Sometimes I-Ds get early adoption because, for instance, a problem needs solved, and implementors don't or can't wait for IETF political processes to complete. Sometimes those I-Ds result in RFCs and sometimes in multiple RFCs and sometimes it is indeed an ugly mess. E.g., SIP DIVERSION and HISTORY-INFO.
Nevertheless, I value access to historical information over pedantic tidiness.
-----Original Message-----
From: John C Klensin <john-ietf@xxxxxxx>
Sent: Thursday, January 13, 2022 12:35 PM
To: Gorman, Pierce <Pierce.Gorman@xxxxxxxxxx>; Scott Bradner <sob@xxxxxxxxx>; IETF <ietf@xxxxxxxx>
Subject: RE: Backdoor standards?
[External]
--On Thursday, January 13, 2022 17:38 +0000 "Gorman, Pierce"
<Pierce.Gorman@xxxxxxxxxxxx> wrote:
> Also FWIW. Early adopters and independent adopters occasionally
> implement what is described in I-Ds and it can be helpful to have the
> I-Ds persist as a usable public reference (IMHO).
Yes. However...
(1) See my note to Ed.
(2) And sometimes, when they do that, the IETF decides to change important details. That results in a new I-D and ultimately an RFC for which implementations of an earlier version cannot
conform. Either the adopters accept that change, makes changes
to their implementations, and convince their users to accept those changes or they don't. In the first case, the early drafts become irrelevant. In the second, we end up with, politely, a mess. At their worst, such messes involve arguments that the IETF specs should reflect implementation realities even if the IETF has determined that those implementation cause serious risks to users or the Internet infrastructure (often invoking the "running code" principle) and/or that particular kinds of implementation are inherently more virtuous than others and that the standard should follow them.
There are some other issues as well, but the primary reason for all of the language about limited validity, references, citations as work in progress [1] is precisely to avoid those messes.
john
{1} See the "Status of This Memo" in any current Internet-Draft as well as Statements and documents cited earlier.
