I think we should be willing to admit that an Internet-Draft
often serves two distinct roles:
(1) As a formal part of the IETF document process.
(2) As an informal record of intellectual content.
Role 1 is obviously the primary purpose of an I-D; they are
created in the service of the IETF document process, not
just to document ideas. But the unavoidable fact is that
they *do* document ideas, as well as the trajectory by
which those ideas are accepted (or rejected).
It seems foolish to me to actively destroy a record of the
intellectual history of the IETF and its contributors.
Even the bad ideas, once they are written down, ought to be
archived, at least so that we can say "we tried that 3
years ago, and it didn't work" (or, with apologies to
Santayana, "Those who cannot remember the past are
condemned to re-invent it"). And even those documents that
are ultimately rejected as "bad protocols" might contain
"good ideas."
Others have already mentioned the utility of an I-D archive
for establishing intellectual priority, both for patent
disputes and for scientific publication. If I'm reviewing
a paper by X who claims to have invented Y, and I know that
Y was published by Z in an I-D long ago, I'm certainly
going to insist that X give due credit to Z.
For role 1, we require
(a) forward progress - i.e., I-Ds expire after 6 months
(b) a public distinction between I-Ds and finished IETF work
(c) no normative references to I-Ds from finished IETF work
(d) no normative references to I-Ds by other authorities
Role 2 depends on a permanent archive (and, with all due respect
to JNC, without a needlessly punitive mechanism to make it
inconvenient to reference archived I-Ds).
There is no inherent conflict between that archive and the
need to preserve the distinction between archived I-Ds and
active ones. We don't want people making normative
references to any kind of I-D, active or expired; creating
a convenient, stable archive really doesn't affect that.
And it's disingenuous to pretend that if the IETF doesn't
archive I-Ds, nobody else will. In other words, the IETF
cannot successfully police (b) and (d), and it shouldn't try.
All we can do is educate people.
The IETF is perfectly capable of policing (a) and (c) above,
through its standard procedures.
Perhaps the IETF should consider adding an explicit warning
to each I-D when it enters the archive:
STATUS OF THIS MEMO:
THIS DOCUMENT IS AN EXPIRED INTERNET-DRAFT. USE OF
THIS MEMO FOR ANY PURPOSE EXCEPT AS A HISTORICAL
DOCUMENT IS STUPID, WRONG, AND EVIL, AND WILL LEAD
TO PUBLIC DERISION.
-Jeff