--On Saturday, September 22, 2012 09:33 -0500 Pete Resnick
<presnick(_at_)qualcomm(_dot_)com> wrote:
...
An I-D will only be removed from the Public I-D Archive with
consensus of the IESG. There are two situations when the
IESG will take this action. First, to comply with a duly
authorized court order. Second, to resolve some form of
abuse.
This second basis looks sufficiently broad and vague to
invite its own abuse and certainly inconsistent application.
Did IETF counsel express comfort with this language?
Counsel actually wanted us to broaden the language, thinking
"abuse" was too limiting.
Yeah. I think the new language is a considerable improvement
and I can live with it, but only because I think "abuse" can be
interpreted very broadly should someone be able to convince the
IESG there is a problem or unfairness associated with leaving
things up. My preference would be to go back to what I believe
was the spirit of the 2026 "...no formal status, and are subject
to change or removal at any time", leaving the IESG with
discretion to consider individual cases and decide what to do
without having to assign them to a category (whether "court
order" or "abuse").
Three observations in that context:
-- Whatever the language of BCP 78/79, or even the IPR
language of 2026, _permits_, doesn't imply that we
should do it, only that we (perhaps) can
-- I think the IESG should be more sympathetic to
removal requests for earlier documents that were posted
with the understanding that there would be no
IETF-maintained authoritative public archive than to
ones posted after the public archive became the
practice. That is not to suggest that all requests
about early document should be approved or that all
requests about later ones should be denied, but, as a
general guideline, I think it would be helpful.
-- I also believe that the IESG should be more reluctant
to remove a document whose successors are still under
active development (especially the immediately-previous
or otherwise very recent drafts for which diffs are
important) than documents that are only of historical
interest. If nothing else, it is easy to make a case
that having those recent documents with active
successors in the archive is serving the needs of the
IETF, while it is somewhat harder to defend keeping a
document that is of historical interest only on that
basis.
None of the above has to be written into a policy, but I believe
that the interests of the overall community and suitable respect
for the desires of authors, would be better served by such
guidelines than by, e.g., trying to figure out a specific and
yet sufficiently flexible definition of "abuse".
john