Note: this is a part of my formal response to IESG's Last Call
that I have chosen to make public. I have also, privately, made
a more detailed response available to IESG, which I reserve the
right to make available to others. Everyone who has received or
will receive a copy of the private response has been/will be
requested, and expected, to keep it in confidence.
What I'm saying in these messages is entirely my own opinion. I'm
not representing any one else, nor acting on anyone else's
request.
-Keith
I strongly disapprove of the proposed PR ("posting rights")
action with the proposed penalty. I believe it is
improper from a procedural viewpoint, unnecessarily vindictive,
without obvious justification, and an over-reaction at best,
proposing sanctions that are disproportionate to the actual
offense(s), if any. I believe it will do unjustified personal
harm to Dan Harkins, perhaps limit his income-making potential,
career, or work opportunity, and damage his reputation without
sufficient cause or factual support. I also believe that this PR
action, whether or not approved, has the effect of increasing
toxicity in the IETF community, and harming its credibility and
ability to function as a fair, consensus-building standards-making
organization. I respectfully disagree with those who assert that
this effort will make IETF less toxic or more fair.
It's possible that some lesser sanction than has been proposed in
the Last Call message, say more limited in scope and/or duration,
would be more appropriate. At the same time it may not be
possible to undo the harm that this proposal has already done.
I reread all of the messages cited in the Last Call and disagree
that those messages, individually or collectively, warrant a PR
action. That's not to say that I view every word in every
message as entirely appropriate or harmless. But I don't think
IESG has established enough of a pattern of abuse or disruptive
behavior on Dan's part to warrant such an action.
For many of the cited messages, I had to read the message multiple
times to even guess what IESG might have been concerned about,
since they did not make their concerns explicit. I'm not at all
sure that my guesses matched their concerns.
For the purpose of this Last Call response, I have not attempted
to review messages other than those cited in the Last Call, except
to establish context for the messages that were cited. Based on
private emails that I have received, I believe that some others
have based their responses, at least in part, on messages other
than those cited in the Last Call, and/or their personal knowledge
of Dan. While I do not wish to constrain anyone else's responses,
I personally believe that it's inappropriate for IESG to evaluate
such accusations outside the specific offenses which they, as
accusers, claim that Dan has committed. Accused persons cannot
effectively defend themselves against every possible
interpretation of every message they've sent to any IETF
discussion, for the indefinite past.
I. Problems with the accusations
1. The accusations do not cite any specific text that Dan
has posted, other than entire messages. Perhaps more importantly,
the accusations fail to explain why any of these messages violates
any specific IETF Consensus rule. A PR action should
not be justified by vague assertions, even if the action is
supported by other people commenting on the Last Call. Without
specific accusations, it's really impossible for the accused, or
anyone else in the community, to effectively call those
accusations into question. Those who wish to argue against the
proposed action have to second-guess what IESG was unwilling to
say (or perhaps, could not agree on).
2. More broadly, it seems like there's been a general refusal from
IESG and/or the Chair in all discussions about the use of language
in IETF, to be specific about what behavior(s) they believe are
worthy of sanction, or why. Instead there has, in my estimation,
been a continued insistence on ambiguous restrictions like
"unprofessional" or "uncivil" or "disruptive" behavior that are
subject to broad and arbitrary interpretation.
I do not believe that ambiguous rules are inherently wrong. I do
believe that imposing significant penalties on an individual based
on perceived violation of ambiguous rules is wrong. It's
difficult for those so accused to defend themselves, precisely
because the rules are so vague and arbitrary. Leadership and
Last Call responders alike can also effectively interpret those
rules however they wish, including varying the interpretation
according to how much they dislike the accused.
3. While this Last Call might not have been intended in that way,
and BCP 83 can be read to actually require it, there's something
sordid in inviting people to "pile on" to support such unsavory
ideas as censorship. It reminds me of a witch trial in which a
mob demands severe punishment for a vaguely defined crime like
being a witch, which cannot effectively be refuted, and for which
any accusation from a member of the mob is accepted without
question.
II. Evaluation of the cited messages relative to the
accusations
4. It seems like the bulk of the accusations against Dan are that
he has expressed "racism". I do not find such _expression_ in
the cited messages.
If the Last Call had cited messages in which Dan had promoted
abuse of, harm to, derision of, or prejudice against, people of
color or any ethnic background or culture, I'd emphatically
support the action. If he had called into question that racial
prejudice or inequities exist, or that they do substantial harm,
I'd also support the action. The existence of race-based
prejudice is well-established independent of anyone's claimed
experience, and denying that such prejudice exists is profoundly
insulting to people who routinely experience such prejudice .
But I haven't seen that he's done any of those things or even
hinted at them. Rather, I think he's questioned some people's
theories about the nature of racial prejudice, and the insistence
that those theories be treated as fact, or that they influence
IETF policy, without being subject to question. I do not
believe that any of the cited messages were racially prejudiced or
intended to promote racial prejudice.
I do not personally believe that disagreeing with someone else's
theory of racial inequity, or denying, belittling, or probably
even ridiculing such a theory, is an _expression_ of racism [*].
In the absence of unambiguous evidence supporting that theory and
disproving other theories, resulting from robust scientific
research, it is reasonable to call such theories into question.
Reasonable people should be able to agree or disagree with a
theory, and it should be possible to discuss, investigate, and/or
refine such theories until they're substantiated or found to be
lacking merit. Of course it's not merely the theory that was
being called into question, but the insistence that everyone must
adhere to that theory, attacks against people who do not support
that theory, and the insistence that the theory influence IETF
policy without question.
[*] Admittedly there's a somewhat a fine line here. I believe
that everyone has the right to try to interpret reality, and form
theories about reality, based on their own experience of it.
(Though they should be aware that nobody's experience is
representative of the whole of reality.) Indeed, I think
forming such theories is a fundamentally necessary and human
activity. Our brains are wired to build models of what we observe
and what is important to us. People should not be compelled to
believe or support others' theories. Nor should people disrespect
a theory that's widely held by some group, any more than one
should disrespect a religion or adherents of that religion.
There should be room to disagree, when necessary, without
expressing disrespect.
Similarly, it's not wrong to point out that the word "racism" is
sometimes used in a way that itself is problematic, or at least
ambiguous. While I understand that "racism" is sometimes used to
describe specifically structurally-based inequities, I also
observe that the word is not universally used in that way. And
in practice it's hard to tell when the term "racism" is intended
to be interpreted with such precision. It's absolutely
reasonable and useful to discuss structurally-reinforced
inequities in the treatment of certain groups, differently from
other forms of prejudice. But the varying usage of these terms
makes it difficult to distinguish one kind of "racism" from
another and respond appropriately, outside of certain circles in
which those terms are effectively disambiguated. IETF is not one
of those circles.
5. Dan has not always been infallibly polite. But neither have
his accusers, or really, have most IETF participants that I've
known over the years. IESG has not shown that Dan's messages in
their aggregate, are significantly less polite or more disruptive
than those of many other IETF participants who are tolerated or
even elevated to positions of authority.
Also, it's inappropriate to try to evaluate Dan's messages without
evaluating the context(s) in which they were written. While I
don't see "they started it!" wars as productive at all, we all
know that it's difficult to respond to personal attacks
respectfully, calmly, and rationally, every time. I believe that
all IETF participants who find themselves subject to such
attack have a duty to try to avoid adding heat to the discussion
(such as by adding insult to a reply), and that the community
leaders (IESG, WG chairs, etc.) have an even greater duty
than ordinary participants to avoid doing so.
7. I believe that this proposal is itself harmful to the consensus-driven IETF process, because it will likely have a chilling effect on all IETF discussion, especially because this proposal lacks specific justification for the proposed action.
I believe that this action is likely intended to serve as a warning to others, but it sets a dangerously ambiguous warning. If, for example, a PR action were proposed because one IETF participant accused another of being a "________" <insert insult>, approval or disapproval of such a PR action would at least establish a clear line for others to follow: don't call someone a "________" (or anything similarly insulting). By contrast, approval of this PR action would not establish a clear line, but instead send a more vague message that (for example) it's not okay to disagree with IETF leadership, or persons elevated by IESG leadership. While that kind of dysfunction is unfortunately common in corporate environments, it is not one of the virtues of "professionalism", and not consistent with IETF's mission. IETF participants need to be able to freely disagree with the views of IESG and other leaders, and freely discuss IETF processes, within reasonable bounds of decorum that apply equally to everyone.
If IESG believes that there are specific subjects which are unacceptable to discuss in IETF, IESG has the power to exclude those subjects in WG or mailing list charters, subject to the usual processes. Similarly working group chairs have the power to set aside discussions of topics that are not appropriate for the working group's charter. If individuals disagree, they have recourse through the usual public charter discussions or, when necessary, established appeal processes. I see no reason to sanction individuals for mentioning subjects that haven't been explicitly excluded, and are within the defined scope of their discussions.
8. The Last Call provided no information as to how the cited messages were selected. One responder suggested that the IESG had "tried alot with the individual but no change for two years", which is approximately the period of time that the cited messages represent. I wonder whether the IESG was collecting messages for two years to support a potential PR action. Surely the primary purpose of a PR action is to address abusive or disruptive behavior in a timely fashion. Was it really appropriate to wait two years? Or was there instead a desire to wait long enough to collect sufficient ammunition?
The communications between IESG or their proxies with the accused, prior to the initiation of the PR-action, were not disclosed to us. While the lack of disclosure may serve a valid purpose (why air dirty laundry if the parties can quietly agree on a resolution?) the lack of transparency becomes problematic once either party initiates action via some public process, and claims that they've followed the required process.
I know from my own experience as an AD that any communication from an AD can be taken as a threat, because ADs effectively have power to render many years' worth of your work useless, or delegate change control of that work to others, and/or harm your company's business interests. It's therefore difficult for an AD's unsolicited input to be seen as helpful, even if the AD is genuinely trying to help participants express themselves clearly. It's even harder if the AD accuses or suspects the participant of malice, rudeness, disruption or acting in bad faith.
In general, even though IESG claims that:
it's not possible to evaluate the veracity of such claims while those communications remain private. In particular, it's not possible to evaluate whether either party made a good faith effort to come to a resolution, or even to understand the other party's position. (This may be more of a problem with BCP 83 than with either IESG or Dan.)Multiple attempts have been made to enter into a private discussion with this individual, both by IESG and community members, to communicate disquiet with his conduct on the lists. These attempts to restore respectful and courteous conduct on the lists have been rebuffed with communication that can be considered both antagonistic and hostile, and the pattern of behavior observed has continued.
It has often seemed to me, that if someone wants to harbor or nurture or encourage some prejudice against an individual or group, they'll find plenty of "evidence" to support that prejudice, and will tend to disregard any evidence to the contrary. One reason is ordinary confirmation bias, to which pretty much everyone is susceptible. Another reason for that is that the prejudiced individual generally makes their prejudice evident to the individual or members of that group, and that influences how that individual or group members react to the prejudiced individual. (This can, of course, work both ways.)
III. Summary of my personal evaluation of Dan's messages
8. Most of the messages cited in the Last Call were on topic,
and within reasonable scope of, the discussions in which they
appeared.
9. Of the fifteen messages cited in the Last Call, I found:
- one message that I judged to be entirely inappropriate
(though not a violation of any of the Guidelines for Conduct or
any other IETF Consensus rule I'm aware of);
- two messages that contained pointless but relatively minor personal "digs";
- two messages that were parts of threads of insults traded between Dan and others;
- two messages that are in my opinion so old that they really shouldn't be evaluated to support this PR action (and one of those I did not find objectionable);
- one message that was almost entirely devoid of substance
relative to the volume of pedantic irrelevant content;
- two messages that I thought would have been inappropriate if
they had been directed at a specific person or people, but I
couldn't tell that they were; and
- approximately five messages in which Dan was responding to
insulting statements, personal attacks, or inaccurate
characterizations of his speech, made by others
Given that I find only a few of the cited messages even
borderline inappropriate, and that they were spread out over a
period of more than two years, I do not consider these messages,
in the aggregate, sufficiently offensive, insulting, or disruptive
to support the accusations made, or justify the proposed remedy.
If any of these messages actually were disruptive, IESG has not
supplied evidence to support that claim.
10. Dan has shown a willingness to discuss "hot button" issues,
or to take up less-popular positions on such issues. I don't
fault Dan for that in the least, as these are important issues
that need to be examined from all sides if they are to affect IETF
policies. However, extra care to minimize the potential
for flame wars is often needed when discussing such issues. I do
not find that due care was always employed in constructing those
messages.
Lack of due care is not a violation of our Guidelines for Conduct.
However, RFC 7154 does state:
English is the de facto language of the IETF. However, it is not
the native language of many IETF participants. All participants,
particularly those with English as a first language, attempt to
accommodate the needs of other participants by communicating
clearly, including speaking slowly and limiting the use of slang.
One could argue that Dan has not always made sufficient effort to
communicate clearly, for example by relying on references unlikely
to be familiar to works of English language literature which are not
known throughout the world, and expecting the meaning of such
references to be clear to such readers. 11. From the selection of messages cited, it is
difficult to escape the impression that Dan is being singled out
for his willingness to express "politically incorrect" opinions on
controversial topics, and perhaps for not tailoring his messages
to be likely understandable by the spectrum of readers of his
messages, more than for anything else. That's the thing that
nearly all of the cited messages seem to have in common. Other
explanations are possible but I don't see patterns in the cited
messages to justify other possible explanations that come to my
mind.
12. In the past there was an internet-draft that explicitly called IETF participants "racists" without supplying any justification. I judged that internet-draft to be inconsistent with the IETF Guidelines for Conduct, patently offensive, and (in my opinion and the opinion of many others) toxic and harmful to IETF. But while that draft was revised to reduce the use of such language and be more vague about who it was accusing, to my knowledge no version of that document was censored, nor were its authors sanctioned in any way. Instead, it was amplified by IETF management at the time, who gave it preferential exposure in a working group and tried to have that draft used as a basis for IETF policy.
On the other hand, when Dan wrote satirical drafts implicitly
criticizing efforts on the part of IETF leadership to support that
kind of prejudice, that draft was censored by IESG. In
addition, the IETF Chair has publicly accused it, without any
specific justification, of being "racist". That accusation was
itself, IMO, inconsistent with the recommendations in RFC 7154.
(I note in passing that RFC 7154 refers to itself as "Guidelines
for Conduct". I believe it was intended more to encourage
constructive behavior, than to be used as a set of rules to be
enforced. The recommendations are, IMO, at least as applicable
to the IETF Chair as to anybody else. But in general, I don't
think they should be used as a basis for sanctioning any IETF
participants. It is necessary to have some "room" between ideal
behavior, and behavior that deserves sanction.)
I will admit that Dan's drafts were not particularly easy to
understand in their entirety, even by me, and were presumably less
likely to be understood by readers not familiar with
English-language satirical writings. But accusations of those
documents being "racist" were unsupported and unfounded. They
were not racist, though at least one was critical of some theories
of racial inequity and attempts to remedy such inequity.
(Disclaimer: I'm making these assertions from memory of reading
those documents, as I do not currently have access to those
documents. Indeed, the removal of these documents from
Internet-Draft archives has resulted in a loss of transparency of
IESG's action.)
In my estimation, it was not disruptive or otherwise in violation
of RFC 7154 for Dan to write those documents. But neither were
the documents, in practice, likely to contribute constructively to
IETF discussion. That, in my opinion, is more due to inadequate
tailoring of the documents for the intended audience, than to any
inherent impropriety of the subject matter, argument, or intent.
(It is worth considering whether the authors of the
above-mentioned "toxic" Internet-Draft made much the same error
than Dan did, that of failing to tailor their draft to the IETF
audience.)
It's possible that satire is never a good way to communicate
subtle concepts to a global audience, because it's too likely to
be (mis)understood as insulting, especially by participants for
whom satire is unusual in their native language literature or
cultures. This is an unsurprising result of using a literary
device that treats something that's harmful as if it were
desirable, with the expectation that the reader will recognize it
as satire. It's also possible that mere mention of a
controversial topic that seems to be unrelated to the point of a
discussion is likely to be misunderstood as some sort of insult,
unless perhaps the relationship of that topic to the discussion is
made clear.
I do however acknowledge that IESG may be under considerable
pressure to make an example out of Dan, in order to signal their
willingness to adhere to a politically correct view of racial
disparity.
It's not wrong for IESG or others to point out that a message or
document is easily (mis)interpreted as abusive in some way, and to
perhaps ask that it be revised. But punishing the author seems
excessive.
14. More broadly there seems to have been a concerted effort on
the part of multiple generations of IETF leadership to marginalize
individuals who are (perhaps arbitrarily) believed to be "rude" or
"unprofessional", even though those descriptions are hopelessly
vague, and the accused individuals may have been treated quite
rudely merely for expressing different views, sometimes in the
guise of "inclusion". IMO such efforts have contributed more to
toxicity in IETF than any of Dan's messages that were cited in
this Last Call. I understand that IETF's main work product is
consensus, but a declaration of consensus is misleading at best if
it results from suppression of views or of individual
participants.
V. Recommendations for IESG and others
15. I respectfully request that this PR-action be abandoned. If
not abandoned, I request that the duration of such sanctions be
limited to a very brief interval.
16. The primary roles of the Area Directors, and probably also
the IETF Chair, should be as facilitators of technical consensus,
or occasionally mediators (not dictators) of technical compromise,
much more than as arbiters or enforcers of politeness or political
correctness or "professionalism".
There's a fundamental conflict of interest that results from the
same group of people trying to serve both roles.
Less-conventional views are often seen as "rude" or "impolite" or
even "disruptive" simply because are less-conventional or threaten
established interests, or perhaps because they touch on
controversial subjects. If IESG are now the arbiters of
acceptable speech AND the enforcers of their own rules (or their
interpretations of vague rules), this inevitably will have the
effect of suppressing views that differ from IESG's. And that's
not consistent with the mission of a consensus-building
organization, or with the ability of IESG to be seen as fair,
neutral consensus-builders. IETF needs for IESG, as much as
possible, to stay above such battles.
17. IESG should take a good look at its own role (from the past
several years to the present) in making IETF more toxic. There
are other ways to contribute to toxicity than mere rudeness, after
all, including (for example): making threats to individuals'
ability to participate (which might also threaten their
livelihood); spreading disparaging opinions or rumors about
individual participants; making and circulating lists of "problem"
participants; appropriating list moderators and others to act as
"tone police"; marginalization of individuals or their
contributions for non-technical reasons; publicly or privately
attacking participants for their views; or tampering with
consensus processes (like this one) by attempting to limit certain
speakers' ability to contribute.
IESG should seek advice from outside parties with experience in consensus-based
decision-making, (rather than, say, consultants to corporations,
military, or other strictly hierarchical organizations) to advise
it how to facilitate such discussions without encouraging
toxicity, and perhaps make such recommendations available to WG
chairs and the general community.
IESG should work to make itself to be better facilitators of
consensus or compromise, before it tries to sanction more
individuals for their speech.
I know from long experience with IETF, that IESG often narrowly
tailors working group charters in an effort to avoid strong
conflicts of interest within the working group. Perhaps IESG
should exercise at least a similar degree of effort to try to
avoid unnecessary political controversy. We are fortunate in
IETF in that most of the questions we face are between different
purely technical choices for how to solve more-or-less the same
technical problem, and are relatively uncontroversial among those
not directly affected by those technical choices. While we cannot
always always be so fortunate as to avoid tackling controversial
problems, perhaps we can at least try to avoid tackling political
questions that we are unqualified to answer.
18. Rather than trying to penalize or marginalize or "correct"
the voices of those who are frustrated or "in the rough", IETF as
a whole (including IESG, WG chairs, authors, directorates, and
ordinary participants) should try to make sure that all views are
respected and maximally understood, so that everyone involved in
the discussion can give those views due consideration, and so that
participants can have trust in the process. The best IETF
participants have always made an effort to practice and facilitate
that. It doesn't always work, but it's a much better first
approach.
19. BCP 83 should be revised in various ways, e.g.: to require
more specificity in accusations, to limit the scope of review to
specific accusations and IETF consensus-based rules, and to
evaluate only specific messages cited (also considering the
contexts in which they appear), rather than inviting the community
to find its own reasons to attack accused individuals. A PR
action should not be an invitation to a fishing expedition or
witch trial. It would also help to increase transparency of the
portion of the process before a formal PR-action is started; and
to invoke a neutral mediator before formal process is started,
with the goal of minimizing the potential for misunderstanding and
improving quality of communication where appropriate.
20. It's possible that the Guidelines for Conduct should be
revised, or other recommendations made, in at least two areas: (1)
how to respond constructively to attacks or perceived attacks, on
one's person, contributions, others, groups, or the community,
while minimizing the potential for such discussions to get out of
hand; (2) it should make clearer that it's not appropriate to
pressure people who are perceived to be "in the rough" to change
their opinions or to cease contributing to the discussion, so that
the majority can win. A third possibility worth considering would
be to add (negative) rules to the (positive) "guidelines" that
currently exist, so that the relationship between ideal behaviors
and forbidden behaviors would be obvious, and the forbidden
behaviors would have IETF Consensus behind them.
21. Community leaders such as IETF Chair, IESG members, IAB
members, WG chairs, etc., and IETF staff, should be especially
steadfast to avoid making either personal attacks, or nontechnical
criticism of people's contributions, or to try to rig a
discussion, or change the rules of an ongoing discussion, in a way
that would favor some people or opinions over others.
22. I believe IETF might benefit from having some kind of
facilitators to help contributors (including leadership) express
themselves with maximum clarity, and/or defuse discussions that
have gotten out of hand or seem to have potential to get out of
hand. Such facilitators need to be independent of IESG, and have
no authority to impose or recommend sanctions, so that they can be
maximally trusted. They should have some sort of training in
clear communications across different cultures and languages,
constructive engagement, and resolution of the kinds of conflict
seen in IETF. The primary goals of such facilitators should be
to promote respect for all discussion participants and their
contributions, to assist participants in expressing their own
ideas clearly, and to help minimize the potential for
misunderstanding, especially given differences in background,
culture, and language. The facilitators must not be "Tone
Police". They may engage in discussion with the affected
participants, point out problems, answer questions, and offer
suggestions, but should not accuse or belittle participants or
dictate language to them, and they should keep their opinions
about participants to themselves. I'd like for such facilitators
to be available to community members who request it on their own
behalf, and available to community members who complain about
others' contributions, but they should advise the complaining
party before contacting the other party. WG chairs and other
process facilitators may also recommend that participants take
advantage of the availability of such facilitators. All advice
given by facilitators should be kept entirely confidential.
I expect that such facilitators might be paid professionals, or
volunteers for whom IETF provides training, or some combination of
the two.
VI. Closing Statement
In closing I wish to point out that the privilege that someone
enjoys, or the burdens that life presents to them, are not things
that can reliably be determined by appearances. I've seen many
examples of this. Some burdens are obvious to an observer, but
many are not. Most people cannot tell by looking whether someone
has some sort of disability or serious health issue, whether
they've endured or are enduring abuse or trauma, whether they've
suffered significant personal losses, or been personally attacked
for no valid reason. I believe we would all do well to keep in
mind the saying: "Be kind to everyone, for everyone is fighting a
difficult battle" [*]. Insisting that everyone behave ideally,
all of the time, or be subject to sanction, doesn't seem
consistent with this idea.
Also, we are all living in rather trying and uncertain times. The
whole world is still suffering from a global pandemic that has
claimed 6.5+ million lives so far, leaving many more with
long-term health issues, certainly on the scale of a major war.
Many of us have lost friends and/or family to COVID, or to other
diseases whose treatment was impaired by the impact of COVID on
our health systems. Many lives and valuable institutions have
been disrupted, haven't recovered, and may not recover. There
are ongoing armed conflicts in various parts of the world that
have done and continue to do tremendous harm to innocent people
and have the potential to escalate. The potential for use of
nuclear weapons is probably greater now than at any point since
1962. Some of us, or our colleagues, may be directly affected by
some of these conflicts. For those of us in the United States,
the very fabric of our government is still under threat, and the
potential for civil war is greater in the US than at any time
since 1865.
The very Internet that is the work of our organization, has
brought tremendous benefit to the world, but also tremendous
changes that many people struggle to deal with, even when the
change was needed. We don't always think about that because
we're in the middle of it, but many of the stresses that our world
is experiencing are at least in part a result of making the
Internet real and ubiquitous, and the shifts in power that
resulted.
(whew!)
We should take all of this into account when considering how we
treat one another, and ourselves, now and probably for many years
to come. Most of us will be dealing with these effects for the
rest of our lives.
[*] Various versions of this are attributed to Scottish author Ian
MacLaren
-- last-call mailing list last-call@xxxxxxxx https://www.ietf.org/mailman/listinfo/last-call
