In general, I would guess that there would be no problem, ethical, legal
or otherwise, if authors explicitly agreed, retroactively or by
boilerplate, to have the document archived. Thus, one could envision at
least three solutions:
(1) Modify I-D boilerplate to include (or not) a statement like
"The author consents to public archival beyond the expiration date."
or
"The author consents to restricted archival beyond the expiration
date."
(2) Ask for retroactive approval (via some web page check-off, since
this probably satisfies the relevant e-signature laws in the EU and the
US) for authors who consent to public archival. To keep this manageable,
it would probably be an all-or-nothing proposition.
(3) Allow I-D authors to restrict access to I-Ds. Depending on one's
interpretation of the strength of presumption of evanescence (which I
find a stretch - after all, many I-Ds were and are posted in their
entirety to mailing lists and included in other archival records that
are archived forever and publically searchable), this may or may not be
satisfactory.
My guess is that the vast majority of I-D authors don't care whether old
drafts are plastered over the electronic display on Times Square and
would be flattered if another author considers the idea relevant enough
to cite them in a publication. Having a well-known repository is a much
better solution than hoping that Google will ferret out a copy in a
forgotten dark corner of the Internet.
Beyond the IPR issue, articles and books don't have a choice about
including references to I-Ds if they are to be intellectually honest.
Indeed, not including an I-D could well be considered plagiarism or
stealing ideas without attribution. Just labeling it as the equivalent
of "private conversation" is not very helpful or honest, either. Authors
can't wait until I-Ds become RFCs to publish technical work.
--
Henning Schulzrinne http://www.cs.columbia.edu/~hgs