Re: [mpowr] Mailing List Management

--On Saturday, December 20, 2003 21:45 -0800 Harald Tveit
Alvestrand <harald@alvestrand.no> wrote:

> in my opinion Alex Conta and Scott Bradner are wrong.
> 
> There are cases where, for the good of the working group, a
> person should have his or her posting privilleges revoked -
> FAST - as on the order of a day or three.
> 
> This is NOT a free speech issue - it has to do with the
> ability of the IETF to conduct its work in an open and orderly
> fashion - BOTH open AND orderly.
> 
> For these cases, a simple procedure with 100% clarity on who
> makes the decision is needed - not a "three strikes and then
> wait another month" procedure, as any procedure involving the
> whole IESG is likely to be.

Two additions to Harald's comments...

As people have probably deduced from other comments I've made,
I'm very concerned about the ability of the IETF to produce high
quality standards in, with apologies to Wirth, as much time as
that takes but no longer.   I don't think we are doing nearly as
well at that -- along either the quality or speed/efficiency
dimensions-- as we were several years ago.  There are certainly
some reasons for that which are unavoidable, e.g., our typical
standard today really does seem to be more complex than was
typical, say, a decade ago.  But some of it is that "we" have
just gotten tied up in a lot more procedures and rules and
requirements.    And, sometimes because of those procedures and
requirements and sometimes because there is just less sense of
collaboration and community, we seem much more subject to DoS
attacks... whether through volume or abuse and whether
intentionally or not.

Those attacks need be to stopped in a length of time that is
inverse to their severity.  If someone is merely making a few
off-topic postings too many, there should reasonably be warnings
and a serious attempt to be sure that the person understands
what is, and is not, off-topic before anyone starts
contemplating revoking privileges.  With more severe attacks, it
may be entirely appropriate, IMO, to pull someone off a list
instantly and negotiate (or appeal, or whatever) later.

Do we need firm rules or procedures to cover these cases?  I
certainly hope not: not only do I not want to see more
procedures, but because I am certain that they will not work,
especially against someone whose intent really is to disrupt or
prevent work getting done.   I believe we should look at every
potential new rule or procedure and ask ourselves the questions:

	(i) What is the problem this proposed rule is trying to
	solve?

	(ii) Will it actually solve that problem, or does it
	only address the easy cases that can be dealt with in
	other ways and/or some problem that occurred once or
	twice in the past but may not occur again?  Put
	differently, if some protocol lawyer and potential
	deliberate abuser comes along and reads the rule, will
	he or she be able to find a loophole through which to
	cause harm and then claim that it is permitted behavior
	since the rule doesn't say anything about it?    If the
	answer to that question is "yes", we are better off
	without highly-specific rules and procedures in almost
	all cases.

	(iii) Can we solve the problem equally well by clearly
	giving someone discretion and authority (ideally with
	general guidelines that represent clear community
	consensus) and then holding them responsible if they
	abuse one or the other or both?

If the answers to those questions don't clearly indicate that we
need a new rule we are, IMO, almost always better off without
it.  And I think the IESG ought to be going through the current
collection of procedural rules, finding the ones that wouldn't
be approved if the above criteria were applied, and getting rid
of them.

Now, going down this path probably implies that we will have
more appeals, and maybe even a few recalls, in our future than
we have had in the past.  Personally, I would prefer to see that
rather than having us spend large amounts of time getting rules
just right which then, because of the loophole problem or other
characteristics, end up getting appealed anyway.

If the Nomcom does not understand that ADs are going to need to
exercise a good deal of discretion so that it selects only
people for those positions who can be trusted to do so wisely,
then we better given them the message because nothing is going
to work otherwise, no matter how many rules we make.   If ADs
don't understand that WG Chairs intrinsically have a lot of
discretion and that they need to select only those people as WG
Chairs who can be trusted to exercise it wisely, then we need a
different set of ADs -- and, again, the Nomcom needs to
understand that the ability to make those selection decisions
wisely is an important requirement for ADs.

So, should a WG Chair have the authority to put someone off a
mailing list for at least a while?  Yes, I think so (and really
see no alternative that doesn't tie us in knots).  Should there
be some guidelines so that WG Chairs can guess what is
appropriate, yes, if we can get someone to do the work.   Should
someone ejected from a list have the right to get (via appeal or
otherwise), a quick review and response from the AD?  Yes, and I
hope that doesn't require a rule (see below).  Should the WG
Chair consult the relevant AD before pushing someone off a list?
Yes, that would be sensible and reasonable.  Should we _require_
that the WG Chair do that?  No, I don't think so, partially
because, if something really needs immediate action and the AD
isn't available, the action shouldn't need to wait... But a WG
Chair who pushes someone off a mailing list --or takes any of
several other actions I think a WG Chair should be able to take
on his or her own-- ought to expect to get fired if the AD later
concludes that the judgments involved were excessively bad.

So, the needed rules here are:

	(1) A WG Chair can eject someone from a meeting or
	mailing list or take any other reasonable action, if it
	is reasonably and sensibly required to prevent or
	eliminate disruptive activities so that the WG can get
	work done.  In exercising that authority, the WG Chair
	is expected to consult as necessary and reflect current
	community culture and judgment as to what actions are
	appropriate under what types of circumstances.

	(2) Someone who has been ejected from a mailing list or
	otherwise subjected to WG Chair discipline has the right
	to appeal and has the right to expect a timely review
	and response to the first stage appeal to the relevant
	AD.  ADs who do not respond to such appeals on a timely
	basis are a menace to the smooth and efficient workings
	of the IETF and should be swiftly recalled.

What is interesting about those rules, of course, is that they
don't require any changes to existing written formal rules and
procedures, only the application of common sense to those
rules... and changes to some of our habits and assumptions about
how we do business.

Two final coments, lest I be accused about writing too-short
notes :-(

(1) "Free speech" makes a nice chant.  There are any number of
places on the Internet in which it can be exercised.  But,
chants aside, the interpretation of "free speech" necessary to
have a society work is that one person's free speech not create
dangers for, or infringe on the rights of, others.   In the IETF
context, one should be "free" to "speak" on a given mailing
list, or in a WG meeting, if ones postings are reasonably
on-topic, constructive, polite, and civilized.  If not, we need
to remember that the costs of excessive hostile postings (for
any definition of "hostile") include driving people off lists
and participation in discussions whose input really would be
valuable.  We must not get to the point at which it becomes
acceptable to have the requirements for participating in an IETF
WG be not only expertise and a concern about the issues but also
the willingness to suffer high-volume abuse.  If we drive away
the people who do have the expertise and concern but won't take
the abuse, any claims we make to _community_ consensus around
the results of the relevant WGs work become highly questionable.

(2) This issue has nothing to do with debates about the
political theory of the rights/ authority of a group versus
those of it chair/ moderator/ leadership.  Down through the
centuries, people and organizations have always discovered that
open, everyone-talks-at-once, assemblies are a lousy way of
getting anything done and, as a result, figured out some
"leadership" selection mechanism.  And, regardless of how those
people are selected (or for how long, or by whom), they always
end up being give the authority to restrain or eject or isolate
someone who is being sufficiently disruptive... or the
organization doesn't survive and function.  So, while I think
ongoing discussions about how Chairs are selected, and to whom
they have the most responsibility, are worthwhile, I think we
need to accept that, in order to get work done, Chairs are going
to need some authority to steer or moderate discussions,
interpret and enforce agendas, and limit potential damage or
disruptions... or we are basically finished around here except
for the shouting (or whimpering).

      john

_______________________________________________
mpowr mailing list
mpowr@ietf.org
https://www1.ietf.org/mailman/listinfo/mpowr