the RFCs repo)
• an issue on rust-lang/rust
• a blog post

Noting that we have a process for closing RFC PRs but not RFC issues (for now).

[glaebhoerl]

Is it fine if I just continue to post the remaining things I wanted to post? (There are... eight, a couple of smaller features, mostly larger topics which have been discussed in other fora but don't have RFCs.) On the theory that there are already lots of issues like this here, migrated from the rust repo, and they can all be moved together if it turns out that they should be.

I want to point out (in part to inform the meta-discussion; in part as pertaining to the above) that if my alternatives are either to post to the discourse forum or to write an RFC, then I'm going to do neither, because both involve much more effort than I want to expend on these topics right now. (I don't want to have long discussions about them at the moment, and in the absence of that, posts on the forum fall away into oblivion pretty quickly. So it's here or nowhere.)

To quote my motivation from #373:

I don't want to have a design discussion about this right now in preparation for an RFC (I intentionally want to "get it off my plate" so I can work on more important things); I just want there to be a public record of the idea where others can see it and it doesn't get lost. I feel that a GitHub issue is more appropriate for this than the forum.

Basically I feel like the forum is more appropriate for either potentially intense discussions over a short time window in preparation for an RFC, or for relatively free-form brainstorming, while over here is for keeping track of things over a longer time period. I guess the words I'm looking for might be that the forum is for more "ephemeral" things and this is for "persistent" ones.

(Although I personally wouldn't mind very much if the forum were abolished and all discussion moved to non-PR issues in this repository. I also don't mind keeping it.)

[pnkfelix]

@glaebhoerl to me, an RFC issue is clearly appropriate when you focus on (1) a concrete problem (solely) that (2) everyone agrees is present in the language/stdlib. It is easiest to identify such cases when there have already been multiple RFCs put up as attempts to address them. (It would be good to define other "triggers" that clearly justify an RFC issue.)

An issue filed that jumps to the proposed solution may lead to someone closing it because of some problem with the proposed solution, losing sight of the problem, which remains unsolved.

If the problem seems minor in the absence of the elegant solution, then that's a case when I question whether an RFC issue is warranted.

( I have no doubt that the mass import from the other repo filed RFCS that are examples of this. But I aim only focusing on human activity on the RFC repo for now )

As @aturon said, we are still working out the kinks in the process. I just do not want to establish a precedent of "got a cute idea? Throw it on an issue the RFC repo."; that would risk us being overwhelmed, when we are trying to make the process (and thus this issue database) lightweight.

[pnkfelix]

@glaebhoerl I also do not understand your unwillingness to post to discuss. Surely those links would serve as permanent links to your idea, to point people at?

If everyone filed issues here for their sketch of a pet feature or ToDo item, then all those issues would be just as "lost" (by which I think you mean "not prominently displayed", but but I am not sure) as they are or become on the discuss forum, no?

[Centril]

It seems to me this issue has outlived it's usefulness as nothing has happened here for 4 years.