Re: FW: AD response to Site-Local Appeal

[Robert Elz <kre@munnari.OZ.AU>]

Dear IESG members.

When considering Tony Hain's appeal (which was in the message
referenced in the "References" header of this message), there
are a few more points that he didn't mention that you may wish
to take into account when determining whether or not there were
procedural irregularities.

First, RFC2418 (aka BCP 25) (WG procedures) requires...

   For coordinated, structured WG interactions, the Chair(s) MUST
   publish a draft agenda well in advance of the actual session.

The agenda of the March 2003 ipv6 WG meeting was sent to the mailing
list on March 14.   The WG meetings were on March 17 & 20.   "Well
in advance".   Hmmm...

Whether a copy of this was sent to agenda@ietf.org (as required) or not
I don't know, but it appears not, as nothing is present in the ietf/ipv6
directory (the last posted agenda or minutes there are from the November
2002 meeting).

Then

   The minutes shall
   be submitted in printable ASCII text for publication in the IETF
   Proceedings, and for posting in the IETF Directories and are to be
   sent to: minutes@ietf.org

As best I can tell, no minutes have been submitted for that meeting
(even though the next meeting has already been completed).   Draft
minutes were sent to the mailing list (March 28), but nothing further
seems to have been done.  They certainly don't appear in the normal
IETF places for WG meeting minutes (as for example, the November 2002
minutes do).

Given all of this, it would not be unreasonable to ask that everything
done at the SF meeting of the ipv6 WG be eliminated from the collective
record, just as if the meeting never happened.   That's probably the
correct course of action - but we probably don't have to take it quite
that far.


Then, again from rfc2418:

   In determining the balance, the WG must ensure that
   its process does not serve to exclude contribution by email-only
   participants.  Decisions reached during a face-to-face meeting about
   topics or issues which have not been discussed on the mailing list,
   or are significantly different from previously arrived mailing list
   consensus MUST be reviewed on the mailing list.

First from that note, "reviewed" on the mailing list, not "continued
there".

Before we get to that though, there's the issue of whether the "deprecate
site local" was a new issue, or one that had been dragging along forever.

Look at the agenda of the SF ipv6 meeting (that one that appeared all of
3 days before the meeting), and on this topic you'll see ...

 Site-Local Addressing
   Impact of site-local addressing -- Margaret Wasserman (20 min)
   http://www.ietf.org/internet-drafts/draft-wasserman-ipv6-sl-impact-02.txt

   "Limited Usage" Summary -- Margaret Wasserman (5 min)
   [See appendix of draft-wasserman-ipv6-sl-impact-02, above.]

   "Moderate Usage" Proposal -- Bob Hinden (15 min)
   http://www.ietf.org/internet-drafts/draft-hinden-ipv6-sl-moderate-01.txt

   Chairs recommendation and discussion -- Chairs  (30 min)

By all means go and read the two drafts mentioned, you will find nothing
about "deprecate site local addresses", that appears to have been a very
late invention.   That is, clearly this is something, which in the words of
rfc2418 was "topics or issues which have not been discussed on the mailing
list, or are significantly different from previously arrived mailing list
consensus" and so "MUST be reviewed on the mailing list."

What was actually done on the list wasn't to review the decision, but to
augment a vote taken during the WG meeting.

That is, rather than a review of the decision (which is to say, a new
decision process which determines whether the original decision should
stand), the WG meeting was just extended out into the mailing list.

That is not what 2418 requires.   Further, there is very good reason for
that - it is way too easy for people in a face to face meeting, who
haven't really thought through a new issue or proposal, to be swayed by
a persuasive sounding argument, and express agreement, only to end up
changing their minds once a period of reasoned thought has had a chance
to occur.  In many cases people who have no opinion express one, just
because lots of other people are doing the same thing, they have no reason
to disagree, so they go along with the crowd.

That's why we have mailing list review of decisions - on the list there's
no reason for those who have no opinion to say anything at all, and there
is plenty of time (or should be) for contemplation of all of the arguments
before making a decision.

In this case however, when the issue was brought to the mailing list
(message of April 01 - comments about fools omitted here) the
requirement was, quoting from that message:

	This email asks those that were NOT present at the Thursday IPv6
	meeting in SF to express their opinions on a question that was
	asked of the room.   If you expressed an opinion on this issue in
	SF you can skip this message; in any case you MUST NOT respond to
	this query.

That is, no-one was permitted to have had a change of mind.   One may
speculate as to why this "unusual" approach was taken, and perhaps
reach the conclusion that having achieved a large majority (reported in
the draft minutes as 102 to 20) in favour of the "deprecate site local"
decision, they didn't want to risk any of that majority being dropped
away, or any of the people who supported it, not really being on the
ipv6 list, and so not normally a part of the working group or its
decision making process (which isn't to say anyone should be excluded,
anyone can join the list, and if, knowing the issue was to be discussed
there, they had a strong opinion upon it, anyone could have joined the
list to express that opinion).

When the mailing list voting ended
	[aside: Tony is quite correct, a cursory review of the list
	 archives will find messages with subjects like "Consensus on what?"
	 where people were attempting to work out just what it was they were
	 being asked to decide - and note this is well after the WG meeting,
	 if people on the list in early April still had no idea what the
	 question really was, how could anyone who was actually in the
	 face to face meeting?]
for which apparently just 1 week was enough (as the decision was
announced in a message on April 9) despite this obviously being a
highly contentious issue, the result announced was

	The final count (including the room and the 
	mailing list) was:  155 YES, 56 NO (211 Total).

If one were to ignore the people in the room, what that means is that
on the list, the opinions were 53 to 36, which is about as close to a
split vote (clearly nothing like any kind of consensus) as is ever
going to be achieved.

Of course, we will never know now just what the count might have been
had the instruction not been given to people who were in the SF meeting
to say nothing on the issue.   Of course, there's also no list anywhere
public of who the 122 people who voted in SF actually were, I have no
idea if the chairs know, but I'd suspect not (just doing something which
is going to pretend to be an accurate count of 122 votes near the end of
a busy WG session is close to impossible to get right, making a list of
names is almost impossible).   That means there's no way to tell how many
of them voted again (on either side of the issue).

This whole process was clearly flawed.

Further, even that assumes that taking votes is the way the IETF
operates.   That would not often be regarded as a normal way to do
things.   Normally, WGs take "straw polls" to determine whether
just about everyone agrees with a point, or disagrees, or whether
there's a serious split.    Here, had this been regarded as a straw
poll, then even with 155 to 56 numbers, "serious split" and "no consensus"
would have been the only reasonable result.

The IESG must instruct the ipv6 WG chairs that they erred in determining
that there was consensus on this issue, that the procedures were flawed,
and that the IETF does not count votes to determine consensus, or rough
consensus.

This may not seem very important, as there is no apparent outcome from
this decision.  What is important though is that the WG not be railroaded
into agreeing to a later document (which is rumoured to be going to be
written one day) with the argument that the WG already reached consensus
that this be done, and that no re-opening of that issue will be permitted
(which is what I anticipate may happen).   That is, if this planned
document ever appears, the only discussion allowed may be whether or
not it says the right things to implement the supposed consensus, and not
whether or not the document itself is a good thing to publish.   That
cannot be allowed to happen, and certainly not based upon the flawed
processes in this particular instance.

Finally, while not strictly relevant to the current question, you may
also wish to know that on April 10 (the day after the decision was
announced), I sent a message to each of the WG chairs (and copied to the
WG mailing list - Message-id: <11997.1049967846@munnari.OZ.AU>) in which
I asked ...

	You can expect an appeal against the decision that there is rough
	consensus here.   What there is, is a debate.    You can treat this
	message as an official request to the WG chairs that they reconsider
	the stated position.   That's the first step ...

The WG chairs did not bother to reply to any of the issues raised in
that message.  Or at least, they have not so far.

kre