Hi John,
Please see my responses inline below.
- Larry
On 5/31/15 8:51 AM, "John C Klensin" <john-ietf(_at_)jck(_dot_)com> wrote:
--On Friday, May 29, 2015 22:07 +0000 "Larry Kreeger (kreeger)"
<kreeger(_at_)cisco(_dot_)com> wrote:
Hi Brian and Benoit,
If authors are added purposefully to a draft without their
consent, it is indeed disturbing, but I hate to see IETF
cycles spent dealing with investigations if they can be
avoided in an automated manner.
Isn't one way to deal with this to build additional mechanisms
into the draft submittal confirmation procedure? I believe
that currently, only one of the author's needs to approve
submission/update. If this were changed to require all
authors to confirm the original submission, then it would save
people-hours investigating and dealing with this. When
submitting an updated draft, only a single author would need
to approve if the author list didn't change. Of course, newly
added authors would need to approve as well if the updated
draft included a new author.
Larry,
In principle this is very attractive (and has been suggested
before). In practice, it illustrates how much trouble we get
into when we start making very specific rules and then
incorporating them into even more specific tools. Consider, for
example, a draft that is actually a revision of an existing RFC,
changing a few details, perhaps adding a critical explanatory
paragraph, but leaving virtually all of the original text (and
perhaps all of an original protocol specification) intact. Our
existing guidelines, usual practices, and
draft-carpenter-whats-an-author-01 all suggest leaving the
original author as an author on the revised version. After all,
she is that one who, in the words of Brian's draft, made "a
substantial creative contribution to the document", indeed may
have written most of it. But suppose that author has dropped
out of the IETF and cannot be easily reached?
In case it wasn¹t clear, I was proposing that only original and
new authors would need to approve a submission. So, an update
(whether large or small) that doesn't add to the author list
doesn't require any additional approvals.
Suppose that
author, if reached, had a significant objection to the changes,
the idea of a revision [1], or the author of the new text?
If they really don't like it and which to withdraw themselves as
an author, I think that should be allowed with no additional
approvals except from the withdrawing author.
I have an opinion about how situations like that should be
handled, but it involves subjective elements, edge cases, etc.
It won't work well with a tool. Could your idea be made to work
with a "manual posting by secretariat" exception? Yes, but then
we would have be punishing the people who are behaving well with
delays at least and maybe creating the need for the Secretariat
or IESG to do an investigation.
Yes, delays for additional approvals to be made are inevitable. I
suppose the judgement call comes down to. 1) How often is this
happening? 2) How bad is it to deal with when it does happen?
If it very rare, and the consequences to dealing with it are not
burdensome, then I agree, don't burden the submission/update tool
with additional approvals. I have no idea how wide spread the
problem is.
Perhaps that would be the best opinion anyway (these cases are
not common although they may become more so as authors age), but
my point is that this sort of problem is more complicated than
one at which simplistic tools can be thrown without creating
other problems.
john
p.s. Brian, your draft really needs to address I-Ds that are
revisions of existing RFCs even if all you say is "more
complicated".
[1] The IPR rules do not permit the authors of an RFC in the
IETF Stream to block publication of an IETF Stream revision or
update of that document. The current discussion seems to
suggest that they should be able to insist on having their names
removed as authors, which is fine. The question of whether, if
they wrote most of the original (and retained) text, they can
also insist on not being acknowledged seems to me to be fairly
clear (and clearly "no") in the IPR rules although the issues
such a request raises are examples of why I object to wading
into the Acknowledgment swamp with a parenthetical note. But,
if their wishes cannot be reliably obtained... well, we better
not have a restrictive rule or tool that assumes that reliably
obtaining those wishes is easy and straightforward in all cases.