In my view, this draft buries the lede, by focusing on “expiry”, which has led much of the discussion down the path of talking about the mechanics of document expiry, etc. While it’s necessary to get these things right (or at least, not wrong), they aren’t the high-order bit IMO. The meat is in Section 2.2, which continues the misdirection by being titled "Removing the Expires field from Internet-Drafts”. The real key is the deletion of "It is inappropriate to use Internet-Drafts as reference material”. It’s impossible to ignore that for quite a long time, in practice we (or some large subset of “we”) haven’t genuinely believed those words even though we’ve continued to mouth them. The examples provided in Section 1 are convincing evidence. It’s problematic that our process documents say one thing, and our process-in-practice does something contrary. The options seem to be: - Continue to pretend there is no problem. I don't love this option. It’s one thing to pretend the king has clothes on, it’s quite another to continue to do so after it’s pointed out that no, he doesn’t. - Force reality to match BCP 9 by strictly enforcing the “not to be used as reference material” position. I think King Canute’s example is instructive as to how successful this is likely to be. - Update BCP 9 to match reality. Since the first and second options aren’t very appealing, this is where I find myself. As far as I’m concerned this could probably be handled by the single deletion I’ve quoted above. That approach would, admittedly, lead to questions like “if a draft is ‘expired’ what exactly does ‘expired' mean, then?” TBH, I’d be OK with the ambiguity, though, and it sure would simplify the document. (For that matter, I can see a related argument for eliminating "inappropriate to cite them other than as 'work in progress'”. Someone already mentioned that if a draft *isn’t* making progress, what does it even *mean* to call it a “work in progress”? It’s a work that’s failed to progress, rather. But I guess it’s close enough, and it does seem important to give some kind of clear signal to people outside our community that caveat lector, an Internet Draft is not any kind of standard.) If the authors/community want to fix the details about the mechanics of draft expiry, just exactly how long “long” should be, etc, that’s OK with me. If the authors/community want to cut it down to (what I claim are) the bare bones, that would be fine with me too. However we do it, for the reasons above I agree with the basic premise, that we should face up to the fact that like it or not, Internet Drafts *are* used as reference material, including within our own process, and update BCP 9 accordingly. Thanks, —John
