Individually-submitted proposed RFCs (was: Re: what is a prob lem)

[John C Klensin]

--On Tuesday, 26 November, 2002 14:37 +0100 "Wijnen, Bert
(Bert)" <bwijnen@lucent.com> wrote:

>> Being flexible with the timeout during IETF meetings, and the
>> immediately proceeding period, seems completely reasonable to
>> me.  But the RFC Editor believes, or believed a few months
>> ago, that IESG had instructed them to not submit documents,
>> nor even initiate an internal review on those documents,
>> during that period.  That is _not_, IMO, reasonable.
>> 
> Although in practice it boils down to the same thing, does it
> not.

Nope.  Assume things are running very smoothly, and a document
arrives shortly before IETF (say, within the first I-D cutoff
window).  Now, the RFC Editor can, in principle:

* Review that document quickly, and send it off to the IESG with
a four week timeout, assuming an automatic extension request.
This results in the IESG needing to respond during its first
teleconf after the IETF meeting.

* Review the document, but send it to the IESG for review
immediately after IETF, with the normal two-week timeout.  This
results in the IESG needing to respond in either its first or
second (or, with bad timing, perhaps third) teleconf after IETF,
depending on whether an extension is requested.

* Not start reviewing the document at all until after IETF,
which is what (at least a few months ago) the RFC Editor
believed the IESG had "requested".  If the review occurs
promptly at that time, the document would probably reach the
IESG during its "reading period" for its second teleconf (it
would be nearly impossible to make it into the first one),
which, with an extension, means action at the third
teleconference.

* Not start reviewing the document at all until after IETF, then
notice that IESG had generated several new documents for RFC
Editor processing immediately before and during IETF.  If they
then applied the other IESG "request" and processed those new
documents rathe than reviewing the queued individual submission,
the latter is likely to be delayed a month or two before even
being reviewed and only then go to the IESG.  That implies
processing at the fifth or sixth telechat, if that soon.  Worse,
if the RFC Editor finds issues that require discussion with the
author and potential submission of a new I-D, that process
doesn't start until a month or two after IETF (two or three
months after the request) and the document, apparently, then
reenters the queue from the beginning, behind new IESG-processed
documents that have appeared in the interim.  That scenario
starts to approximate "forever".

And it is the last of these options that we seem to be following
now.

regards,
     john


> RFC-Editor can:
> - ask for timeout of 4 weeks (what they do today) and wait 6-8
>   weeks around a IETF meeting
> - wait a few weeks before it actually asks for a 4-week-timeout
>   review.
> 
> As Patrik also explained... I think the timeout mechnism often
> does not work very well. I have seen quite a few docs that stay
> in "IESG-review" for more than 4 weeks.