ietf
[Top] [All Lists]

Re: Does the IESG have the authority to do less than 3683?

2006-01-22 08:26:39
--On Sunday, 22 January, 2006 11:30 +0100 Brian E Carpenter
<brc(_at_)zurich(_dot_)ibm(_dot_)com> wrote:

fwiw, my feeling is that if we did bend the rules that way,
we'd be at strong risk of an appeal. I think the rules are
in a bit of a mess.

Brian,

I'm disturbed by several aspects of this, most of which have
little or nothing to do with Jefsey, his behavior, and what
should be done about it...

() While it is not an excuse or precedent for this case, the
IESG has a long history, some of it fairly recent, for far more
creative interpretations of assorted rules and statements in
written documents than this would call for.  I assume it is not
what you meant by the above, but, if our real criteria for what
the IESG (or IAOC, or IAB, or particular individuals) can or
cannot do are levels of "risk of appeal" or "risk of recall",
then we are in a much bigger mess than our rules.

(2) We seem to be having a pattern of problems with all of our
procedures.   It appears to me that we are patching things
without regard to the systems in which those things are embedded
or the side-effects of the patches.  We have seen it, IMO, with
our apparent inability to get IPR policies right without a new
iteration starting as soon as the latest version is published,
with  the exclusion of IAB and IESG members from those who can
initiate recalls, and with, apparently, replacing a series of
graduated moves with the huge gap between 30 day suspensions and
the rather sweeping and nominally permanent 3683 action.  

In at least the latter two cases, I believe the outcome is
accidental.  Certainly there was no evidence during Last Call on
the relevant documents that the community was aware enough of
the consequences and tradeoffs involved to discuss them.
Speaking personally, I know I read the drafts that became 3683
on the assumption that it added one more tool to the collection
(and hence was harmless at worst) rather than that it excluded
any pre-existing options.  I assume others did too.

Perhaps this demonstrates where, as a community, we are skilled
and where we are not.  Perhaps we need to do as good a job of
looking at potential consequences and side effects of procedural
changes as 
we try to do for protocols.  Whatever is needed, it is becoming
fairly clear that the focus on the next narrowly-defined problem
or the next patch in {ipr, newtrk, pesci, techspec} is a good
recipe for making more mistakes of this type.

(3) One of the consequences of several of the situations in
which we have found ourselves is that we seem to be replacing
assumptions of good sense and the exercise of discretion with an
increasing array of rules.  I consider that undesirable in its
own right.  Worse, the rule-making process seems to always have
us fighting the previous war.  However, operating on the basis
of giving flexibility and discretionary authority to, e.g., the
IESG, IAOC, or other "leadership", requires that those people be
exceptionally sensitive to the views and temper of the
community.  If that sensitivity does not appear to be present;
if, instead, the community sees signs of "we can ignore some
rules but interpret others narrowly as it suits our convenience"
or the arrogance of decision-making based on "if we don't like
that idea or find it threatening, it doesn't have a chance" or
"if they don't like it, they can recall us",  or "let's divide
this up into enough separate discussions to prevent anyone from
getting a complete picture so we can claim support for whatever
we feel like doing", then more rules, many of them probably
proposed or applied as hasty patches, are almost inevitable.
And thus we spiral downhill.   

(Disclaimer: I do not have any particular person's behavior or
any particular event in mind in the paragraph above, only what
appear to me to be patterns over the last several years.)

(4) As for the present issue, I observe, as have others, that we
have now spent more energy, and created more disruption,
debating the appropriateness and severity of Jefsey's behavior,
and whether a particular response is appropriate, than the
behavior itself has spent or caused.  In a way, the comments and
postings have been helpful: we have had much of the conversation
about the appropriate application of 3683, and the lack of
in-between measures, that, in retrospect, should have occurred
when we started to make changes.

So, let me make a few suggestions for getting us unstuck and
back to useful work.

(i) If Jefsey's recent behavior justifies it, generate more 30
day suspensions.

(ii) Let's establish a convention (not a rule -- we would just
screw it up or get tangled in it) that, if a suspension action
is taken against someone whose native language is not English,
we attempt to deliver the suspension notice/ complaint in that
person's language as well as English if reasonably feasible.  We
should not, if possible, have people suspended for reasons they
don't understand or can later claim they don't understand.  On
the other hand, we have accepted (even if Jefsey has not) that
the ability to communicate in English is a practical necessity
for participation in IETF and should not paralyze ourselves by
trying to apply this particular courtesy.

(iii) Again, as a convention, let's agree that, if someone is
suspended more than once for the same basic pattern of behavior,
that the suspensions should explicitly note the history of one
or more prior ones and warn that the community will not tolerate
behavior that results in serial suspensions indefinitely.
Multiple suspensions and such a warning should not become
requirements for taking a 3683 action (or any intermediate
action), but represent a reasonable courtesy and attempt to get
the behavior adjusted, which we should certainly prefer to
ever-more-drastic sanctions.  We should not get ourselves into
the position of spelling out what penalties will be imposed of
the behavior recurs: as others have pointed out, the purpose of
all of these procedures is not to punish offenders, or to permit
offenders to measure the advantages of the behavior against the
potential punishment, but to permit and facilitate the
community's getting work done.

While I am sympathetic to Bert's position that more types of
sanctions would actually hurt us, I'd prefer to have some
intermediate levels of options for another reason.  3683 is not,
as I read it, intended to become a legal proceeding, with the
entire IETF acting as jury and discussing, not only the action
but the character of the presumed offender.  My personal
perspective is that I'd like to see 3683 actions under only two
circumstances.  (i) some sequence of actions have been taken
which are so egregious and so obviously motivated by malice
toward individuals or the community that the community recoils
in collective outrage (or should do so) or (ii) there has been a
repeated pattern of suspensions for similar behavior from
different mailing lists and by different people, that the claim
can be made and supportive that the individual involved is
consistently disruptive and an IETF-wide response is in order.
Frankly, in the latter case, I'm not convinced that the
individual involved has any "right" to disrupt the IETF list
with his or her defense.  This business of letting someone who
has been disruptive try to turn the issue into a personality
matter started by the person who finally develops the energy to
come forward and request a strong action is not healthy for any
of us.

(iv) Let us, quickly, follow up on some of the suggestions that
have been made recently to create intermediate steps by
generating a draft, under the RFC 3933 model, to permit some
intermediate measures.  I'm particularly fond of reinstating the
authority of a WG Chair, with AD approval and subject to appeals
that do not suspend the action while they are being
investigated, to apply exponential back-off, if only because 
   suspend for 30 days
   have new incidents on days 31 and 32.
   spend two weeks going through the warn-review-and-suspend
proces again
   suspend for 30 days
   loop
model is intolerably inefficient if work is really being
disrupted. 

If we care, we should be able to have such procedures in place
under the 3933 model within 30 days.  If people are worried that
we also need a mechanism for pushing back on repetitive appeals,
someone should start working on an experimental draft there too.

Note that, if we get such a procedure in place 25 days from now,
and then find a need to apply them 32 days from now based on
prior suspensions, nothing is being done retroactively.

And then, sometime, we need to figure out why we keep doing
these things to ourselves procedurally and try to cure it.

     john


_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf

<Prev in Thread] Current Thread [Next in Thread>