So let's consciously endeavor to ensure that sigificant non-standards documents -- responsible position papers, white papers, new ideas, etc. -- become RFCs.
Sigh. Even more RFCs. Pretty soon we're going to need a 32-bit RFC number space.
(Making Internet Drafts into an archival series seems like a terrible idea to me, but that is a different topic.)
Actually it's pretty much the same topic, as there needs to be a way to preserve drafts that are important in some way or another. One way to do this would be to make all drafts that are worth preserving an RFC. This means drastically lowering the standards for what can be published as an RFC. (Note that this brings us closer to what RFCs used to be 15 years or so ago.) Another way to do it would be to simply archive all drafts. I agree this has the unpleasant side effect that all those drafts that become obsolete (or are so from inception) stay around forever. But it's still better than the current situation, where making something an RFC means an incredible amount of work for many people, but not doing it means that ideas that may have taken days or weeks to write down are pretty much lost forever. Even the emails announcing the availability of new drafts are archived longer.
I think there is some middle ground where expired drafts that are referred to in non-expired drafts are kept, as well as any drafts that are implemented and possibly drafts that a certain n number of people feel are interesting in some way.