Marshall,
I don't disagree that having more than one person involved
in making an exception makes the exception-making process more
visible. It makes it harder as well.
Part of the complexity (and surprise) involved in allowing
exceptions is when the meeting attendees assume late submissions
will not be allowed and budget their reading time accordingly -
only to find out that they were wrong.
Never mind the case where a submission occurs really late
- what about the submission that occurs just in time for an IETF
meeting-goer to print it out (or E-Copy it to their local system)
for reading on the plane. If they (mistakenly) believe that the
draft will not be allowed because it is really late, they will
most likely print/copy and read other drafts.
In one recent case, a raft of late IDs were allowed, most
likely because the affected WG would otherwise have had little
reason to meet. Many, if not most, people had not read these
drafts because of - among other things - day-job pressures and
the need to read and determine comments/questions on other work.
In my opinion, it would have been better if those drafts
had not been allowed - especially since we have now established
a precendence that deadlines for IDs in that WG are irrelevant.
Attending future meetings of that WG is very likely to be a bit
too interesting.
IMO, more important than the long term visibility is the
short term commitment of the WG and - particularly - WG chair(s).
Does the fact that a slot is assigned for discussing an ID mean
we must assume that it will be discussed, even if there's nothing
to discuss, or is a more explicit statement of intent required?
In either case, we don't want the requirements of a decision to
be made so difficult that there is a significant delay in getting
a decision. If there is a noticeable delay, we do not find out
what the decision is until even later than we become aware that a
decision may be in the works. If this happens a lot, then there
is no way to plan our reading in advance of a meeting.
--
Eric
--> -----Original Message-----
--> From: Marshall Eubanks [mailto:tme(_at_)multicasttech(_dot_)com]
--> Sent: Tuesday, November 29, 2005 1:56 PM
--> To: Gray, Eric
--> Cc: dcrocker(_at_)bbiw(_dot_)net; ietf(_at_)ietf(_dot_)org
--> Subject: Re: EARLY submission deadline - Fact or Fiction?
-->
--> Hello Eric;
-->
--> Note that this is not about approving documents, but about
--> approving
--> the
--> discussion of documents. The threat model is weaker.
-->
--> Suppose this was abused. Suppose, as is the case recently in our
--> House of Representatives,
--> that new 400 page documents came up for discussion after
--> having been
--> posted a few hours before.
--> Suppose that the WG Chair was intent on pushing these
--> through. If I
--> was active in the WG, I
--> might complain. If there was reason to discuss this at a higher
--> level, say on this list, then
--> my complaint would include the paper trail ("at IETF-99 we had to
--> discuss the 400 page HTTP over BGP
--> draft only 1 hour after it was posted"). I would like having this
--> paper trail. It's a lot easier to show
--> this abuse than proving, say, that the Chair claimed consensus for
--> when there was actually confusion and doubt.
-->
--> So I think that the threat model is weak, and that in some sense
--> abuse will be self correcting. But I do think that there should be
--> two people concurring, thus I think that the AD's should have to
--> agree too.
-->
--> Regards
--> Marshall
-->
--> On Nov 29, 2005, at 1:26 PM, Gray, Eric wrote:
-->
--> > Marshall,
--> >
--> > This would work reasonably well, if we could be sure
--> > that the "planetary alignment" case was unlikely. Unhappily,
--> > it is not. What happens to the credibility of this approach
--> > when the ID submitter, WG chair and AD are all from the same
--> > company?
--> >
--> > --
--> > Eric
--> >
--> > --> -----Original Message-----
--> > --> From: Marshall Eubanks [mailto:tme(_at_)multicasttech(_dot_)com]
--> > --> Sent: Tuesday, November 29, 2005 1:23 PM
--> > --> To: dcrocker(_at_)bbiw(_dot_)net
--> > --> Cc: Gray, Eric; ietf(_at_)ietf(_dot_)org
--> > --> Subject: Re: EARLY submission deadline - Fact or Fiction?
--> > -->
--> > --> It would seem to me that this could be pushed to a
--> degree onto the
--> > --> AD's -
--> > -->
--> > --> late submissions (up to the point allowed by the machinery)
--> > --> would be
--> > --> accepted on the request
--> > --> of the WG Chair and the concurrence of one of the WG AD's.
--> > -->
--> > --> I have certainly seen cases where such flexibility
--> would have been
--> > --> useful. Getting two people to concur in the
--> > --> request will make it a lot harder to abuse the process.
--> > -->
--> > --> Regards
--> > --> Marshall Eubanks
--> > -->
--> > --> On Nov 29, 2005, at 1:10 PM, Dave Crocker wrote:
--> > -->
--> > --> > Eric,
--> > --> >
--> > --> >> One of your comments seems to apply to the effectiveness
--> > --> >> of having an early submission deadline. What is
--> the point of
--> > --> >> monkeying around with early submission deadlines
--> when they are
--> > --> >> not very effective anyway?
--> > --> >
--> > --> > 1. They add administrative hassle to working groups.
--> > --> >
--> > --> > 2. They lead to having the authoritative version of
--> documents be
--> > --> > outside the Internet-Drafts mechanism, at least for awhile.
--> > --> >
--> > --> >
--> > --> >> Sometimes a WG
--> > --> >> chair schedules time to talk about an ID that
--> doesn't exist at
--> > --> >> the time of the schedule announcement and - sometimes -
--> > --> still has
--> > --> >> not been submitted by the time of the meeting. ...
--> > --> >> So, one question is whether or not it is appropriate to
--> > --> >> allow this practice to continue.
--> > --> >
--> > --> > The broad question is how much freedom a working group
--> > --> should have
--> > --> > to formulate its own procedures. In the not-so-distant
--> > --> past, the
--> > --> > IETF was quite friendly to wildly different working group
--> > --> choices.
--> > --> > More recently our respond to cases (or, yes, patterns) of
--> > --> > misbehavior has been to make a rule that restricts
--> everyone with
--> > --> > respect to procedural details.
--> > --> >
--> > --> > My own view is that we need to facilitate working
--> group progress
--> > --> > while ensuring working group legitimacy (fairness,
--> > --> timeliness and
--> > --> > relevance). I believe that progress is facilitated
--> by letting a
--> > --> > working group do as much self-organizing as it can, while
--> > --> keeping
--> > --> > the working under pressure to be productive. We need to
--> > --> do this by
--> > --> > watching for a working group going astray, rather than by
--> > --> imposing
--> > --> > micro-managing, rigid rules.
--> > --> >
--> > --> > I believe this view is compatible with the core of yours:
--> > --> >
--> > --> >> IMHO,
--> > --> >> I think the common - usage-based - definition is
--> that WG chairs
--> > --> >> get to arbitrate the meaning of these terms as it
--> applies to ID
--> > --> >> submissions for their own WG, especially when
--> submissions are
--> > --> >> late. ...
--> > --> >
--> > --> >
--> > --> > I suspect we differ on:
--> > --> >
--> > --> >> On the other hand, if someone wants to reduce the WG
--> > --> >> chair's role in arbitrating the legitimacy of an
--> ID submission,
--> > --> >> then it is up to them to make sure that their
--> submission is in
--> > --> >> compliance with formal submission deadlines,
--> format, etc. - so
--> > --> >> that no "exception" is required.
--> > --> >
--> > --> > If I understand correctly, you want to retain a deadline,
--> > --> but give
--> > --> > the wg chair authority to override it. This certainly is
--> > --> > reasonable, but I think it is not practical because it adds
--> > --> > administrative overhead (and probably delay) in the
--> > --> Internet-Drafts
--> > --> > processing mechanism.
--> > --> >
--> > --> > A simpler rule is that the working group gets to decide its
--> > --> > deadlines and what will be discussed at the meeting.
--> > --> (All of this
--> > --> > is predicated on moving towards fully automated I-D
--> issuance.)
--> > --> >
--> > --> >
--> > --> > d/
--> > --> > --
--> > --> >
--> > --> > Dave Crocker
--> > --> > Brandenburg InternetWorking
--> > --> > <http://bbiw.net>
--> > --> >
--> > --> > _______________________________________________
--> > --> > Ietf mailing list
--> > --> > Ietf(_at_)ietf(_dot_)org
--> > --> > https://www1.ietf.org/mailman/listinfo/ietf
--> > -->
-->
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf