ietf
[Top] [All Lists]

Review of: draft-farrresnickel-harassment-06

2015-03-11 12:09:25
Summary:

This documents defines harassment and establishes IETF policies and
practices regarding it.  It is an essential effort.

   1.  The document's language appears to repeatedly confuse harassment
actions with other forms of interpersonal difficulties.  It needs to
draw a bright line about the handling of harassment, versus complaints
of harassment deemed to be something else.

       All discussion of harassment, itself, needs to invoke a
perspective that it is a unilateral -- not bilateral -- action.  That is
harassment is essentially an offense against the community.  The Subject
is in no way a contributor to the problem nor, really, a contributor to
its remedy.  Given them a voice about the remedy is different from
holding them responsible for negotiating it or otherwise being treated
as any part of the cause.

   2.  The document constrains itself to classic, formal harassment.
That is, actions against protected classes.  It should expand its scope
to cover other forms of unilateral attack, often called bullying.  The
view that the IETF already has policies and procedures against such
behavior is objectively incorrect.  Further, the IETF has no pattern of
practice against such behavior.  Quite the contrary.  Worse, an
environment that tolerates -- or even promotes -- such behaviors lays a
foundation for the offense of harassment.

   3.  The mechanisms for choosing and training the Ombudteam and for
handling appeals should be considered and explained more carefully.  The
current text appears to have significant problems about these topics.



Detailed Comments:


Abstract

IETF Participants must not engage in harassment while at IETF 
meetings, virtual meetings, social events, or on mailing lists.
This document lays out procedures for managing and enforcing this
policy.

Although this document does not assert use of the IETF's normative terms
(RFC 2119), it should refrain from using the reserved words.  Those
words are so thoroughly ingrained in IETF culture and conversation,
non-normative use of them invites confusion.  Note the nicely-convenient
suggestions in:

     https://tools.ietf.org/html/draft-hansen-nonkeywords-non2119-02

However, given that the document really does make normative assertions,
it ought to use the vocabulary common to the IETF for this.


In any event, I suggest an even more forceful and active-voiced opening
sentence:

     The IETF does not tolerate harassment at IETF meetings, virtual
meetings, social events, or on mailing lists.  This document defines
harassment and lays out procedures for managing and enforcing this policy.


1. Introduction


The IETF has general policies for managing disruptive behavior in
the context of IETF activities.  In particular, [RFC7154] provides a
set of guidelines for personal interaction in the IETF, and [RFC2418]
and [RFC3934] give guidelines for how to deal with disruptive
behavior that occurs in the context of IETF working group
face-to-face meetings and on mailing lists.

However, there is other problematic, often more interpersonal, 
behavior that can occur in the context of IETF activities (meetings, 
mailing list discussions, or social events) that does not directly 
disrupt working group progress, but nonetheless is unacceptable 
behavior between IETF Participants.  This sort of behavior,
described

Long, somewhat awkward sentence.  Shorter and simpler is better:

   However in the context of IETF activities (meetings, mailing list
discussions, or social events) there is other problematic behavior,
which is often more personal. Even if it does not directly disrupt working
group progress, it is nonetheless unacceptable behavior among IETF
Participants.


in the IESG Statement on an "IETF Anti-Harassment Policy" [1], is
not

Folks, as a consensus document of this type, the core of the topic needs
to be fully contained in this document and not rely on definition or other
specification from another IETF group, such as the IESG. If it needs to
draw from an external document, then it should be one from experts, such
as Stewart Bryant suggests.

However since this is a normative specification it ought to include the
definitional text directly, even if it cites an expert definition.
.
The IESG does not define harassment.

Worse, with this document's use of a citation through a URL to a web
page, the IESG could choose to change the definition whenever it wants.

For a consensus document like this, the community MUST be the authority
on definition of harassment or else its adoption of a standard, legal
definition.



As for the definition in the IESG page:

      Harassment is unwelcome hostile or intimidating behavior -- in
particular, speech or behavior that is sexually aggressive or
intimidates based on attributes such as race, gender, religion, age,
color, national origin, ancestry, disability, sexual orientation, or
gender identity.

The language here covers classic protected classes, but not other forms
of personally hostile behavior that serves to intimidate or coerce other
participants, as outlined in Section 2.2 of:

     https://tools.ietf.org/html/draft-crocker-diversity-conduct-02

In a culture with a very long history of actively tolerating and
occasionally encouraging personally offensive behavior and, worse,
behavior designed to intimidate participants -- and especially an
organization
with such a poor track-record for holding accountable those who
indulge in such behaviors -- the exclusion of this larger class of crippling
behaviors means that we are essentially addressing only symptoms and not
a root problem.

In practical terms an environment that ignores ingrained patterns of
intimidation behavior can do little of significance about behaviors
against identified classes.

Lest there by any misunderstanding of the above:  hostile actions
against protected classes have to be stopped; but so do patterns of
hostile actions
against others.  A culture that persists with the latter maintains a
foundation that breeds the former.


easily dealt with by our previously existing working group
guidelines

To anticipate a rejoinder to my pressing for a broader scope:  Nothing
in our current documents or practices easily deals with bullying, either.



2. Definitions


The following terms are used in this document:
...
The IESG statement on harassment [2] gives a general definition of 
harassment as "unwelcome hostile or intimidating behavior -- in 
particular speech and behavior that is sexually aggressive or 
intimidates based on attributes like race, gender, religion, age, 
color, national origin, ancestry, disability, sexual orientation, or 
gender identity."  This document adopts that general definition, but

Follow the definitions format for the term.  Just provide a definition
and/or cite an expert definition.

 It is not necessary to cite the IESG effort at all.  It's actually
distracting to cite it.


does not attempt to further precisely define behavior that falls 
under the set of procedures identified here, nor does it attempt to 
list every possible attribute that might be the basis for
harassment, except to note that it may be targeted at an individual,
directed at a specific group of people, or more generally impact a
broader class of people.  In general, disruptive behavior that occurs
in the context of an IETF general or working group mailing list, or
happens



Resnick & Farrel        Expires September 6, 2015               [Page
3]


Internet-Draft         Anti-Harassment Procedures             March
2015


in a face-to-face or virtual meeting of a working group or the IETF 
plenary, can be dealt with by our normal procedures, whereas

As worded, this creates an anomaly that probably isn't intended. The
focus on 'disruptive' ignores bullying behavior, if it doesn't happen to
be perceived as disruptive. So intimidation that doesn't invoke a
protected class is OK, as long as it doesn't bother anyone but the Subject?


harassing behavior that is interpersonal is more appropriately 
handled by the procedures described here.  However, there are 
plausible reasons to address behaviors that take place during 
working-group meetings using these procedures.  This document gives 
some guidance to those involved in these situations in order to 
decide how to handle particular incidents, but the final decision 
will involve judgment and the guidance of the Ombudsteam.

Any definition of harassment prohibited by an applicable law can be 
subject to this set of procedures.

What does that sentence mean?



3. The Ombudsteam

The document should first state what an Ombudsteam is, in terms of
responsibilities and authorities.  The mechanics are important, but less so.


This section describes the role of the Ombudsteam in terms of the 
appointment of Ombudspersons, their qualifications and training, the 
length of the term of service, any recompense for their service, and 
how they may be removed from service.  The general operational 
procedures for the Ombudsteam are described in Section 4, Section 5, 
and Section 6.

3.1. Size of the Ombudsteam


The Ombudsteam shall comprise no fewer than three people.  From time 
to time, the size may fall below that number owing to changes in

may -> might


membership, but the team will be rapidly brought up to size through 
new appointments.  The team may be grown to a larger size as 
described in Section 3.2

3.2. Appointing the Ombudsteam


The Ombudsteam is appointed by the IETF Chair.  The appointment is 
solely the responsibility of the IETF Chair who may choose to
consult with members of the IETF community.

Since there are multiple members, I suggest appointing through multiple
sources, to get a degree of heterogeneity of organizational affiliations
and accountabilities.




3.4. Qualifications and Training


It is not expected that there will be candidates with all of the 
necessary ombudsperson skills and training who also have a clear 
understanding and familiarity with the IETF processes and culture. 
The Chair might choose someone with a great deal of professional 
experience evaluating and mediating harassment disputes, but little 
exposure to the IETF, or could select someone with more exposure to 
the IETF community, but without as much experience dealing with 
issues of harassment.  Since all of these attributes may be regarded 
by the IETF Chair as essential for the team, the IETF is committed
to providing training (or funding for it) as deemed necessary for 
appointed Ombudspersons.  In determining the appropriate training, 
the IETF chair and Ombudsteam shall take professional advice and
will consult with the IAOC with respect to the overall IETF budget.

As phrased, this defers the question of training.

The nature of this work is far outside the experience base of typical
IETF participants.  (Not all, of course, but many and probably most and
maybe nearly all.)  Good intentions and a logical mind are not enough,
nor is basic management training.  This topic is a difficult specialty.

Relying on spontaneous, real-time training, by random members of other
members, when there is a complaint to handle, creates a process that is
far too ad hoc for something this sensitive and this important.

So I strongly suggest mandating some basic, appropriate professional
training for anyone placed on the team who is not certified or extremely
experienced with this sort of work.

Again: having taken a management course or having supervised some folk
is not enough; this is a specialty, even within Human Resources and it
is common for this topic to be handled quite badly, even by HR folk.  As
an example, it is common to try to treat this as a disagreement or
misunderstanding between equals, rather than as a direct and often
illegal act of aggression by one against another.



3.6. Recompense


An Ombudsperson shall receive no recompense for their services.
This includes, but is not limited to:
...
o  Travel or accommodation expenses

The IETF will, however, meet the costs of training when agreed to by 
the IETF Chair as described in Section 3.4.

Given possible, additional (and spontaneous) travel for ad hoc meetings,
This effectively excludes all but the most well-funded participants.



3.8. Disputes with the IETF Chair regarding the Ombudsteam


If an individual should disagree with an action taken by the IETF 
Chair regarding the appointment, removal, or management of an 
Ombudsperson or the Ombudsteam, that person should first discuss the 
issue with the IETF Chair directly.  If the IETF Chair is unable to 
resolve the issue, the dissatisfied party may appeal to the IESG as
a whole.  The IESG shall then review the situation and attempt to 
resolve it in a manner of its own choosing.  The procedures of 
Section 6.5.4 of [RFC2026] apply to this sort of appeal.

Why no appeal sequence above that?

The IETF Chair and the IESG form a collaborative unit.  While the IESG
is certainly capable of deciding against the Chair, its direct and
on-going affiliation with the Chair creates a degree of conflict of
interest.

It's probably OK to leave the IESG in the appeal chain, but it seems
problematic for it to be the only recourse.



4. Handling Reports of Harassment


Any IETF Participant who believes that they have been harassed, or 
that any other IETF Participant or group of IETF Participants has 
been or may have been harassed, may bring the concern to the 
attention of any serving Ombudsperson.  This can be done by email to 
"ombudsteam(_at_)ietf(_dot_)org" [3], or can be done directly to a chosen 
Ombudsperson.  Direct contact information for the Ombudsteam, 
including the email addresses to which "ombudsteam(_at_)ietf(_dot_)org" is 
forwarded, can be found at https://www.ietf.org/ombudsteam .

When a Reporter brings an incident of potential harassment to the

  potential -> existing or potential


attention of the Ombudsteam, a single Ombudsperson shall be

So, the use of 'shall' sounds normative...


designated as the primary contact person (the Lead Ombudsperson) for 
the report.  When the Reporter contacts a single Ombudsperson, that 
Ombudsperson shall be the Lead Ombudsperson for the report unless 
mutually agreed between the Reporter and that Ombudsperson.



Resnick & Farrel        Expires September 6, 2015               [Page
6]


Internet-Draft         Anti-Harassment Procedures             March
2015


The Lead Ombudsperson may share any information regarding the report

Is 'may' meant to be normative?

(Yes this is pedantic, but again, we have established practice in the
IETF for normative language.  Is there a reason to ignore it?)


with the rest of the Ombudsteam except when an Ombudsperson is 
recused (see Section 7).  If a Reporter believes that a member of
the Ombudsteam should recuse themself, they should make this known to
the Lead Ombudsperson as soon as possible.

Seems like the need for recusal is independent of whether the person
chooses to do it themself.  That's merely one path towards recusal.

Instead:

   should recuse themself, -> should be recused,

Recusal also probably should be discussed in its own subsection, rather
than as a side comment in the sub-section on reporting harassment.


The Lead Ombudsperson will discuss the events with the Reporter and 
may give advice including recommendations on how the Reporter can 
handle the issue on their own as well as strategies on how to
prevent

Handle the issue on their own?  Really?  Harassment?

This line of thinking is entirely inappropriate.

It is like saying that when someone reports that they've been beaten up
the police may give advice including recommendations on how the
complainant can handle the issue on their own.

My guess is that there is confusion between 'harassment' and
'disagreement' or confusing between a complaint of harassment that is
legitimate, versus one that is not.

That is, an essential part of processing a complaint of harassment is to
determine whether harassment occurred.  If it is determined not to have
been harassment, it's possible that the complainant can and should
pursue other avenues on their own.

If, however, harassment has taken place, it is probably illegal in many
jurisdictions to counsel that they handle it on their own.  In any
event, it is entirely inappropriate.  In fact, it essentially repeats
the harassment, by providing explicit institutional failure to deal with it.


the issue from arising again.  The Lead Ombudsperson may also 
indicate that the issue would be best handled using regular IETF 
procedures (such as those for dealing with disruptive behavior) 
outside the context of harassment, and in this case the Lead 
Ombudsperson will provide assistance in using the relevant IETF

If you mean that this isn't harassment and therefore needs to be handled
in some other fashion then say so explicitly, including the the basis
for that assessment.


procedures.  Otherwise, with agreement to proceed from the Subject 
(or the Reporter if there is no individual Subject), the Ombudsteam 
may initiate detailed investigation of the matter, and may 
subsequently impose a remedy as described in Section 5.

4.1. Ombudsteam Operating Practices


The Ombudsteam is responsible for devising and documenting their 
operating practices.  These practices must be discussed with the
IESG and published in a publicly visible place (such as on the IETF
web site).  Discussion with the IETF community is encouraged and,
while IETF consensus is not necessary, significant objection to the 
processes that are not addressed should result in an [RFC2026] 
Section 6.5.3 appeal and/or a recall petition against the IETF Chair 
(and the rest of the IESG if appropriate) if they do not address the 
concern.

The 'should' here is odd.  I believe the intent is to declare that
appeal is permitted according to RFC 2026.

It's possible that there is (also?) an intent to impose on the IESG a
requirement that it attend to 'significant objection' in a responsive
manner?  That certainly would be reasonable to have in this document.


The practices must include at least the following high-level 
components:

Each member of the Ombudsteam is expected to be present at the 
majority of IETF meetings and to be available for face-to-face 
discussions.  The Ombudsteam is expected to arrange itself so that 
there is coverage of every IETF meeting by at least one 
Ombudsperson.

Coverage?

Perhaps what is meant is that there must be at least one Ombudsperson at
every IETF meeting, in person, to be available for hearing reports and
that the community must be informed of who is available and how to
contact them?


All information brought to the Ombudsteam shall be kept in strict 
confidence.  Any electronic information (such as email messages) that
needs to be archived shall be encrypted before it is stored using
tools similar to those used by NomCom.

Is that sufficient protection?  Consider who can see NomCom messages.
Is similar exposure for this information acceptable?


When conducting a detailed investigation of the circumstances 
regarding the complaint of harassment, the Ombudsteam may contact the
Respondent and request a meeting in person or by a voice call.

" and request a meeting in person or by a voice call." seems like
unnecessary detail.



Resnick & Farrel        Expires September 6, 2015               [Page
7]


Internet-Draft         Anti-Harassment Procedures             March
2015


The Respondent is not obliged to cooperate, but the Ombudsteam may 
consider failure to cooperate when determining a remedy (Section 5).

The Ombudsteam shall endeavor to complete its investigation in a 
timely manner.

Any individuals who make a good faith report of harassment or who 
cooperate with an investigation shall not be subject to retaliation
for reporting, complaining, or cooperating, even if the
investigation, once completed, shows no harassment occurred.

Is retaliation acceptable when the report was not in good faith?  Is it
acceptable when they do not cooperate?

(Yes, I mean the question seriously, given how the draft has been worded.)


Anti-retaliation is noted here to alleviate concerns individuals may
have with reference to reporting an incident they feel should be
reviewed or cooperating with an investigation.

In all cases the Ombudsteam will strive to maintain confidentiality
for all parties including the very fact of contact with the
Ombudsteam.

When investigating reports of harassment and determining remedies,
it is up to the Ombudsteam whether they choose to act as a body or 
delegate duties to the Lead Ombudsperson.

5. Remedies
...
If the Ombudsteam determines that harassment has taken place, the 
Ombudsteam is expected to determine the next action.

In some cases a mechanism or established IETF process may already 
exist for handling the specific event.

Huh??  For harassment???


In these cases the Ombudsteam may decide that the misbehavior is best
handled with the regular IETF procedures for dealing with disruptive
behavior

Again, this confuses 'misbehavior' or 'conflict' with 'harassment'.


and may assist the Reporter to bring the issue to the attention of 
the working group chair or IESG member who can deal with the 
incident.

In other cases there is a spectrum of remedies that may be 
appropriate to the circumstances.  At one end of the spectrum, the 
Ombudsteam might choose to discuss the situation with the Respondent
and come up with a plan such that there is no repeat of the
harassment.  With the agreement of both parties, the



Resnick & Farrel        Expires September 6, 2015               [Page
8]


Internet-Draft         Anti-Harassment Procedures             March
2015


Ombudsteam can also help to mediate a conversation between the 
Respondent and the Subject (or the Reporter if there is no individual
Subject) in order to address the issue.

In no circumstances is "mediation" in any way appropriate or relevant in
a case of actual harassment.  Again, the document is confusing
harassment with various forms of interpersonal difficulties.

     Harassment is an attack by one person against one or more others in
a protected class.

     Bullying is an attack by one person to intimidate one or more others.

Under no circumstances do either of these warrant 'mediation', any more
than someone who had beaten another person up warrants mediation.


At the other end of the spectrum the Ombudsteam could decide that the
Respondent is no longer permitted to participate in a particular IETF
activity, for example, ejecting them from a meeting or requiring that
the Respondent can no longer attend future meetings.  (The Ombudsteam
can not impose that a Respondent who is in a IETF management position
be removed from that position.  There are existing mechanisms within
IETF process for the removal of people from IETF management positions
that may be used as necessary.)

This course of remedy appears to confuse disruptive behavior with
harassment, presumably thinking that the harassment is only relevant to
the attacked person.



5.1. Purpose of Remedies


The purpose of the anti-harassment policy is to prevent all
incidents of harassment in the IETF.  The set of procedures
documented here serves to provide a mechanism whereby any harassment
that occurs can be reported and handled both sympathetically and
effectively.  The policy also sends a clear message that the IETF
does not tolerate harassment in any form.

Actually, the policy does not do that.  Effective enforcement of the
policy does that.


However any remedy is imposed to try to make sure that the incident 
does not escalate and to ensure that a similar situation is unlikely 
to occur with the same Respondent in the future.

"Does not escalate"? Not sure what that means.

In any event, I think the intent here is more along the lines of:

   ...to make sure that the harassment is immediately terminated and to
ensure that a similar behavior does not recur.

Harassment is not situational. Also the "with the same respondent in the
future" implies that it would be OK for harassment to occur with someone
else.  I suspect you don't intend that...



Because the handling of incidents of harassment (including the 
imposition of remedies) is confidential, an imposed remedy cannot 
itself serve as a deterrent to others, nor can it be used to "teach" 
the community how to behave.  ([RFC7154] gives guidelines for
conduct in the IETF.)  Furthermore, a remedy is not to be imposed for
the purposes of retribution.  However, the knowledge of the existence
of



Resnick & Farrel        Expires September 6, 2015               [Page
9]


Internet-Draft         Anti-Harassment Procedures             March
2015


a range of remedies and of processes by which they can be applied 
serves both as a statement of the IETF's seriousness in this matter, 
and as a deterrent to potential offenders.

While the details of a harassment case must not be made public, it could
be reasonable and useful for a sanitized summary to be reported.

This accomplishes two things.

     First, it provides some instruction to the community that the
policy is serious.  For real serious, not just good intentions with no
teeth.

     Second, it permits a degree of community oversight.

Both of these are important benefits and are currently entirely lacking.


The Ombudsteam is expected to apply the above considerations, as
well as proportionality and reasonableness, in selecting a remedy.
They are asked to consider the impact of the remedy on the Respondent
as well as on the Subject.

6. Disputes with the Ombudsteam


If either the Subject (or the Reporter if there is no individual 
Subject) or the Respondent is dissatisfied with the decision of the 
Ombudsteam, the dissatisfied party should first contact the Lead 
Ombudsperson and discuss the situation.  If the issue cannot be 
resolved through discussion with the Lead Ombudsperson, the issue
may be raised with the IETF Chair.

If necessary, the IETF Chair may recuse themself from any part of

{insert repeated concern for the 'themself' perspective.}


this process (see Section 7) and request the IESG to select another 
of its members to serve in this role.  This IESG member is known as 
the "delegated IESG member".

The IETF Chair (or the delegated IESG member if the Chair is
recused) will attempt to resolve the issue in discussion with the
dissatisfied party and the Lead Ombudsperson.  If this further
discussion does not bring a satisfactory resolution, the Ombudsteam's
decision may be formally appealed.  The appeal is strictly on the
issue of whether the Ombudsteam exercised due diligence in both their
decision as to whether harassment had taken place, as well as in
their determination of any appropriate remedy that was imposed.  In
particular, the purpose of the appeal is not to re-investigate the
circumstances of the incident or to negotiate the severity of the
remedy.

So a dispute about the assessed facts or a dispute about the remedy are
out of scope for appeal???


All elements of the appeal, including the fact of the appeal, will
be held in confidence, but will be recorded and held securely for
future reference.

By whom?  Under what circumstances?  With access permitted by whom?


The appeal will be evaluated by the IETF Chair (or the delegated
IESG

   will be evaluated by an Appeals Group, comprising the IETF Chair...

member) and two other members of the IESG selected by the IETF Chair 
(or the delegated IESG member) and confirmed by the appellant.  This 
Appeals Group shall convene as quickly as possible to evaluate and 
determine the appeal.  Where the impacts are immediate and related
to participation in an ongoing meeting, this shall happen in no more 
than 24 hours after receiving the appeal.  The Appeals Group may ask 
the appellant and the Lead Ombudsperson for statements or other 
information to consider.  If the Appeals Group concludes that due 
diligence was exercised by the Ombudsteam, this shall be reported to

This gives three IESG members access to detailed information about the
complaint.  Again, this will be a group of people with little-to-no
training in the handling of such matters.

Is this the proper set of folk for handling an appeal on this sort of
matter?



7. Conflicts of Interest


In the event of any conflict of interest, the conflicted person 
(member of the Ombudsteam, member of the Appeals Group, IETF Chair, 
etc.) is expected to recuse themselves.

A conflict of interest may arise if someone involved in the process 
of handling a harassment report is in the role of Reporter, 
Respondent, or Subject.  Furthermore, a conflict of interest arises 
if the person involved in the process of handling a harassment
report is closely associated personally or through affiliation with
any of the Reporter, Respondent, or Subject.

A complaint against an IESG member would seem to require recusal by
everyone on the IESG.


For the avoidance of doubt, recusal in this context means completely

  Delete: " For the avoidance of doubt"


stepping out of any advisory or decision-making part of any process 
associated with handling a harassment report, remedy arising from a 
harassment report, or appeal into the handling of a harassment 
report.  That means that a recused person has no more right to 
participate in or witness the process than any other person from the

That's not quite what the paragraph's first sentence says or means.  It
explicitly lists two limitations but does not preclude hearing details
of the case.




-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net


<Prev in Thread] Current Thread [Next in Thread>
  • Review of: draft-farrresnickel-harassment-06, Dave Crocker <=