Proposal for addressing ongoing patterns of disruptive behaviours in our community

Angie, thanks for taking our conversations in the CWG and putting them into words. It's great to be able to bring this forward and discuss more widely with the community.

Spent the weekend at Kiwifoo. Was reminded of the practice of "Non Violent Communication" NVC and grateful to @eliza411 for introducing it to me.

I think we might want to consider modifying the action plan slightly by weaving this in.

The basic principle is pretty simple. It involved empathetic listening and honest expression of observations, feelings, needs and requests.

More here: https://www.cnvc.org/Training/the-nvc-model

Our "Action Plan" is effectively the "request" phase of NVC - We ask "would you be willing" to take certain steps...

In terms of the process outlined here, I like it. The main things I like about it are that it gives people a clear action plan, and spells out the consequences if it is not followed (in a way that lands them back in the CWG queue). I guess my question, looking at this from outside the working group is "do the people in the working group have the knowledge, experience, or training to effectively implement a good action plan for someone?" If I was someone being subjected to this process, I may very well wonder who is coming up with this plan, and do they know what they're doing? I like the idea of incorporating, or making it more transparent in the document, specific methods or processes that would be followed, like NVC. I also wonder if this brings things to a place where the community should fund actual conflict resolution training for the main people who need to handle these cases.

Somewhat related the question of who is coming up with this plan, do we some ideas for how we will handle a situation where a person can't come to an action plan agreement with the CWG? Does that skip them immediately to "out" or do we have a way to make sure a common action plan can be agreed upon?

Some of the language here ('disruptive behaviour', 'avoiding conflict') is a bit vague for me.

Not all abusive/bullying/victimizing behaviour is disruptive.

Not all disruptive behaviour and conflict involves victimization or bullying - in fact disruption and conflict can be necessary to tackle those things.

Also if there's a dispute between two people and only one of them makes a complaint to the CWG (or a complaint about the dispute comes from a third party but focuses on a particular interaction from one participant), does the CWG look at the history and actions of both parties and potentially follow this policy with both? From the language this only appears to apply to people where a formal complaint has been made.

+1 to the work so far and to the comments in #7 and #8.

I worry a bit about the risk this creates of having 3 people file reports against someone and what the timeframe should be. People who are more involved are more likely to get a report against them. I don't know that I have a solution to this risk (or if it's really a risk) but I feel a little nervous that this structure could be used as a tool to abuse someone.

One anecdote on Catch's last paragraph in #8, our community code enforcer has stopped by to give me a warning for having too many weeds in the alley. He explained that when he gets complaints from one house about another house he will always just use it as an excuse to survey the area looking for infractions. This helps him avoid situations where a code-violation-report is being used as part of an interpersonal grudge.

@add1sun, @catch, @greggles - thanks so much for coming forward and sharing thoughts.

Really good points here from all of you. The CWG met yesterday and resolved not to dive into respond in detail just yet. But we are paying attention, I just wanted to make that clear :)

I also love greggle's example in comment #9 and share the concerns in comment #8. I think it will only happen rarely, but when it does, it will be tough. One thing I would love to see is the similar policy from other projects. Have you all reached out to Wikipedia and other projects that may already have a process in place that we could learn from? I would be happy to reach out to a few places and see if I can get examples. Sharing the policies you learned from may also add legitimacy to your document for the wider community.

CWG might need to further define or reference what it means by "strike out" or "forcibly remove members from the community".

For example,

1. If a person was removed from community, is this permanent, or after (3 years), this person can come back in?
2. Does CWG privately maintain a list of people who was removed? (No one but CWG should have access to the list?)
3. What would happen to the work (issue, code, documentation, translation, etc) of the person "being removed from the community"?

I had a similar question; it's not clear from reading this whether it's meant to refer to a permanent ban or a temporary ban (and if temporary, how long) or something else?

Permanent bans can only be done in extreme cases, and I would expect those to be more of the "one strike and you're out" variety. (I mean like situations involving violence or similar, and the D.A. would have to be involved if that ever happened anyway.) An accumulation of three or more comparatively minor incidents plus failure to follow an action plan does not warrant a permanent ban - maybe a temporary ban, but even then I would question whether it should ever be more than a short "enforced cooling-off period" kind of thing.

Thanks CWG for starting this discussion! You people are saints. I could not be a member of the CWG -- I would just want to bop everyone on the head -- so just let me say I am very appreciative of your willingness to serve in what for me would be an impossible position in the community. And this seems like an excellent addition to your toolbox.

Besides the ideas mentioned above (all very good points), I had one other thought about the proposal as it is in the issue summary:

I can see that the idea is to come up with an Action Plan, but I think there needs to be a more defined process of how this action plan is arrived at. Something like:

a) The CWG writes up the Complaints and Consequences of Complaints sections, in NVC language [the CWG members who would be doing this would get some NVC training, which I think is a very good idea anyway]. I think if you really want to use NVC language, you might want to combine these sections into one, for ease of writing NVC-language statements. The language would follow the "observation" communication guidelines of NVC, whose aim is to report straightforwardly the concrete actions (things that happened), how someone felt in reaction, and the needs/desires/values of the person who was affected. So it would be something like:

When you did _____, a member of the community felt ______, because he/she wants/needs ______ as part of the Drupal community.

b) Then the CWG could write up an action plan section in the form of NVC also:

When _______ happens, we ask you to instead ______.

Note that in NVC terminology, the person who is feeling hurt is supposed to be saying all of (a) and (b) as one statement, something like:

When you do ___, I feel ___, because I really want ____, so I'm asking you to please _____.

In this case, we don't want the person in the community who was feeling bad to have to confront the offender. But still if you want to use NVC the communication of the plan should not be in terms of "You must ___ or else!" but instead in terms of "The community is asking you to ____". Just a thought...

And by the way, having an Action Plan section of the Action Plan is somewhat confusing. :) Maybe think of a different name for the overall Plan, or maybe the Action Plan section needs to be called something else?

c) At this stage I think if I were in the position of receiving this plan, I would feel very backed into a corner if it was presented as a fait accompli. So maybe the next step would be to present just (a) and (b), with (b) being a draft, with room for discussion, it might give the person the opportunity to make additional suggestions, modify the plan so that it would be more clear and/or more constructive, and for the person to get some help in understanding what he/she needs to do to help the rest of the community members feel better in these situations.

Maybe I'm being too generous or optimistic, but my thought is that even someone who is essentially behaving as a troll would hopefully have compassion for other community members and be able to think of constructive solutions to the problem, when faced with straightforward statements of what has been happening and how other people were feeling as a consequence. Or at least, they should be given the opportunity to discuss the plan of action, and make sure they understand it, and hopefully arrive at feeling it is fair and a good idea, before being presented with the "Or else!" part.

d) Something else that's missing from the description in the issue summary: There's kind of an implication, but if you define a process for creating one of these action plans, there needs to be a spelled-out remedy if the person doesn't even agree to participate in the process, or if the process falls apart at some point (like at the point where they're supposed to agree to the Action Plan part of the Plan).

e) OK, back to the process. At this point, the person has agreed to comply with the Action Plan. At this point I think they would need be presented with the consequences if they don't. And whatever is in the Consequences section, it needs to have a clear way to measure whether the plan has been complied with or not, such as "If we get another complaint and it's deemed by the CWG to indicate that this action plan is not being followed, ____". Not just "If you don't comply ____", which would be vague as to how "don't comply" is being measured. Maybe there would be a standard wording for this part?

That's all I can think of...

Thanks so much to all of you getting involved in the discussion. I’m generally encouraged and enthused by the comments as the not only align with the ongoing thinking into this process, but also provide some great input. I’m going to summarise some points and try to pull action points out of the responses so far.

@add1sun there are three really great points here and I think we have the following actions:

1. We need to be clearer somewhere about the CWG members backgrounds in community leadership and conflict resolution as there is considerable experience already available.
2. We should explicitly mention within the process that where the issues in question step outside of the capacity or expertise of the group we could and should seek outside help.
3. The plan should mention the methodologies and inspirations used throughout the process and in developing it.

@catch this is a difficult one and we had already discussed within the group that by no means should all conflict should be considered negative within the community context and that the language is hard to get balanced. We should state the context of the CWG conflict resolution process is mostly where a person or people have reported something that has caused them distress, and this should be reflected in the language. I think its also important to link back to George’s recent blog post that gives more information about the usual process for individual complaints where the situation from both "sides” is taken into consideration always. Based on this the action plan is for an individual, but there is no reason why we wouldn’t have two concurrent plans for people who are consistently hurting or disrupting each other. While complaint driven, the CWG is by no means complaint exclusive.

@greggles its unlikely that three reports about the same incident or within a very short space of time would result in the introduction of an action plan. CWG support to help individuals to solve any conflict is the first port of call and through this process we get insight into both sides of the story, hopefully lessening the risk you mention. That said the risk is there and should certainly be considered.

Both @catch and @greggles make it clear that we need to ensure that all interventions very clearly check all versions of incidents, and in a way that remains as void of presumption and judgement as possible before discussing and deciding solutions.

@holly we should certainly do further research into the way others deal with this kind of issue. We took the ideas from our own experiences of corporate action mixed with those of the non-profit and community development arenas but exposure to the approaches of other specific projects in the tech world should definitely be looked at.

@skyredwang and @David I think we can refer here to the conflict resolution policy that explicitly describes this as:

• An imposed vacation (i.e. asking someone to "take a week off" from a mailing list or IRC). They'll be asked to take this vacation voluntarily, but if they don't agree then a temporary ban may be imposed to enforce this vacation.
• A permanent or temporary ban from some or all Drupal spaces (mailing lists, IRC, etc.). The group will maintain records of all such bans so that they may be reviewed in the future, extended to new Drupal fora, or otherwise maintained.

We would expect to never have to permanently ban anyone (it hasn’t happened yet!) but rather take one of the lesser options. It is important this action remains at the disposal of the CWG in order to use it in the (hopefully) unlikely event of it ever being required.

@jhodgdon thanks a lot - we have discussed many of your ideas through the process and maybe we need a wider “Process Guidelines” that develop and document some of these ideas? It was always envisaged that the individual's Action Plan be a lightly drafted document before being shared, and that it is developed with the person in question collaboratively. Its an essential part of the process and needs to be handled sensitively but helps to reinforce the fact that this is not a punishment. The most important section that we need to ensure is available to them early on is the "consequences of the complaints” so that they can understand what impact their actions have had.

The outcome of not engaging in an action plan should be expected to follow the more extreme aspects of the conflict resolution policy as mentioned in response to @skyredwang. The comment about being able to measure compliance is a really important one and needs to be integrated into the process.

Great input - thanks again.

Is this in responce to specific incidents that have happened, or are we being pre-emptive?

Just updating the issue title slightly to fit with feedback.

This should be postponed on having a working TWG. This policy addresses the consequences (when someone speaks harsh words) but stays totally mum on the causes (not that's new).

Hi @not_chx. Can you clarify how you see the TWG to be important to this issue? I don't see a link, myself.

You want me to drag up every technical conflict that escalated into a CWG matter having no outlet for a technical resolution? It's nothing to me, it's not like I forget or forgive easily but are you sure we want to warm up all those up?

I see the two as completely separate, and this comes from someone who has been on the bad side of that equation more times than he can count. While the causes definitely need to be addressed and discussed, we have complete control over how we respond to them, and how we conduct ourselves as they are being sorted out. The two are connected but separate.

Thanks for the insights in #21, the kernel that seems important is that you feel technical issues cause interpersonal issues.

I agree with the perspective in #22.

If everyone behaved with the perfect composure as described in #22 then why would we need this policy?

So in case of severe technical disagreement whoever can keep their calm for longer will win. In other words, if you have no other purpose than to cause disruption (aka. trolling) then this is the perfect policy: the only thing the other party can do is just to walk away or they will be forced to walk away. Is this what are you saying / want?

The other party has lots of options, but acting in a way that is in opposition to community guidelines is not one of them. If you feel you are being trolled, then you should report such behavior to the CWG and they will take it up. I wish I had done that when it happened to me but I didn't (and the person in question is gone now anyways so whatever.)

What I personally want is peace and for us all to treat each other with dignity. If that comes at a cost to the technical direction of the project then I really don't care. Everyone has choices in how they behave. I'm sorry if you are frustrated and saddened by the technical direction the project has taken, but that is separate from how you react to it and the people behind it. If you can't see that, or if you feel that being on the losing end of a technical battle justifies various types of behavior, then I just don't really know what to tell you.

To be clear, the Code of Conduct does not ask people to behave with "perfect composure" at all times. The Drupal community is a place that is full of passionate people who care deeply about what they do. What the Code of Conduct does do is ask people to treat each other with respect:

The Drupal community and its members treat one another with respect. Everyone can make a valuable contribution to Drupal. We may not always agree, but disagreement is no excuse for poor behavior and poor manners. We might all experience some frustration now and then, but we cannot allow that frustration to turn into a personal attack. It's important to remember that a community where people feel uncomfortable or threatened is not a productive one. We expect members of the Drupal community to be respectful when dealing with other contributors as well as with people outside the Drupal project and with users of Drupal.

I firmly believe that a culture that encourages healthy debate and disagreement is a big part of what gives the Drupal project and community its strength. However, if we can't have those conversations without consistently engaging in personal attacks, intimidation, harassment, and other destructive behaviors that make the Drupal community a less welcoming place, then we've got much bigger problems. Enabling a toxic culture, whether deliberately or through inaction, will eventually lead to a project that has fewer contributors and less innovation.

As Jono Bacon put it in a recent article, in an ideal world, everyone should come to an open source community with the desire to find solutions. If there are people who aren't contributing in a constructive and solutions-oriented way, then we need to provide mentoring, support, and guidance to help guide them to success. But if that effort fails, and they're unwilling or unable to work within the community constructively, "the right thing to do is to ask them to find somewhere else to participate and share their energy and talents." He argues that failing to do so compromises "the very fabric of what makes open source incredible: human relationships connected by a core mission and ethos, and underlined by the spirit and acknowledgement of doing."

The goal of this proposal is to provide that kind of empathetic, mentored, and considerate approach.

> The other party has lots of options

Like escalating the technical problem to the TWG which is nonfunctional? If not that then what? I am very eager to hear!

> What I personally want is peace [...] If that comes at a cost to the technical direction of the project then I really don't care.

OK, I will stop commenting here.

But, no, I have one more comment to make over #27: this means Drupal is no longer a meritocracy. There is nothing wrong with that -- I just wanted to shore up one other piece of the new governance structure that replaces the meritocracy and felt it is important to be in place before we close the book on that chapter.

There are plenty of presentations, blog posts, videos, podcasts, discussions about how Drupal, the technology world, etc. have never been an actual meritocracy (or at least not a meritocracy that works strictly on the merit of the patches/ideas). Claiming that the status-quo-pre-working-group was a functioning meritocracy is inaccurate.

IMO the working groups are making progress toward making Drupal *more* meritocratic (e.g. preventing abusive behaviors from pushing out contributors with good ideas but low tolerance for abuse).

Well. In light of today and in light of the past, although I kept my cool so far, now hear me out: this policy is poison. Discussions online realtime and in the issue queue get completely misunderstood out of context and make any sort of disagreement completely impossible. You will chase out everyone even remotely passionately about anything.

You try to baby everyone so much that everything that can be misunderstood will be misunderstood. "A detailed list of the incidents that have been reported about the individual." -- what you will foster is a witch hunt. The Drupal community already has an excellent track record of finding people who can find some words offending even if those words are not directed to them.

Just think of it: are we better off with Morten off the DA? Absolutely not! And yet, it happened. That's where you will end up with every conflict, people leaving.

Yes, I actually did above:

> This should be postponed on having a working TWG. This policy addresses the consequences...

How many issues did we have which essentially were "- We made a decision. Please stop derailing the issue - But I disagree with you!" stalemate. There's nowhere to go from there. This discussion should continue "- Fine, if you disagree that much and can prove it then escalate but please stop derailing us for now. - OK, I wil do that" but it can't because there's nowhere to escalate technical disagreements and only issue status/severity ping-pong and frustration follows.

#29 got zero replies and as far as I am concerned, there are no options besides walking away and let smooth talking snake oil salesmen win and get to go on to talk at every conference in the name of the Drupal community and even get keynotes at DrupalCon. Maybe there are, but in the last month noone said what.

Like escalating the technical problem to the TWG which is nonfunctional? If not that then what? I am very eager to hear!

I didn't answer #29 because the seems so obvious: make sound, logical, technical arguments in a manner that shows respect and compassion for fellow community members. As I re-read comments from other people in this thread before #29...they suggest similar behavior. So your claim in #34 that #29 was not addressed...I don't get because it was addressed before you even asked it. Basically you are going in circles and not listening to people. It feels like you are just trying to discredit this proposal with sweeping statements.

You insist on conflating this policy with a working TWG but haven't, as far as I've seen, explained why they must be connected. You keep insisting they must be intertwined in spite of a lot of very reasonable people saying they don't. My only conclusion is that you are not explaining well or you are holding on to something that you should let go.

> I didn't answer #29 because the seems so obvious: make sound, logical, technical arguments in a manner that shows respect and compassion for fellow community members.

And when that fails?

I had a very long and eventually fruitful conversation with KatteKrab here and I have a very good summary how I see these issues. I will pick a particular one but only because it's the most recent I could recall and not because of the dramatis personae in it.

So https://www.drupal.org/node/2391025 -- nothing good comes out of an issue like this where one person thinks he has a valid technical concern so much as to override the opposing opinion of a subsystem maintainer and two core committers. This is a classic escalating cycle of violence and you need a hobbesian Leviathan to step in and break it up. I called this entity the TWG but I am not too picky -- it needs to be technically savvy, fair and its decision not questionable. (If thinking in the terms of courts, subsystem maintainers and core committers are already two levels you already lost)

[edit: flowchart temporarily pulled, pending update]

Flowchart from Community Summit at DrupalCon LA:

Flowchart from Community Summit at DrupalCon LA (as a text list):

• Incident Report filed
• Evaluate in relation to first/initial CR (conflict resolution) process
  • Has CR process been followed?
    • If No, is it applicable in this case?
    • If Yes, follow CR process.
      • Did it work? If Yes, Yay!!
  • If CR process is not applicable in this case, or it has been followed and did not work, then:
    • *Review Evidence (from Incident report): Is there clear and credible evidence of a pattern of bad behavior?
      • If No, Document it and wait.
      • If Yes,
        • Co-Create action plan*
        • Implement action plan*: Did it work?
          • If Yes, Yay!
          • *If No, Re-evaluate situation: Are there other options?
            • If Yes, repeat action plan cycle.
            • If No, Punitive action and/or escalation.

* (asterisk) indicates: These processes need to be better defined. Who's doing it and what are the rules? (may need their own flowcharts)

A redraft has now been posted to the CWG Policies section of the our governance handbook.

We'll get a pretty URL for it soon, but in the meantime, see: https://www.drupal.org/node/2504989

Thank you to all of you for engaging with this issue in the way you did! Very helpful.