Disruption of work

Maybe disruption of work should be a little more defined. Taking things very literally, people (legitimately) -1'ing my patches disrupts my work, but that's obviously not the type of disruption meant here. Bawolff (talk) 00:34, 7 August 2015 (UTC)Reply

@Bawolff: I based that on "sustained disruption of talks or other events" at wmf:Friendly space policy; I've rephrased to make it more clear (and a little more narrow for now). Mattflaschen-WMF (talk) 03:15, 7 August 2015 (UTC)Reply

Added a heading. --MZMcBride (talk) 23:27, 10 August 2015 (UTC)Reply


Again, as the WMF Harassing Policy for physical spaces, this implicitly reverses the burden of proof for non-listed harassment and creates first and second class victims. I don't want to support any such discrimination, much less assume any liability for the results, so if this or a similar policy is enacted, it should point out who is responsible for it and that conforming to it does not condone behaviour complying with this policy. --Tim Landscheidt 01:17, 7 August 2015 (UTC)Reply

Where do you draw the conclusion that this creates two classes of victims? It's quite clear that the list is not exclusive; whether or not it's listed, if it's harassment, it's forbidden. I think not listing an examples of harassment and simply saying "Do not harass contributors or users." is far too vague. I'm very open to suggestions of concrete things that are missing from the list, though. The document already notes the Community Advocacy team's existing role in this area, and the final document will have more clarity as to who to contact.
You cite a pre-existing policy (I think you're referring to wmf:Friendly space policy, which this is partly based on). That has a similar phrase, "Harassment includes but is not limited to". Do you recall any harassment that occurred but could not be dealt with due to not explicitly being listed? Mattflaschen-WMF (talk) 03:15, 7 August 2015 (UTC)Reply

This needs some more thought

I very much like this policy but:

  1. I feel like it could afford to be a lot more explicit. One source of influence (which also draws a lot from the Ada and CoC elements) is the jQuery Coc (COI: I've provided some comments to it). It's very clear about the things that are not okay, and more importantly it's very clear about the processes around the escalation paths, which this isn't.
  2. Community Advocacy is absolutely the wrong entity to be handling requests here; they are awesome but do not trend towards a technical background. Why aren't the Engineering Community team handling this? This is quite literally their job - building a healthy community.
  3. Who drafted this and does that group include members of groups traditionally marginalised within technical environments? If not, why not? Asking people to comment publicly is all well and good but it ignores how toxic these venues can get. We could lose a lot of incredibly useful submissions based around lived experiences by not including those groups early in the drafting process. Ironholds (talk) 16:55, 7 August 2015 (UTC)Reply
I particularly agree with points 1 and 2. One of the things I like about the Contributor Covenant's version (which I think we should fork and modify for this; writing an effective CoC is hard, and we should take advantage of existing work!) is the specificity: methods of contribution, spaces the policy applies to (particularly "in public spaces when an individual is representing the project or its community"), unacceptable behavior, means of reporting, and consequences for violations.
I think that CA is a great resource to escalate to for really bad violations (think assaults at conferences, doxxing, etc.), but I think that we need people who are familiar with technical workflows and the normal structure of technical discussion to take point on this. Harassment thrives on plausible deniability, and someone who's not familiar with the normal operation of the space in question won't be as effective at identifying it and handling it effectively. Fhocutt (WMF) (talk) 19:31, 7 August 2015 (UTC)Reply
This is still a draft, and I encourage people to keep editing it, including using other CoC's as inspiration (or maybe rebasing it; see below), and including being explicit about what is not permitted. I'm in communication with both the Community Advocacy and Engineering Community teams, and the exact roles (if any) for each are still being worked out. The document does not currently say that CA will be the go-to contact person for this (hence the ?); it just states their pre-existing authority, partly so it won't be misinterpreted as taking that away.
I'll let the drafters speak for themselves if they choose (partly because I don't want to imply this is their ideal draft). The drafters included at least one participant from a traditionally marginalized group, but more input to the draft (from everyone) is needed. I'm not just asking for comments, but actual edits to the draft. Mattflaschen-WMF (talk) 04:29, 8 August 2015 (UTC)Reply

Two questions

  1. Why is this needed? (What generally comes up now as problems, how do existing channels fail, and how will this resolve that?)
  2. Whose consensus will it come down to in order to actually approve, enforce, and modify this?

Thanks. -— Isarra 18:09, 7 August 2015 (UTC)Reply

I'm very curious about the answers to these two questions as well. --MZMcBride (talk) 19:53, 7 August 2015 (UTC)Reply
Re Question 1: This is a pretty good writeup: https://adainitiative.org/what-we-do/conference-policies/ Greg (WMF) (talk) 00:38, 8 August 2015 (UTC)Reply
This is needed because harassment is a common problem in technical communities, and unfortunately, we are not uniformly an exception to that. Saying existing channels "failed" completely is overstating it, but they have not performed as well as they would have with a clear, binding policy. This will provide a clear and explicit (albeit not 100% complete) list of forbidden behavior, and make clear how it is dealt with (understanding this may be a multi-step process in some cases). We are still determining how this will become binding policy and who will participate in enforcing it. Mattflaschen-WMF (talk) 04:43, 8 August 2015 (UTC)Reply
Sure, but it would be great if we could develop policies based on case studies of where things have failed before (Especially case studies of where we have failed) as opposed to more generic bad things can happen. Bawolff (talk) 21:46, 8 August 2015 (UTC)Reply
Why? Why wait for something bad to happen to call it out instead of saying "these kinds of things are bad, don't do them here". Greg (WMF) (talk) 01:02, 9 August 2015 (UTC)Reply
We already have that. See meta:DICK. -— Isarra 02:36, 9 August 2015 (UTC)Reply
this is a pretty succinct response to why "don't be a jerk" isn't enough to create a friendly atmosphere - the fact that the guidance you're linking to (which is just that; guidance. It isn't even enforceable) opens with a big picture of Wikipetan is...well, it somewhat undermines the point here. Ironholds (talk) 17:11, 9 August 2015 (UTC)Reply
'Don't be a dick' pretty thoroughly addresses the examples given. -— Isarra 18:42, 10 August 2015 (UTC)Reply
I'm not sure we even have the capacity to thoroughly identify where things have failed (discouraging contributors is usually not a very obvious phenomenon, because the persons getting discouraged leave, and no one is left to talk about it), so it's probably easier to build on the work of other communities which did that kind of investigation and came up with guidelines. It seems unlikely that the MediaWiki developer community would be so fundamentally different from the Ubuntu or Gnome or Python community that we somehow need completely different rules.
That said, I can share a case study, although I don't think it's the worst or most important type of harassment. During the second half of 2014, I have been triaging feedback on mw:Extension talk:Media Viewer/About, which some people have decided to use as a channel for venting over their dislike of deployment governance. You can get a picture of it just by skimming the page, although some of the meanest comments have been reverted.
I have eleven years of experience with the Wikimedia movement, much of it in some community leadership role; I have been advocate for a few fairly controversial initiatives. I generally consider myself to have thick skin. Even so, the experience was quite alienating. It did significantly influence my opinion of the Wikimedia community in the negative direction (even though most of the comments are anonymous, and anyway treating a handful of self-selected commenters as representative of a large community is totally unjustified, but that's just how human brains work), and it took me a while to recover from that. I'm fairly sure that if this would have been someone's first and defining large-scale interaction with editors, they would have quickly come to the conclusion that the community is a problem that needs to be fixed, not a partner to work with. I also wonder how encouraging the tone of that page was for someone who actually went there with the intent to report a bug or have a reasonable discussion. From what I have seen (from a distance), the feedback pages of other large-scale initiatives do not fare much better. --Tgr (WMF) (talk) 05:51, 10 August 2015 (UTC)Reply
But that's not development, that's other projects' responses to development. How would this help with that? -— Isarra 18:42, 10 August 2015 (UTC)yReply
At the very least it means don't come over here to development feedback pages and behave horribly towards others because to do so on your local project would likely get you into trouble there. As mentioned several times, there is no civility policy here which allows for enfettered nastiness and rage, often drowning out the important ideas and criticism related to the product itself rather than the ephemeral debates. This is extraordinarily unhealthy to the development process. I hope you can see how it helps in that context. Keegan (WMF) (talk) 18:27, 11 August 2015 (UTC)Reply
Completely agree with Tgr's comments, particularly using the Media Viewer feedback page as an example of how such an uncivil environment is detrimental to software development as well as the humans behind it. Keegan (WMF) (talk) 18:27, 11 August 2015 (UTC)Reply
"Sure, but it would be great if we could develop policies based on case studies of where things have failed before". I'm not going to call out specific people for specific incidents here in public, but if you want proof such incidents have occurred, feel free to email me privately. However, I'm sure there are incidents I don't know about as well. Isarra: meta:DICK is not binding; it's not even a guideline, just an essay. Mattflaschen-WMF (talk) 05:28, 11 August 2015 (UTC)Reply
Normal practice is indeed to discuss such things in public. This is a public project. Why change this now when it's most important? -— Isarra 01:29, 12 August 2015 (UTC)Reply

I guess, the biggest question I have along the why is it needed lines, is why (concretely) is the friendly space policy not enough, and what is the intended relationship between this policy and that one. Bawolff (talk) 21:48, 8 August 2015 (UTC)Reply

This. -— Isarra 02:28, 9 August 2015 (UTC)Reply
Not an author, just a random supporter of the ideas behind this policy, but; my feeling would be that the friendly spaces policy is very deliberately highly general, because it's designed to be cross-applicable to a lot of spaces, and is centred on "real-world" spaces. It does its job very well! BUt one of the costs of this is that it is not tremendously detailed; it does not set out (except in broad strokes) what is problematic, and it does not factor in tech-specific forms of microaggression or macroaggression. Another is that it does not provide any enforcement mechanisms for these kinds of online spaces. By having a code of conduct for technical spaces, both are addressed; we can be pretty specific in what we call out that isn't applicable to other spaces, and have an enforcement mechanism that is centred on technical spaces. Ironholds (talk) 17:11, 9 August 2015 (UTC)Reply
Bawolff, if by friendly spaces policy you are referring to wmf:Friendly space policy, I don't really understand your question. That policy is quite clearly about acceptable behavior at IRL events, focuses on face-to-face interaction, and its enforcement relies on the organizational structure that such events have (but most online spaces don't).
Are you saying you don't see the need for having a policy on behavior in online spaces? Or are you saying that the friendly space policy is or could be usable in online spaces? How? --Tgr (WMF) (talk) 04:48, 10 August 2015 (UTC)Reply
For my own part, you're right - I really don't see the need here. My experience with mw and wikimedia development as a whole has been such that these areas have overall come across as far more civil than most 'content' projects I've interacted with, and yet the latter tended to have policies out the wazoo. If anything these policies seemed to be giving some users more leeway for being giant asswads.
But maybe there is something here, and if so, it should be clearly documented so it can be properly addressed - because then maybe it would actually work, unlike the ones on other projects. -— Isarra 18:42, 10 August 2015 (UTC)Reply
I think the main problem with the content policies is that (they/some of them) have such policies, but don't actually enforce them (there are probably aspects that those policies are missing too, but enforcement is the biggest issue). This draft policy tries to be specific about enforcement. Mattflaschen-WMF (talk) 05:32, 11 August 2015 (UTC)Reply
Content projects tend to rely either on enforcement by consensus or enforcement by a large group of people none of whom are really responsible for it. Those methods work well for enforcing rules against unpopular violators; they don't work very well on popular ones. Verbal aggression from friends and vested contributors is regularly overlooked. --Tgr (WMF) (talk) 06:32, 11 August 2015 (UTC)Reply
In the case of technical spaces, the 'vested contributors' are often WMF staff. -— Isarra 01:29, 12 August 2015 (UTC)Reply
The main reason wmf:Friendly space policy (FSP) is not enough is that it not apply online at all. FSP is purely for in-person events ("Wikimedia Foundation events" to use its terminology). The relationship is that, for in-person technical events, you would have to comply with both of them. Mattflaschen-WMF (talk) 05:32, 11 August 2015 (UTC)Reply
Maybe this policy should focus on online issues, and FSP should focus on in-person issues. E.g. Stuff about unwanted photographs (An in-person issue) would be in the FSP, and this policy would be about things that commonly come up online. With the understanding that if you're at a tech event, both policies apply. Bawolff (talk) 09:40, 12 August 2015 (UTC)Reply


Change :Be catalytic – Tell what they should do instead. Occasionally this may be encouraging and motivating, better than a dense list of "rules". to: Be catalytic – Recommend what they could do instead. Occasionally this may be encouraging and motivating, better than a dense list of "rules" Jadeslair (talk) 07:15, 8 August 2015 (UTC)Reply

  Done Thanks for the suggestion. Mattflaschen-WMF (talk) 04:57, 11 August 2015 (UTC)Reply

Unnecessary section

The part about the Community Advocacy team and global bans should just be removed (or alternatively replaced with language like "the Global Ban Policy also applies to technical spaces"), to avoid any potential confusion that the code of conduct makes a different set of standards for global bans than what is in the policy. The language of the particular section is also pretty unfortunate and reads like "the CA team will be policing technical spaces and getting anyone who violates the above globally banned", IMO. wctaiwan (talk) 03:34, 10 August 2015 (UTC)Reply

The meta:WMF Global Ban Policy only includes extreme situations, and the WMF refuse to talk about those situations.
That will be appropriate for some 'Code of conduct' issues, however it seems like the 'reasons people will be banned' aspect of this CoC policy is intentionally more inclusive of incivility and bad behaviour than the Global Ban Policy, however the CoC is limited to parts of 'Wikimedia' that are 'paid work' spaces for people, esp. WMF staff.
I do worry that this CoC will mean 'the editing community' feels 'mediawiki.org' and other technical spaces has a more pro-WMF banning policy than meta or content wikis, and therefore withdraw from participating in these spaces. John Vandenberg (talk) 06:36, 10 August 2015 (UTC)Reply
I hope that it's the opposite, and people see that codes of conduct are not something to be afraid of, but just reinforcement of good practices. Mattflaschen-WMF (talk) 05:59, 11 August 2015 (UTC)Reply
I have no issue with that wording. I'll put it up and we'll see how it goes. Mattflaschen-WMF (talk) 05:59, 11 August 2015 (UTC)Reply

"including but not limited to"

As "technical spaces" could be almost anything, for example bug requests are discussed on email lists, main noticeboards on Wikipedia and Commons, etc. this is a potential bear trap. The list needs to be specific, and the communities it encompasses need to unambiguously agree with it and have the opportunity to discuss it and amend it. -- (talk) 05:02, 10 August 2015 (UTC)Reply

We could consider removing "including but not limited to" from the 'virtual' section. Mattflaschen-WMF (talk) 06:31, 11 August 2015 (UTC)Reply
On the presumption that unpaid volunteers are allowed to amend this draft in advance of whatever official WMF authorization you have in mind, I have been bold and made the change (diff). -- (talk) 14:50, 11 August 2015 (UTC)Reply
I've asked people to help edit and work on the draft repeatedly, so it's not really a presumption. While I'm okay with removing it from 'virtual' (we could add any missing spaces we notice), removing it from the 'examples of harassment' is much more problematic. It's easy for someone to claim their abusive behavior is technically not listed. Common sense as to what constitutes harassment is still sometimes necessary. This is a necessary list, but not a complete one. I also think it should be next to 'physical'; however, there are possible alternatives there. I also think "Wikimedia-controlled" is overly narrow (#wikimedia-dev is definitely a Wikimedia technical space, but is it Wikimedia-controlled despite being hosted by Freenode servers?). Mattflaschen-WMF (talk) 02:56, 12 August 2015 (UTC)Reply
By definition, #wikimedia-dev will become Wikimedia controlled if this policy is made part of its channel notice, as will any other channel that opts in per the IRC discussion below. In this way there is no need to include an open-ended wording in this policy that may be read as all channels where wikip/media is in the channel name (as any of them might discuss technical matters and have technical staff participating), as we can state it in a more specific way which avoids that reading. -- (talk) 07:40, 12 August 2015 (UTC)Reply

"Unwanted photography"

Clearly this needs clarification. In the U.S., for example, it is entirely legal to photograph any person in a public place at any time. My main work here is for the Commons. I don't normally ask in advance before taking photos in public, nor do most U.S. contributors to Commons. Are we, or are we not, saying that if that person is a participant in a WMF project that they may now have a veto on having their photo taken and posted here? If, for example, that person is giving a speech, do they now (unlike anyone else) have a right to say that someone cannot photograph them during that speech? What if they were one of the many people bicycling naked in the Fremont Solstice Parade and decided they didn't like a photo someone took of them? Would they then have special recourse unlike anyone else in the parade? I don't think that's what we mean to say, but perhaps it is. If so I strenuously object; if not, there needs to be some clarification as to what is meant by "unwanted photography." - Jmabel (talk) 14:27, 11 August 2015 (UTC)Reply

Certainly clarification is needed as the policy is open ended and non-specific at the moment.
Keep in mind that conferences are not necessarily public places and physical meetings might be in many countries, so there are several ways of interpreting personality rights and the nature of intrusion or reasonable expectations for privacy even when photographs are taken in a public place (I'm thinking of UK law here, rather than US law). This is where the policy would be much better split so that hard and enforced policy elements are clearly distinguished from best practice guidelines for participants. I doubt that anyone intends to create a policy which removes a citizen's right to free speech or the rights of journalists or "citizen journalists" to report events by threatening anyone that breaks it with punitive, publicly humiliating and permanent global bans. -- (talk) 15:00, 11 August 2015 (UTC)Reply
Jmabel -- actually, Commons has been generally receptive on several occasions to requests for deletion of such photographs (don't remember any from Fremont Solstice Parade, but do remember one from Burning Man) as a matter of "courtesy deletion" (though the photographed person doesn't have a strict right to demand deletion). AnonMoos (talk) 16:05, 11 August 2015 (UTC)Reply
The code of conduct says "unwanted photography", it doesn't say "photography without permission". It also doesn't say anything about publishing photographs (or a "special recourse")—thus personality rights are irrelevant to the discussion, as is Commons, as is the example about the Fremont Solstice Parade. As to your example of someone giving a speech, I think it is totally reasonable for someone to request that they not be photographed while giving a speech. For example, what if they are a Chinese citizen and speaking out against the Chinese government? In such a situation, repeatedly ignoring their request would be rightfully considered harassment, IMO. Kaldari (talk) 21:24, 11 August 2015 (UTC)Reply
Jmabel requested clarification. If someone asks for photographs to avoided for a talk, we respect their wishes as a courtesy; this already happens. However making threats of global bans will not stop "unwanted" photographs of a notable person being taken by journalists or anonymous tweeters if a conference is in a public place, or has public access. Neither should anyone in fear for their safety be given an impression that a code of conduct for technical spaces makes it acceptable to take more risk. Were I organizing a conference I would consider it unethical to have the event used in a way that creates serious unnecessary risk of harm, such as future imprisonment by their government, for any attendee. -- (talk) 01:09, 12 August 2015 (UTC)Reply
First, we need to distinguish between community policy and government law. There is no law in the U.S. against written personal attacks (unless they are libel), but the English Wikipedia still has a Wikipedia:No personal attacks policy. The primary goal of provisions like the unwanted photography one is to ensure people don't completely ignore the photography subject's intentions. If you just didn't know about their preference, I don't think that would be considered a violation. As far as WMF major public figures, I don't know any of them that have such a preference. If they did, and people felt that it interfered with e.g. documenting major speeches at WMF technical events, the right remedy would be to ask them to reconsider their preference.
I'm not sure how an unrelated public bicycle parade could possibly be considered a Wikimedia technical space. It's not based on who the person is; it's based on the space. So if you see Brion Vibber at some unrelated conference, you should follow their policy, not this one. And if you see him at the grocery store, it's between you and him (of course, being respectful is always good). Mattflaschen-WMF (talk) 02:08, 12 August 2015 (UTC)Reply
So the issue is "unwanted photography" in the technical space, not "unwanted photography" of someone you might have interacted with in such a space? Because that is certainly not clear here. Also, "unwanted" is still hopelessly vague, especially if the sanction is a global ban. If it means "taking photographs of someone who has clearly and explicitly said not to take them" it should say so. If it means a photographer can get in trouble because someone decides the expression on their face doesn't flatter them, that is a very different thing. And, no, that's not just a hypothetical: things like that are by far the main reason for requests for courtesy deletions. I'm all for courtesy deletions, but not for sanctions against a photographer for taking (nowhere here does it even say publishing) pictures. - Jmabel (talk) 07:56, 12 August 2015 (UTC)Reply
I think unwanted photography should be defined as taking photographs where the subject has either asked that their picture not be taken, or has a sticker saying no photos (Or I guess if the photographer is doing something creepy like specifically trying to sneak photos of someone). Bawolff (talk) 10:10, 12 August 2015 (UTC)Reply
I've rephrased this line. 'Clearly inappropriate' is intended to mean basically what Bawolff described as creepy. Mattflaschen-WMF (talk) 20:31, 12 August 2015 (UTC)Reply
The document says "follow this policy in Wikimedia's technical spaces", not "follow this policy anywhere you see a MediaWiki software engineer or bug-reporter". Mattflaschen-WMF (talk) 20:31, 12 August 2015 (UTC)Reply

Encouraging others to behave badly

I was wondering if it would be worth adding the following bullet point under the "Do not harass contributors or users" section:

  • Encouraging others to do any of the above.

Personally, I think that encouraging harassing behavior is also a form of harassment. Thoughts? Kaldari (talk) 23:45, 11 August 2015 (UTC)Reply

Is encouraging someone to do X the same as doing X? If so, I've won a bunch of professional sports games... Yaron Koren (talk) 00:14, 12 August 2015 (UTC)Reply
No, encouraging someone to do X is rarely the same as doing X. But occasionally it is. For example, encouraging someone to commit a crime is often committing a crime. Kaldari (talk) 00:32, 12 August 2015 (UTC)Reply
What is and is not regarded as a crime in various jurisdictions is not a meaningful metric, as completely ridiculous things find their way into legal codes all the time. As for the matter at hand, encouraging people to harass others is not inherently harassment. Don't call it what it isn't, just call it what it is - something that is also problematic, and should not be done. -— Isarra 01:19, 12 August 2015 (UTC)Reply
What's a specific example of this? --MZMcBride (talk) 02:30, 12 August 2015 (UTC)Reply
I would be fine with adding that. To address Isarra's point, rather than putting it in the "Examples of harassment" list, it could be a separate line item afterwards, saying "Do not encourage people to harass or personally attack others" or something like that. Mattflaschen-WMF (talk) 03:51, 12 August 2015 (UTC)Reply

Rewrite based on Contributor Covenant?

Fhocutt (WMF) suggested above to use Contributor Covenant's Code of Conduct as a starting point instead. I'm willing to do that. What do people think? Let's make this decision quickly (a week?), so we know what we're basing the draft on. Mattflaschen-WMF (talk) 04:30, 8 August 2015 (UTC)Reply

I do like it. --Brion Vibber (WMF) (talk) 23:35, 9 August 2015 (UTC)Reply
I just read <https://github.com/CoralineAda/contributor_covenant/blob/4905401/CODE_OF_CONDUCT.md> and I like it as well. It seems like a much better starting point. --MZMcBride (talk) 23:33, 10 August 2015 (UTC)Reply
I think this is a good idea. We will need to modify it to take our specific processes, spaces, community standards, and ways to report a violation into account--specificity in these areas is one of the best practices for codes of conduct that are effective in practice--but it's a solid base to start from. Fhocutt (WMF) (talk) 05:44, 11 August 2015 (UTC)Reply
I agree that there's also room for a separate set of guidelines on what respectful, productive, collaborative technical work looks like in practice, and would like to see momentum on this policy carry over to those guidelines as well. Fhocutt (WMF) (talk) 06:01, 11 August 2015 (UTC)Reply
I like it as well. It's more focused and specific. wctaiwan (talk) 04:46, 12 August 2015 (UTC)Reply
+1 here - the original proposal was a good start, but it seems like a better idea to build on something that has already gotten a lot of feedback from a wide variety of sources. —Luis (WMF) (talk) 21:32, 13 August 2015 (UTC)Reply
+1 for the same reason as Luis. Something that's widely used just carries more weight. I think in general people who misbehave don't stop to carefully read codes of conduct. The more likely potential violators are to have encountered other communities that effectively defend themselves with the same policy, the better. Milimetric (WMF) (talk) 00:58, 15 August 2015 (UTC)Reply
+1 ^ YuviPanda (WMF) (talk) 10:35, 17 August 2015 (UTC)Reply

What about this:

  1. Fhocutt (WMF) publishes a first version based on Contributor Covenant's Code of Conduct at Code of conduct for technical spaces with a {{draft}} template at the top and a disclaimer saying that this is a proposal under discussion.
  2. All the better if Talk:Code of conduct for technical spaces has Flow enabled. It allows for better subscription to specific topics and for marking topics as Done (cc DannyH (WMF)).
  3. We keep this page & talk page here for background with disclaimers on top pointing to the current draft and discussions. Once we are confident that anything valuable has been moved to the new pages, we archive these pages here.

--Qgil-WMF (talk) 07:54, 14 August 2015 (UTC)Reply

It could also be someone else--I care much more about getting something effective and workable together than who does it. Otherwise, fine by me. Fhocutt (WMF) (talk) 17:49, 14 August 2015 (UTC)Reply
If there is consensus to rebase on Contributor Covenant (as looks likely) I plan to do so after at least a week from my post above. I will start with a clean copy of Contributor Covenant. Various people have ideas on how to update the draft from there, so I expect quick changes after we have the new baseline. I do not see any reason to start a new page at Code of conduct for technical spaces. We can just update the draft here, which is a more appropriate title since it's still a draft. If someone wants to request Flow, I suggest starting a new section about that. Mattflaschen-WMF (talk) 20:58, 14 August 2015 (UTC)Reply
I've gone ahead and uploaded a clean copy. We can start refining it now. We should pull things we're agreed on from the old version, and some people have ideas for text to add to this base. Mattflaschen-WMF (talk) 00:00, 18 August 2015 (UTC)Reply


IRC is not operated by Wikimedia and has independent and long established harassment policies. If the IRC harassment policy is to be effectively superseded by this Code, then it needs to be presented to the IRC community and the channels that it applies to need to be spelled out. As it stands it is not clear if this applies to all IRC private messaging, if it only applies to an agreed set of "technical" IRC channels (which should then have this Code linked in the channel notices), or whether it applies to all self-identified Wikimedians whenever they are using an IRC channel. The text currently reads in a way that denies the possibility of anonymity on IRC, which appears to run counter to Freenode's IRC policies. -- (talk) 05:02, 10 August 2015 (UTC)Reply

The fact Freenode has a policy on harassment doesn't preclude use from having more specific guidelines. Harassment of women on IRC is well documented, so it's clear existing guidelines don't achieve the intended effect. I agree it would be good for the policy to be announced in the topic -- not for would-be harassers, but for people who are at risk for being harassed. They need to know we will not accept harassment. Valhallasw (talk) 07:48, 10 August 2015 (UTC)Reply
Threatening WMF global bans for those accused of being in some vaguely defined way "non-positive" or disruptive is not an improvement on Freenode's harassment policy nor on existing community agreed policies for conduct enforced by trusted unpaid volunteers. This Code unnecessarily relies on a process with no possibility of appeal, nor any requirement for evidence to be examined by the accused or their elected peers. It puts an unaccountable faceless harassment police force of WMF employees at the centre of our community, rather than the WMF published strategy of putting the volunteer at the centre.
Just to be absolutely clear, the concept of a code of conduct I have no objection to whatsoever. Nor do I object to more being done about harassment (and not just harassment of women). However this code attempts to define "technical spaces" as almost everything where I might write about technical issues, then jumps into a solution of global bans which are completely unnecessary when there are plenty of volunteer driven systems available. Thanks -- (talk) 08:15, 10 August 2015 (UTC)Reply
The point about the global ban policy has been rephrased. It's hard to see how it could apply to non-Wikimedia IRC channels ("self-identified Wikimedians whenever they are using an IRC channel") when it explicitly limits itself to "Wikimedia's technical spaces"; I don't know how e.g. the emacs channel could be considered "Wikimedia's technical spaces". Mattflaschen-WMF (talk) 06:42, 11 August 2015 (UTC)Reply
No IRC channel is actually "Wikimedia's" as they are not run by Wikimedia, yet this policy asserts its authority on those spaces; if by "Wikimedia's" we means any channel with "wikimedia" in it, then that applies to #wikipedia-en and #wikimedia-commons whenever anyone wishes to assert that there was a technical discussion in those places.
Apart from presuming that "technical spaces" means any technical discussion anywhere, there is no definition so it is not limited. Unless the is accurate qualification of the scope of the Code, it seems fair to presume that users will apply it to all of the technical spaces mentioned where they may want to escalate a dispute, and it will therefore have the authority to bypass all local policies (as the ultimate sanction of a global ban is how it will be enforced). -- (talk) 06:48, 11 August 2015 (UTC)Reply
For the IRC channel issue, what about the idea of having each channel's topic say whether this code of conduct applied there? Mattflaschen-WMF (talk) 01:48, 12 August 2015 (UTC)Reply
I am an op on a couple of IRC channels and have set notices for a few. I agree that channels should be encouraged to opt-in to conduct policies beyond Freenode's standard. It is tricky to assess whether the specific community that are IRC users have a consensus for policies like this, rather than a few individuals who happen to be participating that week. Some channels benefit from highly specific behavioural guidelines, for example #wikimedia-lgbt has had a friendly "PG-13" language warning in its notice, which was added not long after it started, to encourage participants to stay friendly for potential minors. I guess discussion with channel participants a couple of times would probably be the way to go, or if there was an appetite to discuss it for a particular Wikimedia channel, a vote could run on a related wiki. However I suggest that in many cases we can encourage the more widely applied safe space guidelines as sufficient, and these have proved pretty uncontentious since their original creation (I think the version we put together at the UK chapter after some issues with real life events was the starting point back in 2012). -- (talk) 02:07, 12 August 2015 (UTC)Reply
Which safe space guidelines are you referring to? Mattflaschen-WMF (talk) 04:01, 12 August 2015 (UTC)Reply
Primarily https://wikimediafoundation.org/wiki/Friendly_space_policy though specific ones exist for some projects, such as meta:Friendly_space_policies and https://wikimedia.org.uk/wiki/Participation_policy which already apply to related spaces. -- (talk) 07:47, 12 August 2015 (UTC)Reply
The Friendly space policy only applies to in-person events, not online. Mattflaschen-WMF (talk) 20:04, 12 August 2015 (UTC)Reply
Yes, though some do specifically apply to virtual spaces, the distinction being slightly arbitrary as wiki-events invariably have significant virtual components and virtual participants for all related activities before, during and after physical events. Strangely enough, the harassment cases that spring to mind in relation to complaints in relation to friendly space policies, have involved virtual world actions such as deriding photographs online from events and outing participants online post-event. -- (talk) 05:19, 13 August 2015 (UTC)Reply
Hi, This closes a major loophole in which Wikimedia IRC channels, notably #wikimedia-commons were used in most inappropriate way. Fae is very well aware of that, if not part if the issue, so it is not surprising that he opposes this policy. Regards, Yann (talk) 21:37, 11 August 2015 (UTC)Reply
Please don't smear me with false allegations to make a point. I flatly and absolutely deny your snide claims about what I know, and heavy hints to make others believe I have some sort of evil motivation to scare me off asking basic questions about this proposed governance related policy. Please focus on the issues, rather than taking discussions off on a tangent through maligning participants. If you have evidence of harassment to present, please, please do present it in the correct forum where I would be happy to be held to account and answer any question based on evidence, rather than poisoning this discussion in an unhelpful and unpleasantly disruptive way. -- (talk) 01:30, 12 August 2015 (UTC)Reply
What I find very interesting in this discussion is the notion that what pertains to the wiki, doesn't apply to IRC. This would further infer that on wiki status would not carry over to IRC, yet, virtually all of the IRC Ops are admins on Wiki, they frequently threaten people with bans from various IRC channels for either on Wiki justifications or for not doing what they, as admins/ops, say. Its also not uncommon for links to on Wiki essays and guidelines to be used as justifications for blocks or bans on IRC. So what this appears to me to be is a classic case of picking and choosing when to apply a policy or rule and when not too. For example, if an editor does something an admin doesn't like, the admin on the IRC channel can block the other person without cause or justification and the WMF won't get involved. This being the case, the WMF really just doesn't appear to care about blatant IRC disruptions caused by Ops/Admins like Yann who frequently block or ban editors from the channel for what they want. This is unfortunate and disappointing to me but not entirely surprising given the WMF's history of lack of control over admins and the abusive environment they enable to turning their backs on editors. Reguyla (talk) 01:50, 12 August 2015 (UTC)Reply
I don't know the history of all the IRC issues being discussed here, so don't misunderstand me as taking sides. But I do completely agree that the Wikimedia IRC technical community is intertwined with the rest of the Wikimedia technical community (including MediaWiki.org). Mattflaschen-WMF (talk) 03:58, 12 August 2015 (UTC)Reply

I believe that #wikimedia-commons as an irc channel is definitely out of scope of this. This is not the place to discuss etiquette guidelines for the entirety of the Wikimedia movement. If people want to do that, they should try their luck at meta. Bawolff (talk)

Without going into the specifics over whether WMF has reach over IRC channels, let me say that many channel operators, especially in the channels concerned, are not willing to ban someone unless they see clear evidence of actual wrongdoing in the channel. So even if WMF can do this, on a day-to-day basis, the existing operators may not be willing to enforce it.--Jasper Deng (talk) 18:44, 13 August 2015 (UTC)Reply

We can agree that inappropriate behavior is equally inappropriate on-wiki, on-Phab, on-IRC, etc. Freenode's policies define the basic rules of Freenode. Our CoC cannot enter in conflict with such policies, but can build on top of them. https://meta.wikimedia.org/wiki/IRC/Channels#MediaWiki_and_technical should define which IRC channels are used as collaboration channels by the Wikimedia tech community (if the list is out of date, it can be updated). I don't think that an IRC channel related to the Wikimedia tech community can opt-out of a CoC or any other policy if it has been approved for all the tech community, just like I don't think a specific Phabricator project or a specific mailing list can opt-out either. In the "Enforcement" section I'm proposing that the first level of handling of complaints should be the own admins/maintainers of the spaces) affected by the complaint. If you agree with this idea, this would mean that a complaint about a situation in #wikimedia-foobartech would be handled first by the own channel operators. If the ops don't feel like solving the complaint alone or if the person claiming doesn't think that these ops will solve their problem, then a second level would take the issue (and that second level may or may not be formed by WMF employees, that is to be decided).--Qgil-WMF (talk) 08:13, 14 August 2015 (UTC)Reply


I understand that this page is currently a draft. I'm curious when the enforcement-related question marks will be filled in. Specifically:

  • Alternatively, contact ? directly about the issue.
  • In the case of persistent disregard of these guidelines or if you are uncomfortable contacting the person directly, contact ? on Wikimedia IRC or send an email to ? and ask them to look into it.

If this proposed code of conduct is implemented, who will be responsible for handling complaints and how will this person or group do so across IRC, Phabricator, Gerrit, mailing lists, the wikis, etc.?

Assuming action is taken by this person or group, such as instituting a temporary or permanent ban of an individual, what will the appeal process be? --MZMcBride (talk) 23:26, 10 August 2015 (UTC)Reply

I very much support moving forward and making progress towards putting this Code of Conduct for Technical Spaces into place. Thank you very much for spending the time and energy to create it and the time and energy to discuss, change and defend it - however I have the same questions as MZ has written out above (specifically about enforcement). It would also be good to know the difference between actions that would result in a temporary ban vrs ones that would result in a permanent ban and why. It is very clear at in-person events when people have willingly opted into the Friendly Space Policy (they must agree to the policy to register), decisions about enforcement are up to the event organizers and (more or less) fully left to their discretion. We have removed people breaking these policies from events in the past without issue. In this case if we don't have clear and **agreed upon** steps to take in order to enforce this we will run into issues. In case people have not seen what has been going on here, you should check it out: https://meta.wikimedia.org/wiki/Grants:Friendly_space_expectations It seems to be working well and very reasonable, but I think they had a slightly less complex problem to solve. Rfarrand (WMF) (talk) 22:58, 11 August 2015 (UTC)Reply
This is an open issue. As far as "It would also be good to know the difference between actions that would result in a temporary ban vrs ones that would result in a permanent ban and why.", I think there is a role for discretion in this policy as well. It's hard to set a precise per-determined penalty for every grade of violation of every type of forbidden conduct listed. Mattflaschen-WMF (talk) 03:42, 12 August 2015 (UTC)Reply
This seems to be the central and hardest question. Who exactly [has to/gets to] decide (A) what is "over the line", (B) what the outcome is in each instance?
The Ada CoC just says "Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by opening an issue or contacting one or more of the project maintainers." - Who is that, in our case? Who gets that burden/power/responsibility/role/unpleasant-chore? Who do we turn to, trust, and expect effort and investigation from?
Speculation: They can't be staff-only, because that would raise concerns about COI issues when the reporter/reported is a staff-member (and by god do some of the staff get verbally abused a lot, by anonymous and non-). They should be privacy-minded/discretion-trained/perhaps-even-NDA'd people, because they might be sent private evidence. They should be empathic and understanding of cultural and linguistic diversity. They should include experienced mediators/communicators, because they'll need to deliver warnings and decide on consequences, in a way that doesn't make the instance worse, nor fragment the communities. Constituting a proper group to review and respond is the critical factor in adopting any such plan that will work. –Quiddity (talk) 17:58, 12 August 2015 (UTC)Reply
I think the points of contact and enforcement can be organized at two levels. The first level is formed by the administrators or maintainers of the space where the problem happens. All our technical spaces have individuals with permissions to ban a user, and that responsibility is acquired through experience and merit. This might not be as simple in practice, but as a principle I think it works. The second level should be used as escalation point if the first level is no good fit for addressing the problem for whatever reason. The simplest option would be to pick a related WMF team (Community Advocacy or Engineering Community being logical candidates). If this simplest option is seen as problematic for being in practice restricted to WMF employees (I personally don't think this is a problem, but such concern has been raised by some people) then a small committee could be formed with 3-5 people chosen among the admins and maintainers active in the first level.
The point here is not who exactly can be point of contact and enforcement, because we have enough qualified people with enough community trust to be in this role. The point here is to assure that there is people officially appointed to this role with community backing. This way a person willing to report knows where to go with confidence and trust, and a person being eventually warned or banned knows that the person warning / banning is doing that in a community role, not on a personal motivation.--Qgil-WMF (talk) 12:38, 13 August 2015 (UTC)Reply
At the beginning of this thread it was asked "what will the appeal process be?" I cannot see this being addressed yet. Until there is a credible agreed appeal process to assure the community that our shared values of transparency and accountability are being addressed, this will remain controversial, regardless of whatever procedure is devised (presumably this process remains the choice of WMF executives, unless there is a shift to putting this in the hands of elected volunteers?) for appointing community prefects/collegiality monitors or whatever they are going to be called who are granted special banning powers. -- (talk) 14:59, 13 August 2015 (UTC)Reply
@: , I have created a section to discuss the appeal process specifically. There is nothing in this CoC and its application that "remains the choice of WMF executives". In my proposal just above I'm not considering anyone "who are granted special banning powers". The Wikimedia tech community already has users that have these permissions ("administrators and maintainers" for simplicity). Even if we decide that we have a second level of enforcement (say a WMF team, an elected committee, etc) this should not grant to its members any "special banning powers". When such group would reach to a conclusion, they would ask the admins/maintainers of the affected space(s) to execute the ban. Note that this distribution of power would be an intrinsic appeal process in itself, because it will be effective only when there is enough consensus between the acting parts. All we agree that the last thing this CoC and Wikimedia tech community needs is a single person or a single entity with superpowers able to take executive shortcuts. In the way this proposal started and develops, I'm not seeing any risk of falling in this situation. If you have concerns, let's address them in the writing and implementation of this CoC.--Qgil-WMF (talk) 07:41, 14 August 2015 (UTC)Reply
Thanks for setting up the section. With regard to your statement "nothing in this CoC and its application that "remains the choice of WMF executives"", I did not make this up, in fact the statement "Following approval from executive or legal staff, individuals who meet the criteria may be banned from all Wikimedia spaces, including technical ones" was added by Mdennis (WMF) on 7 Aug and eventually edited out by Mattflaschen-WMF on 11 Aug. Presumably this statement about global bans is entirely accurate, and while global bans remain an option then "executive or legal staff" will be part of how this policy will be implemented and interpreted. -- (talk) 10:01, 14 August 2015 (UTC)Reply
OK, understood. I take the mention to the Global Ban Policy only as a reminder that it exists. This CoC and its handling should be compatible with that policy but independent from it. If someone in the Wikimedia tech scope is getting more points to receive a global ban, it would not be based on infractions against this CoC but on the criteria of the Global Ban Policy itself, and that would be handled according to the process of that policy.--Qgil-WMF (talk) 11:55, 14 August 2015 (UTC)Reply
The interrelationship needs to be spelt out in the escalation process. As you infer, it would be best to avoid giving the impression that this policy is part of policies or legal documents controlling office locks and global bans, so the changes to this draft in the last week to move away from that hard line stance are a distinct improvement. This is the first time I have read "points" being part of the system. It would be a good step towards transparency if any system like this were made public and volunteers could ask about their own status if they are in doubt, and ask for review or start an appeal, rather than this being like expecting repercussions for asking about black ops. -- (talk) 13:41, 14 August 2015 (UTC)Reply
"Getting more points" was just a casual expression written by this non-native English speaker. I think the process we need should be as transparent as possible taking into account privacy considerations (I can see how many times a three party private conversation can be more effective solving a problem than a fully public and transparent claim/appeal process). In any case, I don't think we need anything like w:Point system (driving) here.  :) I'll pay more attention to my casual phrasing in this discussion. And let me insist, we are designing a Code of Conduct to improve community health that must go through community approval, not a toolkit to recreate a police state.--Qgil-WMF (talk) 13:52, 14 August 2015 (UTC)Reply

Bug management/Phabricator etiquette

If this code of conduct is adopted, what will happen to Bug management/Phabricator etiquette? --MZMcBride (talk) 23:55, 12 August 2015 (UTC)Reply

Parts of the CoC might cover and hence supersede parts of the Phabricator etiquette, while other parts of the Phabricator etiquette are so Phabricator-only specific (e.g. status and priority fields) that they could remain and simply extend the CoC. --AKlapper (WMF) (talk) 10:35, 13 August 2015 (UTC)Reply

"Sexual imagery or verbiage in public technical spaces"

I believe this example has been expressed oddly; possibly "verbiage" is used differently in modern British English compared to American English. Was the intent to say that sexual imagery and sexual texts are harassment, or was it actually intended that "verbiage" of any kind can be interpreted harassment? This reads as if writing a long complaint could now be called harassment just for being long-winded.

The adjective is meant to apply to both nouns. In other words, it's equivalent to "sexual imagery or sexual verbiage". If someone else finds this confusing, we could change it to that. Mattflaschen-WMF (talk) 23:47, 13 August 2015 (UTC)Reply

As a secondary issue, the example concerns me as "sexual imagery" is vague wording, and in practice has been clearly interpreted on the English Wikipedia to include "sexuality", hence material like LGBT event photographs and historic material which might include partial nudity, such as 19th century ethnographic photographs. I understand what is probably intended, something like nudity or sexual photographs which are likely to cause offence when presented inappropriately. Could someone rephrase this so that the example is unlikely to be a cause of arbitrary future contention if we are discussing technical issues with topics like Wiki Loves Pride events or linking to categories for media in a medical education/history project and could include photographs or diagrams with partial nudity, such as the images from the Wellcome Library or diagrams from Cancer Research? Thanks -- (talk) 11:16, 13 August 2015 (UTC)Reply

This is intended to be a code of conduct for technical spaces; it doesn't affect how Wikipedia chooses to cover e.g. 19th century ethnography. I think people will understand that if you are in the process of working on a technical problem, and incidentally have to link to an image or category with nudity, that would not be a violation (but it might be worth adding a NSFW on the bug report or whatever). Mattflaschen-WMF (talk) 23:47, 13 August 2015 (UTC)Reply

Criticize ideas, not people

I am not a fan of that phrasing. First, there is not all that much difference between "your idea is stupid" and "you are stupid". There is an unfortunate misconception in certain circles of the community that the second is a personal attack but the first is fine, and we should not reinforce that. Second, publicly challenging improper behavior can be an important part of upkeeping community norms and socializing newcomers, and the "consequences" part explicitly suggests it as a first step of enforcement, but it would be itself a violation under this rule. --Tgr (WMF) (talk) 21:49, 16 August 2015 (UTC)Reply

I think we can drop this now that we've rebased onto Contributor Covenant. They instead use different language with the same purpose, mainly "Examples of unacceptable behavior by participants include [...] derogatory comments or personal attacks [...]". Mattflaschen-WMF (talk) 00:08, 18 August 2015 (UTC)Reply

Safe spaces, incivility, and passive aggression

It seems that there's general agreement that a lot of outright "in your face" harassment on gender/sexuality/racial/religious/etc grounds is to be avoided -- which makes me happy to hear -- but it doesn't surprise me that there are more negative reactions over the idea of disruption being a problem than .... any other reaction at all.

This doesn't surprise me because passive-agressive dismissive incivility is absolutely rampant in our community [our community == everyone involved in Wiki*edia, yes if you are reading this it includes you], as well as many other FOSS and FOSS-like projects. I know I'm a part of that civility problem too, and try to avoid it but don't always succeed.

I really like the idea of a positive code of conduct that prescribes respect, civility, and a mindfulness that we're working on tools for people to use... Being actually productive is important, and with our communications channels it's easy to get sidetracked into bikeshedding, "well-actually"ing, "RTFM"ing, and accusations of who's the most disruptive.

The most radical act isn't pointing out others' failings, it's admitting that you're part of the problem, and that changing is required. I know I need to change for the better, and I hope you all will join on that journey. --Brion Vibber (WMF) (talk) 23:48, 9 August 2015 (UTC)Reply

I agree. I was a fan of the safe space policy going into effect for developer events, and that has since been extended first to Wikimania, then other Wikimedia events, and now recipients of WMF grants. With the board, Lila, CA, and others stating that looking into ways to address community collegiality is a priority for this year - this seems like a logical effort for the developer community. I agree it is not perfect, but I have yet to see an online community policy that was. ;) I endorse the idea of proceeding with refinements based on this page's discussion - but not delaying things for months and months in an attempt to achieve a perfection that will likely never come. However, Brion is correct that we must all also own up to our own role in this problem - and yes - there is a problem. --Varnent (talk)(COI) 02:46, 11 August 2015 (UTC)Reply
I agree. I can admit that sometimes I have gotten more heated than I should have. None of us are perfect. Hopefully, this serves as a reminder for us to stay respectful. Mattflaschen-WMF (talk) 05:46, 11 August 2015 (UTC)Reply

A tangent on wording; almost anything written online can be interpreted as passive aggressive. This is especially true in an international and multilingual context. For example written British English (at least for anyone educated before the 1980s) leans to self-effacing to be polite, this careful politeness in an online modern or Americanocentric context is easily interpreted as slyness, sarcasm, or passive aggressive. A general rule of thumb should be to double-check, as sometimes simple good faith and echoing back your reading in plain English, will quickly diffuse what might appear to someone from a different cultural background to be passive aggressive or annoyingly sarcastic. -- (talk) 02:14, 17 August 2015 (UTC)Reply

Need clarification on the scope

I'm watching this discussion since someone posted a link in Commons' village pump. At first, I didn't see anything concerning to my activities there (my main project) and wonder how it affects a Commoner. Now I saw more Commoners raising concerns here. So I re-read the draft. It has two parts of scope; physical and virtual. "Physical" is noway relevant to Commons or any other WM projects as we are not physically meeting any. In "virtual", I see only "MediaWiki.org, Phabricator, Gerrit, technical mailing lists, technical IRC channels, and Etherpad". No other WM project like Commons is mentioned. My understanding is also a WM project like MediaWiki has no control over such projects. General policies like Global ban must be defined in Meta. So I wonder from where these concerns are coming?

I can understand MW has the right to block somebody if s/he misbehaved to a volunteer here on/off-mediawiki. But off wiki activities need investigations by some higher authorities like Arbcom. Commons doesn't such a setup and many people make benefit from it. But it is a "common" problem and not relevant here.

I also see "global ban policy" is mentioned in the draft. It only means people who apply global ban can consider the reasons which result a block here too, while considering a GB against a user. It is quite logical and I see nothing particular in it.

Am I correct? Let me correct; if I'm wrong. Jee 02:46, 12 August 2015 (UTC)Reply

Yes, I think this is basically right. Some of the prior questions and comments were when it said, "virtual (including but not limited to MediaWiki.org, Phabricator, Gerrit, technical mailing lists, technical IRC channels, and Etherpad)". Now it says, "virtual (MediaWiki.org, Phabricator, Gerrit, technical mailing lists, technical IRC channels, and Etherpad)" (no "but not limited to", so no ambiguity about which web sites (or parts of them) it applies to). If there are problems in virtual technical spaces that are not listed here, the policy could be amended after further discussion. Mattflaschen-WMF (talk) 04:06, 12 August 2015 (UTC)Reply
Thanks Mattflaschen for the clarification. Tgr's suggestion below is also very good. It is other project's responsibility to monitor their lists and IRC channels. Jee 04:48, 12 August 2015 (UTC)Reply

Would it be fair to say that technical mailing lists are the ones listed under m:Mailing_lists/Overview#MediaWiki_and_technical, and technical IRC channels are the ones listed under m:IRC/Channels#MediaWiki_and_technical? Those lists seem fairly comprehensive (and if not, they should probably be fixed anyway). --Tgr (WMF) (talk) 04:27, 12 August 2015 (UTC)Reply

I'm pretty sure a lot is missing from both those lists. Bawolff (talk) 08:05, 12 August 2015 (UTC)Reply
True, but it is easier and better to update those pages than let people guess. The rebased text says "This is a code of conduct for Wikimedia technical spaces." I think we should spell out those spaces, right there after a colon or somewhere below. The list: mediawiki.org, wikitech.wikimedia.org, wikimedia.phabricator.org, gerrit.wikimedia.org, Wikimedia technical mailing lists, and Wikimedia technical IRC channels. Is the intention to keep this CoC for online spaces only, since events are already covered by the Friendly Space policy? If so, that should be explained as well, for context (now it's in the See also, I think it should be integrated to the text).--Qgil-WMF (talk) 11:26, 18 August 2015 (UTC)Reply
No, it's explicitly for physical spaces as well ("Project spaces can be both physical, such as hackathons and other technical gatherings"), and has been since the beginning of the draft (though the wording has evolved). As far as IRC, I'm not sure whether it's better to use the topic (which can be controlled by the channel) or a Meta list that anyone can edit. Mattflaschen-WMF (talk) 20:41, 19 August 2015 (UTC)Reply

Other examples from industry

There's a nice code of conduct from the Electron project here that has a prominent link in the project's readme. Niedzielski (talk) 04:24, 9 August 2015 (UTC)Reply

FYI, The TODO code of conduct actively puts marginalized people in harm’s way. (...) This Code of Conduct is essentially legalese to protect the corporations and their cis straight white men contributors from accusations of sexual discrimination, racial discrimination, and with the recent EEOC ruling, homophobic and transphobic discrimination. - https://modelviewculture.com/news/todo-group-and-open-source-codes-of-conduct Valhallasw (talk) 07:35, 10 August 2015 (UTC)Reply
That blog post falls very far from the standards of constructive and respectful communication it supposedly promotes, and is full of falsehoods (for example, check the pull request it links to and compare the actual discussion with the way it's characterized in the post). I don't think it should be taken seriously. --Tgr (WMF) (talk) 07:47, 11 August 2015 (UTC)Reply
Another resource which helps explain why codes of conduct can help, is https://adainitiative.org/2014/02/howto-design-a-code-of-conduct-for-your-community/ . Mattflaschen-WMF (talk) 01:40, 20 August 2015 (UTC)Reply

Role of the document

I feel like this code of conduct is trying to be a set of guidelines for best social practices (no off-topic comments, discuss things in the open) and a set of binding rules (no harassment) simultaneously, with no clear delineation. This should be addressed either by structuring things differently and making it clear which things would lead to consequences, or just splitting it into a binding policy and a non-binding set of guidelines. The binding parts would also need to be concrete (in defining unacceptable behaviour) to be enforceable.

In what's been defined so far, I think the part about rejecting patches should just be removed—given the precedent of some developers being very determined to push through features against consensus, people really shouldn't be punished for rejecting patches, provided they do so without personal attacks and such. (Also, please don't use words like "catalytic".) wctaiwan (talk) 04:03, 10 August 2015 (UTC)Reply

Agreed. --MZMcBride (talk) 23:16, 10 August 2015 (UTC)Reply
All of it is meant to be binding with the possible exception of the first sentence, which is kind of a preamble laying the groundwork. I feel like "Communicate about technology in public where possible" is relevant to the document, even though it's kind of non-actionable due to "where possible", because this is an important part of the MediaWiki ethos. "Comments should be made with the intention of constructively contributing to our technologies." is actionable; it means e.g. comments that are purely unconstructive attacks are not appropriate. I would expect this would not be enforced for one-off insignificant off-topic comments, but only to deal with major disruption. It doesn't say patches can't be rejected; a lot of patches need to be rejected, for valid reasons (product, technical, or both). It just says "consistent and unwarranted rejection of patches" (emphasis added) is not appropriate. Mattflaschen-WMF (talk) 06:28, 11 August 2015 (UTC)Reply
I certainly don't think it should be a policy that communication happens in the open (the foundation may wish to mandate that for WMF developers, but this shouldn't be binding on volunteer developers). Unconstructive attacks may be actionable, but as written it's too vague and too broad (e.g. foundation PMs may argue that debating whether a feature is wanted at all is "unconstructive"). "Unwarranted" also leaves a lot of room for interpretation.
I'm not fundamentally opposed to covering both best practices and actionable misconduct on the same page (e.g. first list the goals of a CoC and document best practices, and then list what is considered actionable), but I think the latter needs to be concrete and separate for it to be enforceable at all. wctaiwan (talk) 07:28, 11 August 2015 (UTC)Reply
Not to mention the fact that banning someone from one or more MediaWiki forums (the only real enforcement available for these rules) would presumably make them more likely to communicate with others privately. Yaron Koren (talk) 17:47, 11 August 2015 (UTC)Reply
This is not the only enforcement mechanism given. It explicitly makes clear that in some cases, just talking to them is the best way to go, and in other cases, issuing a formal warning may help. Even in the case of bans, there are different types given. Mattflaschen-WMF (talk) 02:03, 12 August 2015 (UTC)Reply
I wouldn't consider a warning to be enforcement, but that might just be a semantic issue. Yaron Koren (talk) 02:33, 12 August 2015 (UTC)Reply
It doesn't mandate solely public communication for anyone, WMF or otherwise. It just expresses that you should default to make it public (then consider the alternative, where appropriate), rather than go the other way. Mattflaschen-WMF (talk) 02:03, 12 August 2015 (UTC)Reply
Anyway, as I said above, while I think the "Communicate about technology in public where possible" line is relevant, it's not really actionable due to 'where possible'. I don't have my heart set on keeping it. Mattflaschen-WMF (talk) 03:19, 12 August 2015 (UTC)Reply

(outdent) My concern is less about the specifics and more about the enforceability of a code of conduct that is deliberately designed not to be specific. You mention the role of discretion below. I'm fine with, say, discretion as to whether to give someone another chance, or whether something constitutes a pattern of disruptive behaviour, but as it is there isn't enough clarity when it comes to the line between best practices and actual rules. wctaiwan (talk) 04:37, 12 August 2015 (UTC)Reply

@Krenair: @Yaron Koren: Krenair put back "We also pledge to communicate about MediaWiki/Wikimedia technology in public where possible" and made it even stronger. I don't have a strong preference on the first part (though I respect the fact that it was removed before, and understand why), but I think "All written non-security-related discussion about changes to MediaWiki core and extensions and skins which are deployed to Wikimedia sites or included in the MediaWiki tarball releases must be public." is way too strong and I've removed it. Mattflaschen-WMF (talk) 01:23, 18 August 2015 (UTC)Reply
I have no idea what a "pledge" means in the context of a code of conduct. Actually, I have no idea what "we pledge" means here at all - most people aren't pledging anything, so it just sounds like false information. Yaron Koren (talk) 02:19, 18 August 2015 (UTC)Reply
As much as I agree with the importance of public communication, I'd rather keep the CoC within the core goal of "making participation in Wikimedia technical projects a respectful and harassment-free experience for everyone". Someone not communicating publicly will not automatically incur in disrespect or harassment. I would leave it out.--Qgil-WMF (talk) 11:15, 18 August 2015 (UTC)Reply
I agree with Quim. We should keep this document tightly focused on identifying prohibited forms of harassment, rather than general good practices. I added a part about the Wikimedia mission because I think it strengthens the case that prohibiting harassment is core to our values, but on the other hand I think communicating in public is a different principle and should be discussed elsewhere.—Neil P. Quinn (talk) 22:23, 18 August 2015 (UTC)Reply
+1. I think that a separate document on positive ways that we communicate respect for each other and for the movement (somewhat like the positive portions of the Phabricator etiquette) is a very good idea and that this would fit well in there, but I agree that it would be better to focus this on our core goal. --Fhocutt (WMF) (talk) 00:55, 19 August 2015 (UTC)Reply
Okay, in that case I'll go ahead and remove it.—Neil P. Quinn (talk) 22:12, 19 August 2015 (UTC)Reply

Enforcement in new draft

The enforcement and reporting are now way more detailed, which is both a good and a bad thing. It's good because there's less ambiguity, but that much text may distract from the rest of the document. Also, since it comes from multiple sources, there are consistency and redundancy issues to work out. We may want to separate it out to a subpage (both as a break-out page if the overall Code of Conduct is approved, and for discussion now). Mattflaschen-WMF (talk) 01:28, 18 August 2015 (UTC)Reply

(((Comments have been moved to "Right to appeal"))).--Qgil-WMF (talk) 09:13, 20 August 2015 (UTC)Reply
The current section Enforcement contains two lists of "possible responses: that, well, look too similar. Wouldn't it be better to consolidate a single list, and then have a new section "Appeal" with the details of the appeal process?--Qgil-WMF (talk) 09:49, 20 August 2015 (UTC)Reply
@Qgil-WMF: Which two lists are you referring to? The "For a simple case" is for actions a single member can take on their own. The "Possible responses to a reported breach of the Code of Conduct may include" is a list of things the full committee can do. I've tried to clarify this. Mattflaschen-WMF (talk) 02:41, 21 August 2015 (UTC)Reply

First line of enforcement

I would like to float an alternative proposal for first line enforcement. I suggest that there should be a small group (probably either 3 or 5) of people from our technical community (both volunteers and staff, based on their involvement in the tech community). There would be an OTRS or mailing list address pointing to that group. I think the steps would be something like that:

  1. Someone emails e.g. techcoc@wikimedia.org , or contacts one of the members on IRC. If it's OTRS, I think the email could auto-create a ticket, and maybe for IRC it could be created manually from the web interface?
  2. For a simple case (e.g. first offense, minor harassment in a single space), a single member could:
    • Defer to the space itself (e.g. MediaWiki.org admins in the case of an incident on MediaWiki.org, or event organizers for a hackathon), if they seem best-equipped to handle it.
    • Issue a public or private warning directly.
    • Decide to take no action (logging why and notifying the reporter)
  3. For a more complicated case (e.g. repeat offense, when a ban looks like like it will be required, or where there is cross-space harassment), the initial responder could ask other members of the group for feedback, and can also bring in people from the spaces involved (e.g. if it happens on both MediaWiki.org and IRC, bring in involved admins and IRC contacts).
  4. In case of even more complicated or urgent matters, or if the group is unable to reach a decision (e.g. the minority in the group objects to the decision), it can be escalated.
  5. After the initial outcome (which would always be logged, even if it's just 'deferred to XYZ' or 'no action taken'), the reporter should be notified.
  6. The reporter always will be notified of the initial outcome, and has the right to appeal (see separate discussion).
  7. If public action is taken on the basis of the CoC (whether by the group or by local maintainers of the space), the alleged offender also has the right to appeal.
  8. Appeals would be taken to the same group that the committee could choose to escalate to (which group is under discussion).

Part of the goal here is the person who feels harassed does not need to look up different policies and points of contact depending which space it happened in. They can always contact the same email (or IRC). Mattflaschen-WMF (talk) 21:21, 14 August 2015 (UTC)Reply

Also, I suggest that all participants in the first-level group be required to be identified to the WMF. Exactly how the group is formed could be determined if people otherwise support this proposal. Mattflaschen-WMF (talk) 21:36, 14 August 2015 (UTC)Reply
I think this is a good start. I think that the Engineering Community Team is a logical group for the second level response. I'd like to see this group made up of WMF staff, technical volunteers, and one representative from ECT, since ECT has considerable knowledge about community structures and dynamics. I think that an email ought to be a sufficient place to report to--IRC is likely to be more complicated to monitor, and I don't think it's necessary as long as emailed reports get a prompt response. --Fhocutt (WMF) (talk) 02:38, 15 August 2015 (UTC)Reply
I don't have a strong preference to keep an IRC point of contact. People might still informally report stuff there, but the email can be the definitive way. Mattflaschen-WMF (talk) 00:12, 18 August 2015 (UTC)Reply

Some notes and suggestions:

  • Anonymizing reports are terribly important here. Specially when it comes to sexual harassment. we need more explicit statements on this to ensure people to feel more comfortable. I think CA can help us.
    Yes, this should be addressed in the draft. Maybe, in case of private harassment, not allowing release of the victim/reporter's name (even in the course of a public apology) without the reporter's consent. In the case of public harassment (at least where it is not/can not be oversighted), the victim/reporter is already known, and since it happened publicly it can sometimes be important to speak out. Mattflaschen-WMF (talk) 00:26, 18 August 2015 (UTC)Reply
  • In lots of cases people already made contact with local maintainer of space (e.g. someone harassed me in mediawiki.org and initially I will contact admins of mediawiki.org but in some cases they don't handle it well. Lack of experience from admins, influential friends, being "too" valuable, etc.) So we need to determine a way to handle such cases.
    Yes, this is a good argument for having a central point of contact (techcoc), plus an appeal process in case even that fails. Mattflaschen-WMF (talk) 00:26, 18 August 2015 (UTC)Reply
  • Punishment vs. protecting the system: Imagine I offended someone in pywikibot code review. What is the best action? Do you think I should be banned for a while or I should be stripped off from my owner rights or both? this states "Project maintainers who do not follow or enforce the Code of Conduct may be permanently removed from the project team." Should we have something like this?
    We have language about this from Contributor Covenant now. Not sure how much we'll tweak this; we will probably need a Wikimedia-specific enforcement section at any rate. Mattflaschen-WMF (talk) 00:26, 18 August 2015 (UTC)Reply
  • Removal of comments: you didn't define anything about this. e.g. the CoC says: "Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct. By adopting this Code of Conduct, project maintainers commit themselves to fairly and consistently applying these principles to every aspect of managing this project." Ladsgroup (talk) 08:22, 15 August 2015 (UTC)Reply

From the perspective of a maintainer

In other sections, others have already clarified why a code of conduct is a good idea from the perspective of potential contributors and victims, but I'd also like to share the perspective of a maintainer. On the Pywikibot mailing list, we have had our fair share of behavior that is not OK, but I've always been reluctant to actually put the users on moderate, or to ban them from the mailing list altogether. Why? I wasn't sure whether that was the right thing to do, and I was afraid of the backlash that would probably follow.

Having an explicit policy makes policing behavior also much easier. From one side, it's much easier to say 'hey, you are violating the code of conduct', and from the other side, having the policy means we //have to// intervene. It's much easier to act on a previously agreed policy than acting off the cuff. Let's make clear rules, and let's use them to make this environment more fun to work in. Valhallasw (talk) 21:31, 16 August 2015 (UTC)Reply

Mailing list etiquettes already exist. --Nemo 07:17, 21 August 2015 (UTC)Reply
Are you referring to m:Mailing lists/Guidelines? I added this link to See also.--Qgil-WMF (talk) 09:00, 21 August 2015 (UTC)Reply


In the "Unacceptable behavior" section, we start of by saying that unacceptable behavior includes a variety of things, one of which is harassment. We then continue by saying harassment includes a whole list of things, which includes all the other unacceptable behaviors we listed in the first place. How about we simplify this by saying the following?

All forms of harassment are unacceptable in Wikimedia technical spaces, both online and in person. This includes but is not limited to the following behaviors:

  • Inappropriate or unwanted communication, including on-wiki messages, private or public IRC messages, email, texts, and chats.
  • Sustained disruption of discussion, talks or work, for example by personal attacks, constant interruption, trolling or consistent and unwarranted rejection of patches.

Neil P. Quinn (talk) 22:16, 18 August 2015 (UTC)Reply

@Neil P. Quinn-WMF: I see what you mean. I think it's important to state that we do not want our technical spaces to include "sexual language or imagery, derogatory comments or personal attacks, trolling, public or private harassment, insults", regardless of whether they rise to the level of harassment or not--even if these things are drive-by one-off comments, they make technical spaces less welcoming for many people. How about mentioning that harassment can include any of the above and removing them from the list of potentially harassing behaviors instead? --Fhocutt (WMF) (talk) 00:46, 19 August 2015 (UTC)Reply
Take a look now--I think this version is clearer. --Fhocutt (WMF) (talk) 01:17, 19 August 2015 (UTC)Reply
@Fhocutt (WMF): I think that looks good. I'm going to rearrange it to put the section about non-harassment unacceptable behavior below the list of examples of harassment.—Neil P. Quinn (talk) 17:54, 23 August 2015 (UTC)Reply
Oh, wait. Someone else already did that.—Neil P. Quinn (talk) 17:55, 23 August 2015 (UTC)Reply
The edit summary re-adding "rejection of patches" seems to go out of its way to point out that it's happening elsewhere. I don't know of this ever being a problem in Gerrit or Phabricator or Special:CodeReview. Why should this be included? --MZMcBride (talk) 02:35, 19 August 2015 (UTC)Reply
Because it's a good thing to build processes around problems we have seen in similar spaces with similar dynamics and similar base populations so that we have a way of handling them if and when they happen. This question or something very likes it comes up a lot, particularly in regards to codes of conduct; "why do we need to protect X? I've never seen X happen". And the answer is that if X does happen in communities similar to ours it's hubris to believe it couldn't happen here, and we should protect against that - and that the lack of the appearance of X, when there is no process for X, can also be indicative that the lack of process caused an absence of any ability to productively surface it. Or to put it another way, a problem without reporting looks the same as no problem at all. Ironholds (talk) 19:20, 19 August 2015 (UTC)Reply
This sounds very much like you're trying to prove a negative. --MZMcBride (talk) 02:01, 20 August 2015 (UTC)Reply
Not really. Again, this happens very commonly in communities very like ours; I'm not trying to prove a negative, I'm trying to avoid the hubris of assuming we should not ward against things we have seen appear in similar spaces. If it was entirely a thought experiment it would be potentially problematic because then we're just adding stuff in an unfounded way, but by definition worrying about things identified in near-identical entities is not unfounded, it is very well founded. Ironholds (talk) 14:16, 20 August 2015 (UTC)Reply
I'm a bit lost in this discussion. Are you proposing a specific change in the "Unacceptable behavior" section?--Qgil-WMF (talk) 09:03, 21 August 2015 (UTC)Reply

Strings attached

I removed this text about trying to coerce users into apologizing through threats and intimidation. Ironholds seemed to want a discussion about it. Perhaps it's a cultural difference, but it seems/seemed pretty obvious to me that this type of behavior is inappropriate and I don't think we should be encouraging it. --MZMcBride (talk) 02:54, 19 August 2015 (UTC)Reply

That isn't what it says. What it says is "The group may, if it chooses, attach "strings" to this request: for example, the group may ask a violator to apologize in order to retain their membership on a mailing list or their Phabricator privileges." That's not "threats or intimidation"; it's saying that in order for you to retain the privileges that you have after violating the behavioural standards that the group has, you must demonstrate that you understand that you violated these standards and that doing so was wrong. This is pretty fair, to me (I'm not sure what cultural difference there is here; yes, I'm a European, but I've also been living in the United States for quite a while now) - such a demonstration serves as evidence that misbehaviour will not occur again. It's also pretty standard; looking just at enwiki, almost every discussion around unblocks and unbans requires some assurance that the person who was sanctioned understands what they did wrong. Requiring that understanding to continue is...precisely what we do, we're just shuffling the order here. Ironholds (talk) 19:18, 19 August 2015 (UTC)Reply
This is from the jQuery enforcement manual. I think realistically this is something that is already done in certain situations informally (and not just in tech communities); they just chose to spell it out. Without this, either it could still happen informally, or the only option is to directly ban them in a case where the string option would be considered. Mattflaschen-WMF (talk) 21:46, 19 August 2015 (UTC)Reply
I also don't see any problem with requiring users to apologize as a condition of their remaining part of the community. The whole point of this code is to provide a mechanism for coercing toxic users to either change their behavior or leave the community. We can certainly argue about what makes a user toxic or how the mechanism should work, but it seems pretty obvious to me that a reasonable degree of coercion is the whole point.—Neil P. Quinn (talk) 22:22, 19 August 2015 (UTC)Reply
I don't really like the strings language. I kind of feel like it reduces the value of an apology when it is so forced. Apologies should be free admissions of wrong doing. I feel like a forced apology doesn't shows that the user understands s/he violated standards, it shows the person understands that they are being required to apologize in order to continue to participate. However I think its fine for the committee to take the presence of an apology (Or any other action the offender takes to make amends/demonstrate regret) into account when deciding appropriate reprimands or to strongly suggest that somebody apologize. Bawolff (talk) 07:16, 23 August 2015 (UTC)Reply
Demanding apology is an old antipattern. Apologies are valuable because they signal a change of beliefs (as opposed to a mere change of behavior). By making them obligatory most of that value is lost. Besides, giving a group the power to force belief changes is kind of creepy.
Asking the violator to promise that they will honor the code in the future (possibly as a condition of lifting the ban), which is the enwiki practice Ironholds was referring to, is a very different thing and entirely reasonable. Although I am not sure if there is any benefit in doing it in public. --Tgr (WMF) (talk) 08:32, 23 August 2015 (UTC)Reply
I agree that a required apology is not a 100% sincere apology. As I said above though, I think these kinds of required apologies are common in other areas of life. That said, I don't object to removing the strings part, and changing it to something like "The committee may recommend that someone apologize, and take into account whether they do so. They also may require that someone explicitly agrees to comply with the code of conduct before unbanning them." Mattflaschen-WMF (talk) 23:30, 23 August 2015 (UTC)Reply
This is actually not currently in the draft, but it's worth discussing whether to put something else back in (see my last suggestion). Mattflaschen-WMF (talk) 00:08, 24 August 2015 (UTC)Reply


I removed trolling as it seems to be problematic to define. I wouldn't want to see rickrolling to be considered a violation of this document, for example. I've also heard some people reclaim and embrace the term trolling as a positive behavior/attribute/skill. If someone wants to re-insert this, I think it first warrants discussion and perhaps a clearer definition. --MZMcBride (talk) 03:05, 19 August 2015 (UTC)Reply

"In Internet slang, a troll (/ˈtroʊl/, /ˈtrɒl/) is a person who sows discord on the Internet by starting arguments or upsetting people, by posting inflammatory,[1] extraneous, or off-topic messages in an online community (such as a newsgroup, forum, chat room, or blog) with the deliberate intent of provoking readers into an emotional response[2] or of otherwise disrupting normal on-topic discussion.[3]". I think that firmly falls within the set of behavior we don't want to see. In addition, it is noted explicitly in the Contributor Convenant, so I think it makes more sense to place it back while the subject is being discussed. As for Rick Rolling: I think that generally is also counterproductive, and thus something we don't want in our technical spaces either. Valhallasw (talk) 08:02, 19 August 2015 (UTC)Reply
Trolling is defined in Wikimedia terms at Manual:Glossary#T and m:What is a troll?, and in case of doubt w:Internet troll is also an interesting read. I think in most cases it is a behavior simple to recognize. "consistent and unwarranted rejection of patches" is not an equivalent of trolling but a much more narrower case, therefore I think that we should bring "trolling" back instead.--Qgil-WMF (talk) 09:50, 19 August 2015 (UTC)Reply
Agreed, and let me say that people reclaiming trolling as "a positive behavior/attribute/skill" have no place here. Ironholds (talk) 19:09, 19 August 2015 (UTC)Reply
I agree with Valhallasw. There was unanimous consent to start from the Contributor Covenant. While we have made changes from there, 'trolling' is one of the few things the concise Contributor Covenant addresses, which shows they view it as important. I agree it is important to forbid, and I see no reason to remove it from our versions. Rick-rolling is a prank, not trolling. While pranks can waste time (which is why I try to restrict mine to April Fool's), pranking is not the same thing as trolling. Mattflaschen-WMF (talk) 20:59, 19 August 2015 (UTC)Reply
Thanks for re-adding it, Matt. Ironholds (talk) 14:14, 20 August 2015 (UTC)Reply
Trolling is not defined anywhere and has never proved useful as an explicit charge to make to anyone, hence I support the removal of its mention. m:What is a troll? is a mere essay for a reason. --Nemo 06:59, 21 August 2015 (UTC)Reply
The question is, are the behaviors identified as "trolling" already mentioned in clear terms (making the addition of "trolling" redundant) or is "trolling" covering other behaviors that are still not mentioned? I would be fine removing that single word, but only if the behaviors we want to address are well covered in the CoC. Currently we have:
  • Inappropriate or unwanted communication
  • Sustained disruption
  • Offensive, derogatory, or discriminatory comments
  • deliberate intimidation
  • derogatory comments
  • personal attacks
Anything missing?--Qgil-WMF (talk) 09:21, 21 August 2015 (UTC)Reply

Qgil-WMF: If we remove "trolling," as I think we should, we would be left with:

The following behaviors are also unacceptable, regardless of whether they rise to the level of harassment and whether online or in person: inappropriate use of sexual language or imagery, insults, and other unprofessional conduct.

I think "unprofessional conduct" is similarly problematic as it's incredibly vague. And is "insults" needed? Are insults different from "personal attacks" or "offensive, derogatory, or discriminatory comments"? I'd be inclined to remove these three pieces (maybe moving "insults" up into a bullet, though it seems unnecessary). That would leave us with:

The following behaviors are also unacceptable, regardless of whether they rise to the level of harassment and whether online or in person: inappropriate use of sexual language or imagery.

This seems pretty awkward/weird to be in its own separate paragraph. --MZMcBride (talk) 00:45, 23 August 2015 (UTC)Reply

Trolling used to have somewhat positive connotations back in the newsgroup era, but has that changed pretty definitively since then. Partly because its meaning changed to encompass any kind of deliberate manipulation, and to imply malicious intent; and partly because most newsgroups weren't communities in the sense the MediaWiki tech community strives to be one (they weren't about building things), so different behaviors were problematic there. Trolling, as used today, fits well into the list of non-acceptable conducts. If there are non-hypothetical persons in our community who are trying to reclaim it, that might be worth a discussion; otherwise, I wouldn't worry much about it.

Qgil: something like hostile tone maybe? A code of conduct might not be the best place to address that, though. --Tgr (WMF) (talk) 09:06, 23 August 2015 (UTC)Reply

"Tone" was specifically raised under #Creeping bureaucracy. It could be a rationale to block anyone for writing almost anything "non-positive". There are no community agreed policies which rely on bad "tone" in order to justify blocks or other sanctions; it would be worrying if there were. -- (talk) 14:31, 23 August 2015 (UTC)Reply
I thoroughly disagree with the assertion that it is not possible to write about a negative thing in non-hostile tone. I also don't think it differs fundamentally from WP:CIV (how much the latter gets followed is another matter). --Tgr (WMF) (talk) 18:52, 23 August 2015 (UTC)Reply
I agree. I don't think we need a point specifically about tone. Mattflaschen-WMF (talk) 23:41, 23 August 2015 (UTC)Reply
People have defined 'trolling' above. I completely disagree with removing it. There is a reason it's in Contributor Covenant. In both their and my experience, it's a serious problem in communities like ours (and I have seen examples of it in our community). If people really think a definition needs to be incorporated into the CoC, I'm fine with quoting the definition from Wikipedia (which Valhallasw quoted above). I think it's clear that while there may be some overlap, not all the items from that definition (e.g. "provoking readers into an emotional response") is in the list Quim quoted. I think "unprofessional conduct" should be kept; it basically means the kind of disruptive behavior that is incompatible with getting things done in a respectful way. I'm okay with removing "insults", as I think "personal attacks" is a superset of "insults". Mattflaschen-WMF (talk) 22:15, 23 August 2015 (UTC)Reply
+1. Ironholds (talk) 14:45, 24 August 2015 (UTC)Reply

Inappropriate or unwanted communication

"Inappropriate" was mentioned above; I agree that it does not seem very useful - it includes everything but explains nothing. There are generally three functions for the list of unacceptable behaviors: they inform the enforcers; help a concerned community member decide whether something they are about to decide or something someone just posted is acceptable or not; and tell a potential newcomer whether they can trust that the community is a safe space. "Inappropriate X" doesn't seem terribly useful for either goal. And I am not sure what unwanted even means. If it means that once someone tells you to stop communicating with them you must do comply, that is fine, but should be spelled out more explicitly. --Tgr (WMF) (talk) 09:14, 23 August 2015 (UTC)Reply

I think there are some kinds of inappropriate messages that most reasonable people would agree are inappropriate and should not be sent. As far as "unwanted communication", I've clarified that similarly to how we did for unwanted photography. Mattflaschen-WMF (talk) 23:35, 23 August 2015 (UTC)Reply
I think simplicity is more effective. Somehow we have grown an own list that is far more dense and extensive than the original Contributor Covenant, and I'm not seeing a good, practical justification for that. We already have several instances of inappropriate / unwanted communication. That first point could be removed without creating any new gap for actual harassment.--Qgil-WMF (talk) 10:56, 24 August 2015 (UTC)Reply
Specificity, as has been covered repeatedly in this discussion, is a pretty good way of ensuring there isn't wiggle room or ambiguity. I don't feel we should be treating the Contributor Covenant as the baseline here except in the sense of 'the minimum acceptable`. It's widely used because it's not objectionable. That makes it a good template; it doesn't mean we should pin ourselves to the belief that it is ineffable. Ironholds (talk) 16:41, 24 August 2015 (UTC)Reply
@Qgil-WMF: I don't think it's too long (especially if we just consider the length of the sections prior to Reporting and Enforcement). There are advantages to being specific. It helps avoid gamesmanship and wikilawyering. One thing "Inappropriate or unwanted [...] communication" may address that the other points don't as well is written harassment. This might be written sexual harassment, or just some other form of persistent communication you don't want to receive (and ask not be sent to you). Mattflaschen-WMF (talk) 02:28, 26 August 2015 (UTC)Reply
Actually, gamesmanship and wikilawyering might be helped by lengthy policies inviting to have lengthy discussions driving most people away (an actual problem in Wikimedia). If the principles are clear and the committee is trusted, then this process will be accessible and strong. Now the text of the Principles proposed is shorter, but I bet we are offering the same actual coverage as before. "Inappropriate or unwanted communication" has been connected to "following, or any form of stalking", and in that context I think its function is clearer.--Qgil-WMF (talk) 07:39, 26 August 2015 (UTC)Reply
@Qgil-WMF: : Much of the added list comes from policies that have been working in WMF spaces so far: the Grants Friendly Space Expectations and the Friendly Space Policy. Also, what Ironholds said. --Fhocutt (WMF) (talk) 17:26, 24 August 2015 (UTC)Reply
One thing is being specific, another thing is having overlapping concepts spread in different points, making the whole list more dense and confusing than needed. I have just posted #Reorganization of types of harassment, a simpler proposal that doesn't lose specificity, that may address the problem exposed here and in the discussion about trolling.--Qgil-WMF (talk) 06:10, 25 August 2015 (UTC)Reply

Substantive v Procedural

The draft is a confusing mix of Substantive law and Procedural law.

  • Substantive law declares which types of conduct lead to which types of penalty;
  • Procedural law defines how cases are handled.

Please place the procedural law into a separate paragraph. That paragraph should define:

  • the judicial body (e.g. committee membership and powers);
  • how cases are to be filed (e.g. identify the parties, state the facts, state which substantive law was breached, request remedy);
  • the rules of evidence (e.g. original documents and testimony is admissible; hear-say is inadmissible);
  • how cases are adjudicated (e.g. read charges, determine facts, determine applicable substantive law, decide penalty);
  • how decisions are announced (e.g. three brief paragraphs: findings of fact, findings of law, and penalty).

Procedural law is about the judicial system. If the work load grows, it may become necessary to have many such committees. Procedural law is our guarantee that the substantive law is applied in a uniform manner across many cases and many committees.

Uniform handling of cases is an important component of fairness. Moreover, public perception of uniform handling of cases contributes to public perception of fairness. Wp mirror (talk) 03:09, 19 August 2015 (UTC)Reply

Agreed they should be split into different sections but they should be considered part of the same document, jurisprudence aside. This is not intended to be a legal system and should not be read as a legal document; as you say, fairness is important, and uniform handling and transparency are important parts of that fairness, but speaking as a former lawyer another element of fairness is to ensure that justice is efficient. Replicating a legal system down to the level of specifying the determination of "applicable substantive law" and every single possible type of evidence that is and is not acceptable does not necessarily help with this. Ironholds (talk) 19:38, 19 August 2015 (UTC)Reply
An example might help. Procedural law, for many not-for-profit organizations, includes a reference to RONR 10th ed, Chap XX Disciplinary Procedures. The draft contains no such reference. Procedural law for the Conduct Committee should be gathered into one or more paragraphs, and thought through. Otherwise, committee members will find themselves having to improvise when they receive their first few cases. Improvisation leads to an aggrieved party claiming that committee members act in an arbitrary and capricious manner, which in turn injures the public perception of fairness and shrinks to pool of volunteers willing to serve on the committee. Wp mirror (talk) 05:29, 20 August 2015 (UTC)Reply
In short, should we create a section "Code of Conduct Committee" and move there the details about its creation, membership, second level...?--Qgil-WMF (talk) 09:45, 20 August 2015 (UTC)Reply
Yes, I think it's fine to create separate (sub)sections for that. Mattflaschen-WMF (talk) 02:38, 21 August 2015 (UTC)Reply
I have never seen a reference to Robert's Rules (which are incredibly US-centric) in any code of conduct for this sort of community. Could you point me to an example of a code of conduct for a technical or online community that contains such a reference? Quim: yes, I am reading this as a request to create a dedicated section about the entity tasked with enforcing it, and endorse that request (although Code of Conduct Committee may not be the best name, since it leaves ambiguous whether that is an enforcement body or an amending body or or or). Ironholds (talk) 14:13, 20 August 2015 (UTC)Reply
What part of the substantive six-step process do you feel is unclear? I agree the committee may have to choose certain non-substantive manners themselves (e.g. do they meet by private chat room, or telephone, who takes notes?) but the substantive matters should be addressed here in the policy. Mattflaschen-WMF (talk) 00:21, 21 August 2015 (UTC)Reply

@Mattflaschen-WMF: Would anyone object to my starting a page called "Conduct Committee Charter/Draft". The idea is this: There should be two pages:

  • Code of conduct for technical spaces/Draft - contains substantive law (definitions of unacceptable behavior, penalties);
  • Conduct Committee Charter/Draft - contains procedural law (committee structure, membership, responsibilities, procedures).

I would be happy to provide a first draft to get the conversation going. The outline would look like this:

  • Purpose of Conduct Committee
  • Membership and selection
  • Responsibilities
  • Procedures and process

Wp mirror (talk) 00:28, 21 August 2015 (UTC)Reply

They should not be separate pages, they should be the same page, as I thought we'd all agreed on. I am loathe to set more than the high-level elements of the appeals process due to the potential (in fact, the certainty) that we would get bogged down in that debate. The committee should determine, within the restrictions of the code of conduct, their own policy (and transparently document it). Ironholds (talk) 00:58, 21 August 2015 (UTC)Reply
I agree with Ironholds on both points. The committee will quickly gain a sense of what is effective in practice; let's provide a reasonable blueprint and clear goals here and let them do their job. --Fhocutt (WMF) (talk) 01:44, 21 August 2015 (UTC)Reply
Yes, I do object to drafting a new charter from scratch. There are already important procedural aspects in the main document. The key procedural questions should be described in the Code, but I do not support drafting our own very-detailed procedure in a vacuum. Many of the procedural elements (including resolutions) are based on https://github.com/jquery/jquery.org/blob/master/pages/conduct/reporting.md and https://github.com/jquery/jquery.org/blob/master/pages/conduct/enforcement-manual.md . This in turn is based on https://www.djangoproject.com/conduct/reporting/ and https://www.djangoproject.com/conduct/enforcement-manual/ . The key insight here is that we are benefiting from the procedural experience of real organizations (Django, and the others they adapted from, such as PyCon) that have dealt with these problems. Making our own elaborate procedural system based on Roberts Rules of Order, but without practical experience with dealing with this, will lead to an unwieldy procedure that does not solve the problem.
As already noted, I expect the committee to answer some non-substantive procedural questions on their own. It is also possible they will ask for some substantive procedural changes, in which case they may request to amend the CoC. But in either case, we'll be benefiting from practical experience in this kind of committee, which we can't if we draft our own based on Robert's Rules of Order. Mattflaschen-WMF (talk) 03:44, 21 August 2015 (UTC)Reply

Physical contact

MzMcBride removed the parenthetical in "Inappropriate or unwanted attention, touching, or physical contact (sexual or otherwise)" with the rationale that it felt "awkward and unnecessary". To kick off a discussion about that:

Sexual versus non-sexual touching has different levels of acceptability within western society (speaking very broadly, there). Sexual contact is seen as obviously inappropriate for professional spaces, and non-sexual totally fine if both parties are comfortable with it. This split in characterisation should be called out when we're talking about how to handle unwanted contact. Ironholds (talk) 19:30, 19 August 2015 (UTC)Reply

I think the key point is that unwanted or inappropriate touching is prohibited either way (sexual or otherwise). I agree it's worth calling out, since sexual harassment is one of the problems this is trying to address. I've put it back. Mattflaschen-WMF (talk) 21:27, 19 August 2015 (UTC)Reply
Which split in characterization should be called out? I don't follow what you're saying. I also still don't see the need to parenthetically clarify that physical contact can be "sexual or otherwise." --MZMcBride (talk) 02:16, 20 August 2015 (UTC)Reply

This area seems more than slightly over-egged, possibly this is a cultural gap of some sort between Europe and the US. If I were at a lunch at a Wikimedia sponsored hackerthon, and a fellow volunteer started choking on their chicken sandwich, I would like to avoid feeling so awkward about potential embarrassing claims of inappropriate touching, that I never tried to slap their back or attempt abdominal thrusts to clear their airways before they lose conciousness. At the moment I'm thinking that I would step back and wait until someone else helped out, especially if the person choking were half my age (which is quite likely). -- (talk) 09:50, 20 August 2015 (UTC)Reply

In the event that somebody was choking to death would your read genuinely be that the Heimlich was unwanted? Ironholds (talk) 14:06, 20 August 2015 (UTC)Reply
I think this is unhelpful hypothetical argument. I'll respond to this one direct question, but I prefer to avoid getting into replying to more of this type.
The way the draft is being extended in ways that only relate to thought experiments, rather than case history for our community, is of concern and especially the emphasis on "touching" as a reason to apply this policy. My feelings on this are from the last few years of targeted sexual and sexuality related abuse/vandalism on and off the Wikimedia projects (one of these attacks was on my Commons talk page in the last 24 hours, and has been deleted from the page history). This past does make me concerned about the risks of socializing at physical Wikimedia events, and has made me concerned about practicalities of things like sharing rooms when I attended Wikimedia events.
If I take your comment literally, then there is plenty of precedent and I note that preferred description is "abdominal thrusts" as other terms can be confusing or controversial depending on what you read. -- (talk) 15:11, 20 August 2015 (UTC)Reply
I concur the choking example was unhelpful and I'm glad we're in agreement on that. You seem to be making a false distinction, here; that there are "thought experiments" and "things we have seen in our community" and nothing in-between. What we are talking about is not thought-experiments that could never happen, or could happen but never do happen; we are talking about things that absolutely happen in communities incredibly similar to ours in terms of focus, demographics and process. It is hubris for us to take the position that because instances of this have not been reported - in an environment similar to those where it has, and an environment lacking an existing enforcement mechanism in many ways - they cannot, or will not, happen.
I am deeply sorry for your past experiences of harassment in this community, along those axes or along any other axes. Ironholds (talk) 16:43, 20 August 2015 (UTC)Reply
With regard to "axes", sarcastic apologies that are deliberate personal attacks are just as unhelpful. We should all expect better behaviour of WMF employees if the Wikimedia community is to provide them with special trust and authority. -- (talk) 16:56, 20 August 2015 (UTC)Reply
That was not sarcastic, nor do I see where it was a personal attack. I was genuinely saying that I am sorry for your experiences. I apologise if what I wrote was badly written or ambiguous and so came off as an attack in any way; I will endeavour to be more precise before hitting save. Ironholds (talk) 18:05, 20 August 2015 (UTC)Reply
I've now realised the ambiguity might be around "axes"; I meant it in the sense of "the plural of axis" not "the plural of axe". Apologies for the ambiguity :(. All data visualisation and no play makes.... Ironholds (talk) 14:41, 21 August 2015 (UTC)Reply
@: The precedent you cited is from Weekly World News, a fictional tabloid publication. I'm sorry about any harassment you've received on any of the projects. This draft code of conduct does address communication ("Inappropriate or unwanted communication", "Offensive, derogatory, or discriminatory comments", etc.), not just touching, in the covered spaces. Mattflaschen-WMF (talk) 00:09, 21 August 2015 (UTC)Reply


By the way, this whole discussion seems dominated from the start by a cherry-picked pre-determined very small subset of the community, which convened in external places and then started this concerted effort. In our wiki culture, this smells of so-called canvassing, which by definition makes a discussion and decision invalid and void. --Nemo 06:46, 21 August 2015 (UTC)Reply

This proposal, just like most proposals in our community, was started publicly by some people who cared about this topic, and others equally interested have been involved since then. Nobody has been sherry-picking anybody, and nothing has been pre-determined. Trust me, I have been following this process and its collateral discussions and initiatives since phab:T87773#998422 (2015-01-28), both in public spaces and in the casual private conversations where I have been CCed.--Qgil-WMF (talk) 08:05, 21 August 2015 (UTC)Reply
I learned of it from wikitech-l (2015-08-07) which is one of the technical spaces that would be covered by the proposed Code of Conduct. If anyone is aware of a technical space not notified, we should remedy that now. Wp mirror (talk)
Basically what Matt said; I heard about this from the Wikitech thread and to my knowledge it started on a phab task and with a public meeting at Wikimania. That's about as far from canvassing as you could get. Ironholds (talk) 14:35, 21 August 2015 (UTC)Reply
I learned about this from the same Wikitech thread as Ironholds and Wp mirror.—Neil P. Quinn (talk) 05:24, 22 August 2015 (UTC)Reply
I learned about this draft by personal email from a deeply concerned long term technical contributor. They have not contributed here themselves, from my understanding of their history, the most likely reason being that they do not want to be seen to be 'non-positive' about any WMF driven initiative, and with weaker English skills their voice would be lost in the wiki-lawyering. Without the email, I would never have known that this draft had been created. -- (talk) 08:21, 22 August 2015 (UTC)Reply
I don't think I would say canvassing - but it certainly does seem as if it is being pushed by a rather small group of interested people, and being pushed rather hard. I am somewhat concerned in how the discussion doesn't seem to have all that many voices among the long term mediawiki/core developers (Since that is one of the main groups this document plans to regulate), but there is a long thread on wikitech-l, so its not like the relavent parties don't know about it. If the people commenting are the people interested, I guess that is that. Bawolff (talk) 06:48, 23 August 2015 (UTC)Reply
As other people have noted, this work started publicly on Phabricator, and I've sent multiple notifications to wikitech-l. Architecture committee/2015-08-12 indicates the architecture committee members (i.e. long-term contributors) are aware of it; I would expect anyway they're aware from wikitech-l. If you feel it should be sent out to another public forum, please do so. Mattflaschen-WMF (talk) 20:57, 23 August 2015 (UTC)Reply


jQuery's Code of Conduct Committee is a creature of the jQuery Foundation's Board of Directors:

  • the board appoints and reviews the committee members;
  • the committee is empowered to speak on the foundation's behalf to contact witnesses and to stop ongoing situations;
  • the committee reports the outcome of each case to the board;
  • deadlocked cases are escalated to the board; and
  • public statements are issued by the board, never the committee.

This is documented in their Enforcement Manual.

  • appeals go to the jQuery Foundation's President and Executive Director.

This is documented in their Reporting Guide.

No equivalent language appears in the draft Code of Conduct under consideration here. These are issues of governance; and they need to be addressed. Is there any reason why the proposed Conduct Committee should not be a creature of the WMF Board of Trustees? Wp mirror (talk) 17:59, 21 August 2015 (UTC)Reply

Time, existing org structure, and lack of technical background, mostly. The jQuery Foundation is very different from the WMF and the two organizations and their boards focus on different tasks. This proposed committee's purpose (to promote and maintain a healthy technical community) is well within the Engineering Community Team's purview and skillset and I think it is a more natural fit. --Fhocutt (WMF) (talk) 00:38, 22 August 2015 (UTC)Reply
That seems unsatisfactory. Members of WMF technical staff will be actively working in spaces where this code is in force, and subject to its provisions. It cannot be right for the ECT to sit in judgement on, or take appeals from, its own members or fellow-worker -- the clear failure of natural justice implied would completely undermine the credibility, acceptability and authority of the Code in all its applications, not just those that might involve a member of the WMF engineering staff. The authority of the Code must surely emanate from the Board, which is the ultimate authority and embodies the community consensus through its elective structure. It seems unlikely that there will be large amounts of business for them to transact and if it prove burdensome, I am sure the Board is fully capable of managing its own business either by delegation ad hoc or to a standing subcommittee. Rogol Domedonfors (talk) 18:53, 23 August 2015 (UTC)Reply
Presumably the WMF's HR department should also not be involved in WMF employee behaviour, nor their lawyers, nor their managers, since they're all "fellow workers", and the committee should not contain any technical people since then the technical contributors they're tasked with checking the behaviour of will be "fellow workers" too. There are conflict of interest and recusal provisions practically by default for entities like this. The Board does a vast amount of work already and is not a particularly technical body; it also has little to no exposure to work within the technical spaces. Your proposal is that to avoid conflicts of interest we abstract the task away so far it lands on the plate of a group of people with absolutely no expertise or lived experiences that pertain to the space-specific problems. Ironholds (talk) 14:43, 24 August 2015 (UTC)Reply
I am sure that the trained professionals in the HR and Legal departments are fully capable of handling those issues. That is of course of marginal relevance to this discussion and I presume you knew that when you dragged it in.. My proposal -- that is, what I actaully wrote rather than your mis-statement of it -- is based on the belief that the Board has the authority, the independence, the community support, that it is fully capable of understanding the issues of human interaction, fully capable of managing or delegating its own business and fully capable of obtaining technical advice should it see the need to do so. Apparently you disagree and it would be interesting to know why? It seems that you find independence a disadvantage in dealing with these governance issues -- I would see it as a positive advantage. Rogol Domedonfors (talk) 08:39, 25 August 2015 (UTC)Reply
I don't see any problem in asking the Board for their opinion once this draft is more consolidated. If they say "Yes, we want to be the ones enforcing this Code of Conduct directly", I don't see any strong reason not to agree with them. However, I find this possibility unlikely for practical reasons, and we'd rather have an alternative proposal where the Board is either not in the picture or basically nominally, with their duties delegated to an executive body, aka a committee. In any case, the first step is to consolidate this draft, which we are doing at #Proposing Intro + Principles + Unacceptable behavior and related sections.--Qgil-WMF (talk) 09:41, 25 August 2015 (UTC)Reply
I don't have any issue with independence; I have a problem with independence to the point of lacking context, and concerns about the time the board has available. But in any case, I endorse Quim's comment on this. Ironholds (talk) 14:40, 25 August 2015 (UTC)Reply
I think there is in fact language addressing the same questions. It's just that the current draft has different answers. The draft is quite clear about who speaks (e.g. "A private reprimand from the committee", "A public reprimand. In this case, the group chair will deliver that reprimand", etc.), and it's also clear about the appeals process ("To appeal a decision of the committee, or if you have concerns about how it was resolved, contact the Engineering Community Team at ect@wikimedia.org and they will review the case"). The main question not addressed is how the committee will be formed. There is already active discussion about that.
I think ECT has better understanding of day-to-day life in the technical community than the Board of Trustees does, and is thus better suited as the appeals body. Mattflaschen-WMF (talk) 22:47, 23 August 2015 (UTC)Reply


Hi. Re: this edit, "outing" actually has a different meaning in the context of Wikimedia wikis (cf. w:en:WP:OUTING). --MZMcBride (talk) 00:29, 23 August 2015 (UTC)Reply

I agree. Without looking at the link, I would assume the other definition was meant in this context. Bawolff (talk) 06:23, 23 August 2015 (UTC)Reply
Agreed. I've pointed the link to the enwiki project page you linked.—Neil P. Quinn (talk) 18:05, 23 August 2015 (UTC)Reply
The original Contributore Covenant says "Publishing other's private information, such as physical or electronic addresses, without explicit permission". This is plain English, and very clear. I propose to go back to this original sentence. "Outing" and "doxing" are not part of everyday's vocabulary. If you prefer, we can add the concept of "person's identity": "Publishing a person's identity or other private information, such as physical or electronic addresses, without explicit permission".--Qgil-WMF (talk) 10:46, 24 August 2015 (UTC)Reply
The version of Contributor Covenant we based this on does not have that text (see revision link in See also). That said, it seems fine to use more plain language for this. Mattflaschen-WMF (talk) 02:20, 26 August 2015 (UTC)Reply
It should be amended as you suggested, since unlike some communities, in ours name/identity is often private, which is not addressed in your first quote. Mattflaschen-WMF (talk) 02:22, 26 August 2015 (UTC)Reply
  Done taking the wording proposed at #Reorganization of types of harassment since it is directly based on the previous sentence, just removing the very specialized terms.--Qgil-WMF (talk) 06:12, 26 August 2015 (UTC)Reply

Remove "Principles" section?

The "Principles" section as currently written seems both incorrect and pointless. There's a reason why the coding conventions pages, for instance, don't say "As Wikimedia developers, we all believe that tabs should be used instead of spaces." If there are rules, fine; but there's no point pretending there's unanimity behind them.

The first sentence of the section is simply incorrect: "As members of Wikimedia's technical community, we work together to build technology that makes free knowledge available to humanity." There are plenty of MediaWiki extension developers, myself included, who have never worked on code used on any of the major Wikimedia sites; we thus make no real contribution to free knowledge for humanity. (Why this sentence belongs there, even if true, I don't know. If the goals were less lofty, would harassment be okay?)

The second two sentences, starting with "We welcome" and "We pledge", are actually similarly nonsensical. Did everyone pledge? (Did anyone pledge?) What does it mean to pledge? Do you have to pledge repeatedly, or only once?

I don't know why you need a set of principles to explain rules in the first place. It feels like this whole section could just be removed. Yaron Koren (talk) 18:44, 24 August 2015 (UTC)Reply

I disagree that there is no point to this section although I accept that you don't see one. The point of this is the same as the point of any preamble anywhere; to provide a context within which the document can exist and be interpreted. So that people reading the code of conduct can understand not only what specific activities are proscribed but the principles under which the community is organised.
As for the 'there's no unanimity behind them' or 'who pleged?' - that is about the principles, not the individual rules. Excepting the working together section, do you disagree that the community should be a respectful and harassment-free experience for everyone, say? Ironholds (talk) 04:34, 25 August 2015 (UTC)Reply
I think Yaron's point, is that pledging is usually defined as an active action. Almost akin to signing something. Regardless of whether everyone agrees with the statements to be pledged, its not true to say that everyone has pledged it, unless they've actually pledged those principles, which they haven't, and there is no plan to make everybody pledge it. A policy we all agree with, is different from a policy we have all pledged to uphold. Bawolff (talk) 05:30, 25 August 2015 (UTC)Reply
For what is worth, this section is basically an adaptation of the first paragraphs of the Contributor Covenant this Code of Conduct is based upon. "Pledge" is being used there. Today nobody has pledged anything, but the approval of this Coc implies that the Wikimedia Tech community endorses it. To me this reads as a linguistic detail, and in general I think that deviations from the original Contributor Covenant benefit from a strong justification. As per MediaWiki-not-Wikimedia contributors, anybody participating in the Wikimedia tech infrastructure is considered a Wikimedia tech contributor, and does contribute in a way or another to Wikimedia's mission. So yes, it is a preamble, it might sound a bit redundant for those knowing the details, but it is informative and usefl to set everybody in the same page.--Qgil-WMF (talk) 06:21, 25 August 2015 (UTC)Reply
Ironholds - I agree that respect is good and harassment is bad. Now, if you can get everyone who's ever contributed code to a Wikimedia project to agree with that, you'd have a stronger case...
Qgil - I simply don't see how someone who writes, say, a Google Ads MediaWiki extension is contributing to free knowledge for humanity. Right now the wording makes it seem like such people don't even count; is that really "welcoming"?
Seriously, though - instead of "We pledge to make participation in Wikimedia technical projects...", what about "Participation in Wikimedia technical projects should be..."? Yaron Koren (talk) 14:10, 25 August 2015 (UTC)Reply
If there are people who can't agree on that they shouldn't be members of our technical community, full stop. I have no issue with the people who are concerned by individual clauses and the edge cases they might introduce, but if you do not believe that, as you put it, respect is good and harassment is bad you have no place here. I don't have a problem with your rewording. Ironholds (talk) 14:41, 25 August 2015 (UTC)Reply
Well, that's refreshing. People have been wasting time arguing about warning levels, arbitration committees and so forth, when it turns out that all we needed all along is just an immediate lifetime ban. Yaron Koren (talk) 15:10, 25 August 2015 (UTC)Reply
I really don't want to spend my energy with us digging ever-deeper into this, so I'm going to back off at this point, but I really don't see how you concluded that a comment consisting entirely of deliberate facetiousness was helpful. Ironholds (talk) 18:08, 25 August 2015 (UTC)Reply
Yaron's valid point refers to the part of the community focusing on MediaWiki-not-Wikimedia. It's a topic that we are considering in other areas, and can be considered here as well. The incidence on the mission and purpose of Wikimedia is not essential in this Code of Conduct. In the original Contributors Covenant, the "pledge" is used in a slightly different way. All in all, here I don't see strong reasons to depart from the original writing, so I will edit the Principles section proposing a writing closer to the original. Feel free to revert if you prefer to discuss further.--Qgil-WMF (talk) 20:51, 25 August 2015 (UTC)Reply
I think it's better now, yes. Yaron Koren (talk) 22:42, 25 August 2015 (UTC)Reply

Reorganization of types of harassment

The list of types of harassment has grown to a point where I don't think it is serving its original purpose as good as it could. Having two levels (a bullet list and an extra sentence) doesn't help. Having similar concepts in different points doesn't help either. We can be just as specific with a clearer organization and self-explanatory descriptions. Here is an attempt:

Harassment and other types of inappropriate behavior are unacceptable in all public and private Wikimedia technical spaces. Examples include but are not limited to:
  • Personal attacks, violence, threats, deliberate intimidation.
  • Offensive, derogatory, or discriminatory comments.
  • Inappropriate use of sexual language or imagery.
  • Unwanted attention, touching, or physical contact (sexual or otherwise).
  • Unwanted communication, following, or any form of stalking.
  • Unwanted photography or recording.
  • Disclosure of a person's identity or other private information without their consent.
  • Trolling through sustained disruption, interruption, or blocking of community collaboration.
  • Other unethical or unprofessional conduct.
Project administrators and maintainers have the right and responsibility to...

What do you think?--Qgil-WMF (talk) 06:08, 25 August 2015 (UTC)Reply

They're in different sections because some of them are inappropriate but not actively harassing. 'inappropriate use of sexual language or imagery' is grossly inappropriate and should not be tolerated but I wouldn't describe it as harassment necessarily. You've also categorised anything unprofessional as harassment, which, again, is not always the case. Ironholds (talk) 06:17, 25 August 2015 (UTC)Reply
I would also consider some forms of trolling to not be "harrasment" (but still problematic). Bawolff (talk) 06:18, 25 August 2015 (UTC)Reply
Fine, but the goal of this CoC is not to define Harassment as in w:Harassment, but to identify types of "unacceptable behavior" in our community. The distinction between "technically harassment" and "other" is superfluous here. I have fine tuned the description.--Qgil-WMF (talk) 06:27, 25 August 2015 (UTC)Reply
@Qgil-WMF: Some of the changes seem okay. Others I disagree with.
  • I strongly disagree with only forbidding certain forms of trolling, which I think the new wording does. The Contributor Covenant forbids all forms. I don't object to adding a parenthetical definition of trolling, but I also don't believe that's genuinely necessary. If you want to merge in the others (from Sustained disruption), it could be done with "Trolling, for example by" (I'm not sure those are 100% the same as the definition of trolling, but it's probably close enough).
  • Your list keeps "unwanted communication" (albeit moved, which is fine), but removes "inappropriate [...] communication". This is somewhat hard to define, but basically I think of creepy, harassing, or upsetting communication (especially private communication) that is not appropriate in this setting. I would prefer to put that back.
  • No problem with shortening "Offensive, derogatory, or discriminatory comments"; it should be obvious it includes offensive comments related to the listed characteristics.
  • "threats" should be changed back to "threats of violence". If someone threatens to stop editing Wikipedia, that should not fall under this.
  • Would prefer to put back "Inappropriate or" before "Unwanted attention". Similarly, it is possible for touching to be clearly inappropriate without someone explicitly telling you it's unwanted.
  • The definition of Unwanted photography or recording was put in response to feedback, but if most people don't have a problem with this wording it could be kept simpler.
  • Personal attacks might now be ambiguous. We have to forbid both verbal and physical personal attacks (physical is handled by violence, though), but a reader might now think it only refers to physical. Maybe clarify it.
Other than that the changes seem alright. Mattflaschen-WMF (talk) 02:56, 26 August 2015 (UTC)Reply
Take a look to the current version. I think I have addressed your concerns. I have left the "Inappropriate or unwanted" pairs, although I still think "Unwanted" alone is clearer for everybody and more straightforward to act upon. In these sentences "unwanted" clearly implies "inappropriate". The "or" conjunction opens the possibility to deal with "inappropriate but not unwanted" behavior, which might be a can of worms. When not unwanted, what is appropriate/inappropriate differs a lot between individuals and their social contexts. In order to consider a behavior unwanted, we don't need "someone explicitly telling you it's unwanted". Social norms exist that define the lines of the unwanted, which are the same lines that define socially appropriate vs inappropriate. Those lines are fuzzy across cultures and individuals, but adding "inappropriate" to "unwanted" in our CoC does not solve that fuzziness either.--Qgil-WMF (talk) 06:57, 26 August 2015 (UTC)Reply
@Qgil-WMF: I think the new version is okay, as long as we clearly understand that personal attacks includes verbal and written personal attacks, such as insults. Mattflaschen-WMF (talk) 23:19, 27 August 2015 (UTC)Reply

Right to appeal

The right to appeal has been mentioned in several places. As a general principle, of course anybody being accused of breaching the CoC must have the right to appeal. The first ones having to excel at Conduct are the ones using it to file a complaint and the ones using it to enforce it on someone. The exact implementation of the appeal process can be defined only after the processes for starting and handling a complaint has been defined, though. For instance, complaints filed privately probably will have an appeal process that is equally private, unless one of the parts wants to bring the issue to the public. Let's agree first on how to file a complaint and how to handle it?

Let me add that the root of the questions about an appeal process seem to be based on the assumption or fear that this CoC and its enforcement will be authoritative and unfair by nature. If this is the concern, then let's focus on this concern in the first place. This CoC and its application in the Wikimedia technical spaces can only become a reality if the Wikimedia technical community embraces it and maintains it. If there is anything suspicious and unfair in this CoC, it will not be approved, it will not become a policy.--Qgil-WMF (talk) 07:25, 14 August 2015 (UTC)Reply

(((moved from "Enforcement in new draft")))@Fhocutt (WMF): Also, "Only permanent resolutions (such as bans) may be appealed" (proposal in the current draft) should be clarified. Is it intended that only permanent actions (e.g. lifetime ban) can be appealed, or any ban (possibly temporary)? This should also clarify how the reporter's appeal rights are (un)affected. Mattflaschen-WMF (talk) 01:31, 18 August 2015 (UTC)Reply
It's intended that only permanent actions can be appealed, but that the reporter's appeal/feedback rights are unaffected. I'm not terribly attached to this; I'd also be ok with something like "actions with effects that last for longer than three months" as well. The things I'd like to see balanced are not wasting the committee's time with specious appeals (such that the second step to any given action taken becomes "deal with the inevitable appeal") and allowing the flexibility to deal with genuine mistakes, particularly cases where the violator is using the CoC enforcement mechanisms as a way to further harassment of their target and the committee does not initially recognize that. --Fhocutt (WMF) (talk) 21:42, 18 August 2015 (UTC)Reply
Okay, I think three months is a good compromise. I've added that. Mattflaschen-WMF (talk) 20:44, 19 August 2015 (UTC)Reply
I still think this sounds odd: "Only resolutions (such as bans) that last 3 months or longer may be appealed by the reported offender." I think everybody should have the right to appeal, and the Committee should have the tools to avoid that this right is being abused in bad faith. For instance: "Resolutions may be appealed. For simple cases, the committee may decide to start the enforcement during the appealing process to avoid an artificial obstruction to the process. In more complex cases implying a ban of three months or more, the enforcement must wait until the appeal has been resolved."--Qgil-WMF (talk) 09:24, 20 August 2015 (UTC)Reply
It feels like something is missing in the current draft -- notifying or talking to the reported offender. If I'm reading this correctly, it looks like the Code of Conduct Committee reviews the report and makes a decision, and the only time the reported offender is mentioned is the line about being able to appeal a 3+ month long ban. Sometimes the evidence is entirely public and easy to corroborate, and sometimes it's not. When it's not, you probably want to talk to the person and see what they have to say about it. Maybe that's obvious and doesn't need to be in the policy? DannyH (WMF) (talk) 00:22, 21 August 2015 (UTC)Reply
I agree. However, it has to be done carefully. You don't want a scenario like:
  1. Someone feels they are being harassed, and reports it.
  2. The Committee contacts the alleged offender to get their side of the story, naming the reporter (who did not expect to be named).
  3. The Committee ends up taking only minor action (or none at all).
  4. The offender retaliates against the reporter for the report.
The jQuery CoC is ambiguous on this front. The Reporting page says, "All reports will be kept confidential. In some cases we may determine that a public statement will need to be made. If that's the case, the identities of all victims and reporters will remain confidential unless those individuals instruct us otherwise.", but the Enforcement Manual says, "The committee is empowered to act on the jQuery Foundation's behalf in contacting any individuals involved to get a more complete account of events." Contacting those individuals could itself be a breach of confidentiality. I added a sentence that tries to balance these concerns (needing the facts and avoiding revealing information against the reporter's wishes). Mattflaschen-WMF (talk) 01:26, 21 August 2015 (UTC)Reply
@Qgil-WMF: Even if enforcement begins during the appeal process, your proposal does not do anything to deter people from appealing every single case (no matter how minor the penalty). We have to consider ECT's time. Mattflaschen-WMF (talk) 01:19, 21 August 2015 (UTC)Reply
ECT's time is a factor I'm considering, trust me. I worry more about people complaining for having no right to appeal than about people trying to game the process. It is likely that the latter will put themselves in evidence with little community support. As for the former... they have a chance to drag more time and energy from all of us and, who knows, maybe even for a good reason because everybody can make mistakes, a committee too. (Note: I'd rather leave this discussion about appealing later, as proposed at #Next steps).--Qgil-WMF (talk) 08:47, 21 August 2015 (UTC)Reply
Matt -- I agree with the need to strike a balance between involving the reported offender and not violating the reporter's confidentiality. That's a good point, and I think your changes to that section make a lot of sense. DannyH (WMF) (talk) 21:10, 21 August 2015 (UTC)Reply
  • The wording is unclear as to who has the right to appeal -- who is the "you" in "if you have concerns"? Possibilities include anyone sanctioned in the case, any party to the case, anyone involved in any way at any stage, or anyone at all. Which is it to be? Rogol Domedonfors (talk) 20:52, 25 August 2015 (UTC)Reply
    I think it's "anyone at all". But this should not be used as a back-door appeal (by either offender or reporter). It should only be to notify them of opinions, e.g. "I think the committee has developed a pattern of being too lax; this case is an example". Also note that under the current draft, either offender or reporter can ask for reconsideration ("After being notified of the outcome, the reporter or alleged offender may raise objections to the resolution. These will be considered by the committee."), but only offender can appeal to the separate level, and only under restricted circumstances. Mattflaschen-WMF (talk) 03:25, 26 August 2015 (UTC)Reply
Then "you" is the wrong term, and should be in a separate sentence, reading "In addition, any member of the community may raise concerns about a case ...". The wording will still not explicitly forbid it being used as a second appeal. I also note that it is at least implicit, and perhaps needs to be explicit, that only a party sanctioned may appeal. Is it clear that we do not wish an aggrieved party to appeal on the grounds that their complaint has been dismissed or that sanctions are too light? Rogol Domedonfors (talk) 06:59, 26 August 2015 (UTC)Reply
I've attempted to clarify this, partly based on this feedback. Mattflaschen-WMF (talk) 00:09, 28 August 2015 (UTC)Reply
I would suggest that there needs to be the possibility of an appeal against a reprimand which is public and delivered in a way which forms part of a permanent record. Rogol Domedonfors (talk) 06:51, 26 August 2015 (UTC)Reply
  • The grounds of appeal are not limited in any way. That may be right, but it is a choice. Is there any appetite for restricting grounds of appeal, and if not,, itmight be helpful to spell out that an appeal may be made on any of the grounds that procedure was not correctly followed, that the decision was wrong on the factgs or that the penalty was disproportionate. Rogol Domedonfors (talk) 07:04, 26 August 2015 (UTC)Reply
    I don't see any need to restrict the grounds, but if someone does, they can suggest which grounds should not be allowed. I would guess pretty much everyone eligible to appeal will do so, but the 3-month minimum means minor penalties won't go through a time-wasting appeal process. Mattflaschen-WMF (talk) 00:13, 28 August 2015 (UTC)Reply

Membership of the committee and ECT's role

The CoC doesn't explain how the committee is formed and how are their members renewed. Following with Wp mirror's comment on Substantive v Procedural (thank you, I learned something), I'm not sure whether such information should be included in the CoC itself or in an attached document, but we need to define it.

Then, as ECT member I would like to understand what is the role the community wants us to have here. Especially when it comes to escalations, public statements and very ugly cases in general, I think it is useful to have a speaker paid for doing their job, saving some stress to those volunteering in the committee. However, I don't see the need of ECT to be involved in the regular work of the committee. If they would want our help, they could simply "escalate", as defined by the CoC. I think having a committee of equal members chosen through the same process by their own merits would give this team more credibility, independence, and strength.--Qgil-WMF (talk) 10:27, 19 August 2015 (UTC)Reply

Why is there a presumption that the appeal body must be the ECT? An elected group such as Stewards or Oversighters, positions of trust open to both unpaid volunteers and appointed WMF employees, would be in-line with putting the volunteer at the centre. -- (talk) 16:07, 19 August 2015 (UTC)Reply
Neither the Stewards nor the Oversighters necessarily have technical accumen, and this duty is not something they've volunteered to take on. One of the things to make sure of here is that the people doing the job are deeply knowledgeable about technical spaces and discussions and thus capable of recognising tech-specific forms of harassment for what they are.
Quim: it would probably be included in the document. The jQuery CoC, for example (which got passed today!) sets out this sort of thing very specifically, which I think works well. Ironholds (talk) 19:05, 19 August 2015 (UTC)Reply
@Qgil-WMF: If I understand right, you're saying ECT does not necessarily need a reserved spot on the first-line committee, as the document currently lays out. Is that correct? I agree we will have to specify how the committee is formed, either in this document or a sub-document. I have added a note to this effect. Mattflaschen-WMF (talk) 21:08, 19 August 2015 (UTC)Reply
Agree that we will need to specify more about the committee. I think that ECT should have a reserved spot for several reasons: ECT's job is community expertise, and there is substantial gathered knowledge and resources on that team, including easy consultation with other community teams; all-volunteer committees are vulnerable to gaps in institutional memory (and as tough a time as we have with our technical documentation and continuity of project knowledge, we need a way to ensure this doesn't happen here!); and CoC response and enforcement is a difficult and sometimes draining position, and having at least one person whose job this is part of should mean the committee has more support and structure. --Fhocutt (WMF) (talk) 23:41, 19 August 2015 (UTC)Reply
for what is worth ECT also wants to be sure that the role this CoC gives to us is the role the community wants us to have. Any big social problem in the tech community has been and will continue to be a problem that we ECT need to help solving, with or without CoC. How exactly should we involved, you tell us. A pragmatic approach would be to define the process to form the committee (leaving the hypothetical ECT seat out for now) and to know the opinion of those interested to be part of it. At the end, these are the people that need to work comfortably with us.--Qgil-WMF (talk) 09:35, 20 August 2015 (UTC)Reply
Note that I did not mention ECT's participation in the committee, only the nature of the appeal process. I would think it natural that some ECT members would volunteer to be part of the proposed committee, and if they end up going through a community election I doubt they would be likely to fail to get in considering the case. See below for my follow up on the choice of appeal body. -- (talk) 09:56, 20 August 2015 (UTC)Reply
@Mattflaschen-WMF: Yes, I think that having an appointed and immutable ECT member in that first-line would set a default point of gravity that would influence how the committee operates. Also, this permanent seat would not help in cases where pro/anti WMF attitudes are part of the problem (not an unusual situation). Having all the members of this first-line appointed equally through a community-driven process would make this first-line stronger in many ways. Whenever the committee members feel that they need ECT's help (or help from anyone else at the WMF), we would be there to support them and to call anyone else that needs to be involved. No less, no more.--Qgil-WMF (talk) 10:02, 20 August 2015 (UTC)Reply

Let me make a specific proposal for the committee membership, to prompt some discussion.

  1. Once we form consensus to enact the code of conduct, we (the people interested enough in this issue to be discussing it here) discuss amongst ourselves and nominate a slate of five trusted community members. I also think one seat should always be reserved for a member of the ECT, to help provide capacity and institutional memory, but if Quim doesn't think that's a good idea I don't think I can force him :)
  2. That slate gets ratified by the same process we used to enact the code.
  3. People serve on the committee as long as they want (unless they're forcibly removed). When a member leaves, the other members simply use their judgment to replace the departing member with another trusted community member. The code will explicitly mandate that they focus on keeping a wide variety of perspectives on the committee—which primarily means gender and geographic diversity—and that at the very most 2 (3?) members can be employed by the WMF.
  4. At any point, the community could use the same process used to enact the code to form a consensus to replace a committee member (this would essentially be a constructive vote of no confidence, requiring consensus on a replacement to remove someone). This means that as long as the committee does a good job (which, given that it's composed of trusted community members, I expect it almost always will), the rest of the community doesn't have to spend any time worrying about it.

Unanswered questions:

  • Should there be other mechanisms for removing a committee member? I imagine this would come up very rarely, so a community-wide discussion would be sufficient, but perhaps 4 members of the committee could vote to remove the fifth, or the ECT could step in to remove people.
  • What kind of process do we use for the community-wide discussions necessary to enact the code, select the first set of members, and remove them if needed?

My main objective here is a lightweight process. There's no need to institute the equivalent of enwiki's ArbCom elections when we will likely have far less conflict to deal with here. Also, I move that the committee be called the Conduct Committee, so we can call it ConCom for short :)—Neil P. Quinn (talk) 06:43, 20 August 2015 (UTC)Reply

The question of why the appeal body must be the ECT is actually still floating here. There have been no solid reasons put forward as to why no alternatives are possible, or more desirable. -- (talk) 09:33, 20 August 2015 (UTC)Reply

For what is worth, I don't think ECT or whoever is at the "second-line" should be the automatic appeal body. The first-line is in charge of receiving reports, and also appeals. If the committee or the person(s) appealing think that the first-line is no good fit in a specific case, then they can go to the second-line.--Qgil-WMF (talk) 10:10, 20 August 2015 (UTC)Reply
I have difficulty comparing your reply against what is currently written in the draft document. It might be helpful to stick to the same use of words. -- (talk) 10:16, 20 August 2015 (UTC)Reply
Sure, this is because we are all discussing and drafting, still with different opinions. We will get there. What is your take?--Qgil-WMF (talk) 09:28, 21 August 2015 (UTC)Reply
To use the same terminology as the document, in a "case where the committee disagrees with the above initial response", the Committee already has the right to revisit a decision by a single Committee member. I don't think there should be a general right to appeal to the Committee again. However, I will broaden the 'objection' stage to reflect that the alleged offender can also object. Mattflaschen-WMF (talk) 01:38, 21 August 2015 (UTC)Reply
Ironholds answered the question about why ECT above ("Neither the Stewards nor the Oversighters necessarily have technical accumen, and this duty is not something they've volunteered to take on. [...]"). Mattflaschen-WMF (talk) 01:38, 21 August 2015 (UTC)Reply
I strongly disagree with this proposal for membership. My primary concern is very simple; you are assuming that people confident and certain enough to publicly comment in this discussion are adequately representative of the community, its concerns, and the nuances around problematic behaviour - and the members of the community who have left due to the inadequacy of existing codes of conduct, where they exist, and/or people who have not joined yet. That does not seem in any way a safe assumption to make. If we limit the pool of candidates and how they are evaluated to the people on this talk page we end up with a pool of people acceptable to the existing community - the existing community that in some elements is still discussing whether a code of conduct is even necessary much less what it looks like. This repeated idea that those who turn up represent the full gamut of opinions, or the opinions worth listening to, is problematic, and the assumption that the pool of people discussing it represents the pool of people who care about it is inaccurate (I know several people who have strong opinions here but are not comfortable commenting publicly) and honestly rather offensive. People not commenting do not "not care", they are in some cases simply not comfortable, and that should be something that causes us to re-evaluate how we hold these discussions, not dismiss factoring them in. Ironholds (talk) 14:12, 20 August 2015 (UTC)Reply
What Ironholds said. I've personally heard from a fair handful of people who've been intentionally avoiding these discussions for reasons that include discomfort and lack of trust that their opinions will be thought to matter. --Fhocutt (WMF) (talk) 03:27, 21 August 2015 (UTC)Reply
Ironholds, Fhocutt (WMF): My thought behind that idea (and maybe this is naive) is that it's not incredibly hard to find people who would do a good job in the role of a ConCom member. It's certainly not something I would entrust to anybody off the street, and it certainly requires sensitivity and empathy for the type of people we want to include with this code. But I think there are enough good options that even an unrepresentative discussion would come up with pretty good choices. For example, I think that both of you would do a great job—and despite the limitations of this discussion, you're already here.
That said, it's still a fair criticism, and it seems like the provision for private comments in Matt's plan would give people who are not comfortable discussing this in public a chance to make their voice heard. The only problem is that private comments means the ultimate decision on committee membership has to be entrusted to a third party—in this case, ECT. I don't think that would be a bad outcome, since I trust them and think they would make good choices—and since they are ultimately accountable to the community through the Board of Trustees. However, the Wikimedia ideal is fairly direct community semi-democracy and I would be disappointed if we couldn't find a way to make that work in this case.—Neil P. Quinn (talk) 06:05, 22 August 2015 (UTC)Reply
I just reread what I wrote and saw where I put "the people interested enough in this issue to be discussing it here". You're right—that's misguided. There are many reasons why someone might be interested yet not be commenting here.—Neil P. Quinn (talk) 06:17, 22 August 2015 (UTC)Reply
And actually, to modify my comment again, I've actually thought of some ways to make anonymous comments compatible with discussion-based consensus decision making (see #Next steps below). I also don't think giving ECT the ultimate authority to approve the people initially selected by the community would be bad—this isn't very different from the CA team having ultimate control of the global ban process, which I don't find concerning.—Neil P. Quinn (talk) 06:52, 22 August 2015 (UTC)Reply
@Neil P. Quinn: I would be worried if the CA team claim to control global bans, as this has always been a community driven process. You may be conflating Global Bans with WMF Global Bans. -- (talk) 08:02, 22 August 2015 (UTC)Reply
@: You probably know a lot more than I do about the global ban process; as I understand it, it's primarily community driven, but the CA team does intervene in extraordinary circumstances. I like the idea of a similar role for the ECT team here. For example, I wouldn't make them directly responsible for selecting Conduct Committee members, but I might well give them veto power over the community's picks, to be used only when strictly necessary.—Neil P. Quinn (talk) 01:14, 23 August 2015 (UTC)Reply
I disagree with the proposal that 4 members could vote to oust the fifth. That would meant they could do that to prevent that 5th member from escalating (which the current draft explicitly says a minority could) the issue. I also think the idea of the community ousting members at any time risks politicizing it (think 'recall election'). I need more time to think about the rest. Mattflaschen-WMF (talk) 00:33, 21 August 2015 (UTC)Reply
Yeah, on reflection, recall discussions do seem like a bad idea.—Neil P. Quinn (talk) 06:05, 22 August 2015 (UTC)Reply
I also object to the idea of limiting the number of WMF employees to 2/5. I do think there should be non-WMF members, but I don't think you've made a case for why most (at least 3/5) members should be required to be non-WMF at all times. Mattflaschen-WMF (talk) 00:38, 21 August 2015 (UTC)Reply
Agree on both counts. Recall elections get nasty easily, and if we're instituting quotas I don't think that WMF/non-WMF should be the only characteristic we filter on. Let's not open that up. --Fhocutt (WMF) (talk) 03:27, 21 August 2015 (UTC)Reply
  • I disagree on the formation of any such committee. Existing Wikimedia projects processes should be used, such as sysops and stewards. --Nemo 06:40, 21 August 2015 (UTC)Reply
    Existing Wikimedia projects processes have failed pretty thoroughly in this regard. As you yourself noted earlier, wikis which have behavioral policies are even more toxic then the Wikimedia tech space. Replicating that failure one more time does not seem like a worthwhile endeavor. --Tgr (WMF) (talk) 07:07, 21 August 2015 (UTC)Reply
    I'm not aware of said processes being attempted for the cases presented and in the only cases I remember of blocks etc. in the MediaWiki community our processes worked perfectly. --Nemo 07:30, 21 August 2015 (UTC)Reply
    Existing Wikimedia projects processes are designed to be effective on specific tools, and they don't attempt to be cross-tools. There are no stewards specific to "Wikimedia technical spaces", and it will be interesting to ask them what they think about enforcing a CoC in places like Phabricator, Gerrit, etc. I do think some stewards could be very capable committee members, though. As for sysops, only mediawiki.org has close to 200 users with that role. We cannot expect such group to deal efficiently with delicate situations privately. So no, I don't think the existing Wikimedia processes you are referring to have been effective to address the problems this CoC aims to address, and I don't think they are designed for this task.--Qgil-WMF (talk) 09:46, 21 August 2015 (UTC)Reply
@Tgr (WMF): "Existing Wikimedia projects processes have failed pretty thoroughly..." is this really what the WMF believes and is trying to fix here? I have seen some anecdotes and ridiculous amounts of over dramatic grandstanding, but a lack of measurable evidence to show that if today we abandon community driven collegiate processes and force an untested WMF procedure from the outside, that this would result in attractive and healthier Wikimedia project communities, or be more effective for handling harassment. -- (talk) 08:12, 22 August 2015 (UTC)Reply
I am not sure what gave you the idea that I am expressing the WMF's beliefs here. As for my beliefs, I do think the Wikimedia movement has a chronic problem with conduct, both with occasional harassment and with a constant background noise of hostility and microaggressions; and existing processes could not fix it, neither in the tech spaces nor in the content projects. This proposal borrows elements from other tech organizations, who had some success with them. In general, replacing a method that has failed for us with one that worked for others seems common sense to me; and implementing it and then comparing two processes that we actually tried seems more reasonable than trying to somehow measure theoretically how well something will work before we implement it. Maybe it will fail, and then we can scrap it; it doesn't seem like a particularly risky experiment. --Tgr (WMF) (talk) 08:21, 23 August 2015 (UTC)Reply
A big +1 to that. It's worth considering, when people summarise WMF employees' opinions as "the WMF's opinion", that one of the components of being an employee at the WMF is that people are usually (as part of their job) spending more time in technical venues, if they're in technical roles, than is the norm. When the most regular contributors to and users of technical discussions are saying "we have a problem", "they're being paid to say they have a problem" is not the read to have there. Ironholds (talk) 14:39, 24 August 2015 (UTC)Reply

It looks like there is more agreement than disagreement on the idea of giving the Committee the possibility to escalate complex issues to the Engineering Community team. I have added this bit in the draft. Whether ECT has a member in the committee or not is still open, but this can be discussed in the context of how the committee is formed.--Qgil-WMF (talk) 09:28, 2 September 2015 (UTC)Reply

The draft mixes two classes of unacceptable behavior that must be handled differently.

  • Unacceptable behavior
    • Class 1 - violence, threats of violence, intimidation, stalking, inappropriate touching;
    • Class 2 - inappropriate communication, sustained disruption, offensive comments, unwanted photography, outing.

In many jurisdictions, class 1 behavior is a crime, and must be handled by the authorities. Not to inform the authorities would make one an accessory. Only class 2 behavior is properly handled by a WMF Conduct Committee.

  • Procedure
    • Class 1 - inform law enforcement, inform WMF Legal Counsel;
    • Class 2 - report to technical space moderator, file case with WMF Conduct Committee.
  • Penalty
    • Class 1 - WMF Legal Counsel recommends: ban, suspension without pay pending investigation, termination;
    • Class 2 - WMF Conduct committee issues: warning, timeout, suspension, ban.

Wp mirror (talk) 18:10, 19 August 2015 (UTC)Reply

'Class 1' behavior can be handled by the authorities, but it is up to the victim to decide whether to report it to the authorities or not. The WMF and the conduct committee must not report what happened to the authorities, unless explicitly asked to do so by the victim.
There is no need to distinguish the two classes for this code of conduct; our (civil) response to a criminal act is seperate from a potential criminal investigation. Valhallasw (talk) 19:00, 19 August 2015 (UTC)Reply
Agreed with Valhallasw. I'd also say that your argument very nicely demonstrates one of the problems with this kind of distinction; some of the acts you have categorised as "class 1" are not crimes in some places, but are in others, or have totally different standards in different jurisdictions. I absolutely agree that in the case of situations where we believe a crime has been committed, counsel should be notified. I disagree that the need to provide that notification precludes civil enforcement in addition (and would argue there are going to be cases where filing legal reports, straight off the bat, is not in the interests of the victim). This seems like an area where counsel should be consulted if there are legitimately worries that not reporting constitutes making the Wikimedia Foundation (or the committee, or anyone else involved in the reporting chain) an accessory. Ironholds (talk) 19:41, 19 August 2015 (UTC)Reply
Nothing in this document discourages reporting to the authorities (and that is exactly as it should be). However, I think it is still worth reporting to the conduct committee as well (remember, the authorities may not be able to take legal action, but that doesn't mean we want that person in our community). There is also a clear statement, "If you believe anyone is in physical danger, please notify appropriate law enforcement first and email emergency wikimedia.org", which is an additional way the WMF could be informed about such serious matters. Mattflaschen-WMF (talk) 21:19, 19 August 2015 (UTC)Reply

The draft states that one of the penalties is "public reprimand". This does create a need to distinguish the two classes. For a class 1 case, if the Conduct Committee issues a public reprimand, the following charges could be filed against the members:

  • Libel. If the Conduct Committee got the facts wrong (e.g. deceived by a vengeful X), the announcement could ruin the reputation of the reprimanded party, who would in turn seek remedy.
  • Accessory. If the Conduct Committee got the facts right, but did not inform the authorities, the announcement would alert the authorities, who would in turn want to question the members (what do you know, how long have you known about it, why did you not notify the authorities).

For a class 2 case, these concerns do not arise.

From time-to-time, the Conduct Committee will get the facts wrong. Why? No rules of evidence. The committee can not insist upon sworn testimony. The committee will sometimes have to rely upon hear-say. If we wish to claim that no distinction exists between the two classes, then perhaps we should choose penalties less likely to cause blowback. Wp mirror (talk) 08:48, 20 August 2015 (UTC)Reply

Again, you are introducing a distinction that is not found in the actual draft, and these questions are best handled by actual lawyers. Ironholds (talk) 14:08, 20 August 2015 (UTC)Reply

If the proposed Conduct Committee issues public reprimands, then we should expect that, sooner or later, actual lawyers will get involved. Shall we ask WMF Legal Counsel to look over the draft before this happens? Wp mirror (talk) 11:05, 21 August 2015 (UTC)Reply

WMF Legal has reviewed this document for the Foundation. It is required that any reports involving WMF employees in any way be reported to WMF HR and potentially WMF Legal. VBaranetsky (WMF) (talk) 01:26, 28 August 2015 (UTC)Reply

I am puzzled about what "It is required" can mean here. It is in the passive: by whom is it required? Do they mean that they, WMF Legal, require the Committee to pass on all such reports and if so by what right do they claim to require the Committee, which might well consist entirely of non-employees, to do something? Rogol Domedonfors (talk) 06:47, 28 August 2015 (UTC)Reply
@VBaranetsky (WMF): Please rephrase in order to describe who is supposed to report and how those supposed to report can actually contact "WMF's HR". Also please explain what the abbreviation "WMF" stands for as it's the first time it's mentioned in the document. Thanks in advance! --AKlapper (WMF) (talk) 12:42, 28 August 2015 (UTC)Reply
@VBaranetsky (WMF): This intervention is not entirely hepful. Firstly, WMF Legal is not in a position to simply dictate arbitrary unexplained and obscure additions to this code, which is in the middle of a community discussion aiming at consensus. Secondly, the addition is surprisingly vague, as pointed out just above. Thirdly, it is incompatible with at least two other provisions of the Code, which states in other sections that All reports will be kept confidential and mandates of the Committee that before revealing any confidential information from the report, they must get consent from the reporter -- are you now attempting to order the community to alter those provisions? Fourthly, it is potentially incompatible with legal requirements in other jurisdictions, where these events may take place, which mandate certain privacy procedures for the protection of complainants and suspects in certain kinds of crime. Fifthly, it is unenforceable, as WMF has no authority over non-employee actions. This requirement needs to be restricted to areas in which the WMF HR and Legal departments have authority, which is over their own employees. Simply make it an internal rule that WMF employees and contractors are required to report to line management if they are involved in a complaint under this code. That is proportionate and enforceable. Rogol Domedonfors (talk) 19:29, 28 August 2015 (UTC)Reply
This can certainly be edited as needed. My understanding is that under California (USA) law, organizations are liable in various ways when their employees are harassed, whether by employees or third parties (like customers or, in our case, volunteers). I am personally ok with including whatever reasonable provisions are needed to make it easy to comply with employment laws or reduce the probability that the WMF could be sued for allowing a hostile work environment. I do share concerns about confidentiality and not forcing victims to report, however. That said, I suggest waiting to hash this out for when we are focusing on reporting/governance. --Fhocutt (WMF) (talk) 21:57, 28 August 2015 (UTC)Reply
Unfortunately this cannot "be edited as needed" since we have no idea what it is trying to achieve, why it is necessary or what alternative forms of wording or processes would meet whatever it is that WMF Legal "requires". It is probably counter-productive to guess without having a clear statement from them of their rationale. I have proposed a alternative form of words to meet what I guessed to be that rationale, and which preserves the independence of the Committee. It is disappointing that VBaranetsky (WMF) has edited the page again, albeit in a minor way, without addressing those points. Waiting is certainly an option but since there is currently a fundamental contradiction in the Code as written (between confidentiality and requirement to report to WMF), and confidentiality is a rather important principle to get settled, it seems to me important to resolve it sooner rather than later. I may say that my inclination is to accord more weight to confidentiality of the proceedings than to the currently unexplained demands of WMF Legal. If it is indeed the case that this requirement stems from their legal duty to protect their staff, then it would seem that they have no confidence in the ability of this Code to do so, which is rather sad news for the rest of us who are not protected by the WMF in the same way. Rogol Domedonfors (talk) 06:35, 29 August 2015 (UTC)Reply
"Simply make it an internal rule that WMF employees and contractors are required to report to line management if they are involved in a complaint under this code." In my opinion, this (without a check) is not enforceable, even if the alleged offender is acting in good faith. First, confidential information from the report is not given to the alleged offender unless the reporter agrees. If the reporter does not agree, the alleged offender may never know about the complaint (though this may impede investigating it, as the draft notes), but my understanding is that HR still needs to know. Then, of course, what if the alleged offender is not acting in good faith? What happens if they don't report the Committee proceedings to HR? Who will be in a position to verify this is done? Mattflaschen-WMF (talk) 20:16, 2 September 2015 (UTC)Reply
  • A couple of scenarios.
  • A participant in a hackathon at Wikimania 2016 is sexually assaulted by another participant. A third participant is a witness and is able to give clear evidence that the incident was an assault. The victim is embarrassed to describe the nature of the assault but does so in the assurance that the description will remain private.
Now we discover that the witness is a WMF contractor. Under the required rule, the Committee is required to report to the WMF Legal department. They cannot help in any way, and the victim is further distressed by having the details of their assault pointlessly communicated to unknown parties in another country.
  • A participant in a hackathon at Wikimania 2016 is sexually assaulted by another participant. A third participant is a witness and is able to give clear evidence that the incident was an assault. The victim is embarrassed to describe the nature of the assault but does so in the assurance that the description will remain private.
Now we discover that the perpetrator is a WMF contractor. Under the required rule, the Committee is required to report to the WMF Legal department. The victim will now be faced with the details of their complaint being given to a team of lawyers at a multi-million-dollar organisation who have been given this information in order to protect the interests of the perpetrator.

Do we believe that either of these scenarios will encourage victims to step forward and complain formally to the Committee? Or do we believe that the requirement in question will have a chilling effect on the reporting of serious breaches of the Code? What benefit is there to counterbalance the harm done to the community? Rogol Domedonfors (talk) 09:06, 29 August 2015 (UTC)Reply

@VBaranetsky (WMF): : Since your edit does not address where reports concerning people affiliated to other organisations should go, it's been reverted. --Krenair (talkcontribs) 20:39, 29 August 2015 (UTC)Reply
That question is already addressed: the answer is nowhere, since it is twice stated that all reports will be confidential. It is only the WMF that has asserted, without explanation, the requirement to breach that confidentiality and the right to order the Committee to do so. It is not at all clear what will happen should a member of the Committee decline to pass on this confidential information because it violates the privacy laws in the country where an incident took place, or because it violates their personal or professional code of ethics, or because they resist being told what to do by an outside party without authority over them. Rogol Domedonfors (talk) 21:47, 29 August 2015 (UTC)Reply
Rogol: I've replied below. I share some of your concerns, but I also don't want, say, the only information that HR has to work from to come from the offender, as might happen with only having an internal policy that mandates involved parties report. What I mean to say in suggesting we handle it later is that it falls under enforcement/reporting, not under principles/unacceptable behavior/re-use and so it doesn't block Qgil-WMF's suggestion to approve this by stages. --Fhocutt (WMF) (talk) 18:38, 1 September 2015 (UTC)Reply

Double jeopardy

I think it's implicit in some of the preceding discussion, but I think it worth stating explicitly. By double jeopardy I mean the issue of whether sanctions under the Code should be applied in cases where other proceedings and penalties, possibly legal, are also applicable to the conduct in question. For example, some forms of harassment are legal crimes. Some people take the view that one should not be proceeded against or punished twice for the same offence. I personally take the view that the Code is here to protect the community and that the additional privileges and benefits of being a member of this community justify the imposition in principle at least of imposition of additional penalties for conduct that disrupts or damages this communty. I further think that this is the consensus view of this discussion. It might be worth discussing exactly what that means in the case og, for example, criminal proceedings. For example: "If the conduct complained of is or becomes the subject of formal legal proceedings, then the Committee will suspend its final decision until the results of those proceedings is known, but may make an interim ruling for the protection of the community. At the conclusion of those proceedings, the Committee will decide whether to continue with its own decision or to rule that further action is unnecessary.". Rogol Domedonfors (talk) 08:54, 29 August 2015 (UTC)Reply

I agree that we do not want to exempt someone from a Code of Conduct penalty due to a separate legal proceeding. I'm not sure suspending decision during a separate process is necessary, but it might work if the Committee still has its full powers for the "interim ruling". I also disagree with deciding "further action is unnecessary" on the basis that a separate body (e.g. a court of law) has taken action. There are cases where the court could impose a severe penalty but we would still certainly want to take action. But even if the court takes relatively minor action, we may still choose to take action since the Code of Conduct is a different set of rules with different consequences. If by "rule that further action is unnecessary", you mean "decide that the interim decision is sufficient, and simply making the interim decision final should conclude the case" that might be appropriate in some cases. However, the penalties imposed by other bodies (if any) should be irrelevant to the Committee's decisions. Mattflaschen-WMF (talk) 23:30, 31 August 2015 (UTC)Reply
Most legal systems are terrible at handling harassment. This policy is about what behavior we want to accept and support in our community, and I'm fine with other communities and policy systems having their own say about what behaviors they want to accept. Legal proceedings are often long enough and vary enough by location that I don't think there's a reason to mention them explicitly. --Fhocutt (WMF) (talk) 17:59, 1 September 2015 (UTC)Reply

Proposed additions

i suggest the fllowing addition to Principles:

This code is published by Resolution of the WMF Board of Trustees, which acts as the authority for its implementation on behalf of the community.

and under Enforcement:

The WMF Board authorises a Code of Conduct Committee to manage reported infractions of the Code! Constituted as follows ...
The WMF Board will hear appeals from the decisions of the Committee, either directly or through a committee formed ad hoc.

A draft resolution for the Board would be needed. Rogol Domedonfors (talk)

Except it's not, it hasn't, and it doesn't. If there's a desire to have the board as the appellate body or authorising body the board should say "Yes, we're comfortable doing that". Ironholds (talk) 18:07, 25 August 2015 (UTC)Reply
You will need to explain the various values of "it" in your first sentence: the tricolon loses in rhetorical power by the unclarity of the referents. The proposal is however exactly as you say in your second sentence. To achieve the situation outlined in this proposal would indeed involve putting the proposed Code formally to the Board and having them formally adopt it by Resolution. Of course a prelininary informal approach and agreement in principle would be needed first, so to that extent we seem to be in agreement. Rogol Domedonfors (talk) 20:38, 25 August 2015 (UTC)Reply
It's ordered by your original order. If we're in agreement that informal approach/agreement in principle is needed first then somebody should be doing that; we should not be adding language which presumes acceptance and makes the entire draft ride on board agreement. Let's draft without presumption; if the board later decides to adopt it they can pass a resolution that contains whatever modifications are necessary to confirm the board is doing it. Ironholds (talk) 20:51, 25 August 2015 (UTC)Reply

Its a community initiative not something being imposed from above. I don't think the board really has any place here. Bawolff (talk) 20:55, 25 August 2015 (UTC)Reply

This doesn't belong to Principles but to Committee. Rogol Domedonfors, if you are in a rush willing to include the Board in this initiative, feel free to contact them and ask their opinions. I still think it is better to wait to have a more consolidated draft, though.--Qgil-WMF (talk) 21:01, 25 August 2015 (UTC)Reply
The rush is being imposed by those who do not wish to proceed with a draft along these lines without asking the Board in conjunction with those who want the draft in order to consult the Board. So I have anyway, at meta:Wikimedia_Foundation_Board_noticeboard. Rogol Domedonfors (talk) 21:12, 25 August 2015 (UTC)Reply
This does not require a Board resolution, and I disagree with asking for one. The Terms of Use (which is Board-approved) says, "The Wikimedia community and its members may also take action when so allowed by the community or Foundation policies applicable to the specific Project edition, including but not limited to warning, investigating, blocking, or banning users who violate those policies. You agree to comply with the final decisions of dispute resolution bodies that are established by the community for the specific Project editions (such as arbitration committees); these decisions may include sanctions as set out by the policy of the specific Project edition." Mattflaschen-WMF (talk) 03:13, 26 August 2015 (UTC)Reply
Then you will be pleased to see that I did not ask for one. As discussed here, I asked for informal views about how the Board viewed the proposal that thry should be the authority behind this code. If they wish to do so, they would eventually enact it by formal Resolution, but that is a long way off. If they indicate their views informally at this stage, it will make drafting easier. Rogol Domedonfors (talk) 07:08, 26 August 2015 (UTC)Reply

It makes me feel sad to see that Board involvement is asked for. This is exactly the place where I thing the affected community should deliberate, find a consensus, agree on wording and implement such a policy. A code of conduct only works if it is backed up by those who have to follow it. A top-down-manner is not the best idea to create this ownership feeling. Lyzzy (talk) 15:47, 1 September 2015 (UTC)Reply

Being Lyzzy Vice Chair of the Board of Trustees, I think we can take her comments as a good indication that we should keep the Board away from this draft. If they want to have a role here, they know where to find us. :) Thank you Lyzzy for your feedback.--Qgil-WMF (talk) 08:06, 2 September 2015 (UTC)Reply

Protection of Conduct Committee members

A volunteer serving on the Conduct Committee must consider the possibility that an aggrieved party will sue for libel. To attract qualified volunteers to serve on this committee, one should consider how best to protect them. The main methods are:

  • D&O Liability Insurance. Many not-for-profit organizations purchase a D&O policy that extends coverage to chapter members and volunteers. It is not clear if the WMF has such coverage. Nothing relevant appears when one searches wikimediafoundation.org.
  • Limited disclosure. To take away the grounds for a libel suit, many not-for-profit organizations follow RONR 10th ed, Chap XX Disciplinary Procedures, §61: "If (after trial) a member is expelled, the society has the right to disclose the fact that he is no longer a member—circulating it only to the extent required for the protection of the society or, possibly, of other organizations. Neither the society nor any of its members have the right to make public the charge of which an expelled member has been found guilty, or to reveal any other details connected with the case. To make any of the facts public may constitute libel."

For the protection of Conduct Committee members, please:

  • procure D&O liability insurance, and post the terms on your site;
  • specify, in the procedural law for the Conduct Committee, how to report the outcome of cases; and
  • reconsider the "public reprimand" penalty clause stated in the draft.

Wp mirror (talk) 07:12, 20 August 2015 (UTC)Reply

The Wikimedia projects are not organised as a society of members. Again, these concerns are best surfaced to and handled by counsel, not pseudorandom members of our technical community (I include me in that). Ironholds (talk) 14:07, 20 August 2015 (UTC)Reply
What open source code of conduct committee offers D&O insurance to all members (both staff and volunteer)? Unless you can show a significant number of them do, I don't think it's reasonable to pursue this. I also can not think of any cases where someone has successfully sued an open source project or related organization for libel after being banned from participating or reprimanded. Can you? Mattflaschen-WMF (talk) 00:43, 21 August 2015 (UTC)Reply

I have in mind the following libel cases:

The last three are sometimes called cyberlibel cases. Consider the following:

  • If the proposed Conduct Committee got its facts wrong and publicly reprimanded someone for a crime he did not commit, Zenger (1735) would not apply, and WMF General Counsel would seek another defense.
  • If WMF General Counsel disavows the Conduct Committee, then Cubby (1991) might be a defense. But then the Conduct Committee would not have authority to enforce the Code of Conduct.
  • If the WMF Board of Trustees acknowledges the Conduct Committee (e.g. if the Conduct Committee reports to the Board of Trustees, like say the Affiliations Committee does), then Stratton (1995) might apply and WMF would be liable.

As to D&O Liability Insurance:

  • members of the WMF Board of Trustees are indemnified (see Bylaws Article VIII - INDEMNIFICATION);
  • the WMF Chief of Finance and Administration is responsible for purchasing the coverage (see seventh bullet under ADMINISTRATION).

The need for coverage is real.

  • If coverage is not provided, qualified volunteers might be hard to find, and even WMF employees would be prudent to hesitate.
  • If WMF employees are covered, but not volunteers, then committee membership should be restricted to employees.
  • If WMF General Counsel needs to be able to disavow the Conduct Committee, then WMF employees must be excluded from the committee.

This is why potential committee members need to see a copy of the policy. The WMF Chief of Finance and Administration is only one e-mail away. Shall we ask him? Wp mirror (talk) 03:33, 21 August 2015 (UTC)Reply

I asked some specific questions, with the goal of comparing to similar situations (open source or at least similarly-minded organizations with code of conduct enforcement). None of the libel cases you mentioned show that a good-faith ban or reprimand can lead to a successful libel suit. I am not convinced libel suits are a real problem other communities have faced in connection with their codes of conduct.
Similarly, you have not shown that D&O insurance is something other similar committees have had. It's not always affordable or realistic to buy more and more insurance, and it certainly doesn't seem to be legally required (or you would cite an example of a community that had bought it). Mattflaschen-WMF (talk) 04:04, 21 August 2015 (UTC)Reply

Example: The w:Association for Computing Machinery is a publisher, has chapters and special interest groups that also publish, hosts public speaking events, and has on-line communities. The ACM also has a code of ethics, breach of which may result in termination of membership (see section 4.2). When I was a member, the ACM had a D&O Liability Insurance policy that extended coverage to all its chapter members and volunteers worldwide. As an officer of one its chapters, I had a copy of the policy. When the conduct of one of our members created a legal liability, we initiated a disciplinary process. We followed RONR 10th ed., as stated in our Bylaws. Charges were drafted, the board went into executive session, and we do not disclose the facts of the case.

Following RONR plus D&O coverage is why libel is not a problem for the ACM. The ACM is not unique in this respect. Indeed, the formula is widely used in the not-for-profit world. Public reprimands for breach of professional ethics are issued by governments and governing boards (e.g. legal, medical). Non-profits, however, quietly dismiss an unethical member.

D&O premiums are low. Cost is not the issue. Nor is having a list of court cases. Getting people with executive experience to serve is the issue. This is why WMF trustees are indemnified. They would not serve otherwise. Wp mirror (talk) 06:52, 21 August 2015 (UTC)Reply

We should definitely check this with WMF Legal before it takes effect. There may be other approaches to dealing with any potential liability arising from participation on this committee. --Fhocutt (WMF) (talk) 23:10, 21 August 2015 (UTC)Reply
The ACM Is only loosely comparable. It's not an open source organization, but a publisher. Unlike us, they also have a clearly delineated set of 'members'. Moreover, their bylaws say, "The Association shall have power to purchase and maintain insurance or to self-insure on behalf of any person who is or was a director, officer, employee or agent of the Association". You'll note that it says nothing about having the power to purchase insurance for all ACM-related volunteers. In fact, it doesn't even authorize them to purchase insurance for all paying members. That suggests that if they ever covered all volunteers or all members, they no longer do. No one has suggested that 'executive experience' is required to serve on this committee. I am confident we can get people to serve on this committee, including volunteers. Mattflaschen-WMF (talk) 21:38, 23 August 2015 (UTC)Reply
In order to resolve some of these issues, it will be necessary to have a decision from WMF Legal. Is the WMF willing to take steps to give legal protection to members of the Committee (and the appeal body)? If so, this discussion can proceed on the basis that WMF Legal will know what to do and will do it effectively: it may be that they will wish to advise on procedural matters to make that protection most effective and efficient. If not, then the discussion will need to proceed on the basis that individual Committee members and indeed individual community members are legally at risk in these proceedings, and again WMF Legal may be able to help by advising on how to manage those risks (although I do not know whether their remit extends to advising non-WMF people or entities, of course). Personally speaking I would not advise anyone to undertake this is sort activity without adequate legal protection, but other community members may have a different appetite for risk (or better legal insurance). @VBaranetsky (WMF): as one who has been taking an interest in the legal relationship between the WMF on the one hand and this Code and its Committee on the other, can you please help here? Rogol Domedonfors (talk) 20:33, 28 August 2015 (UTC)Reply

Is the CoC introducing any novelties here? User bans and other types of measures already exist, and with them the risk of upsetting someone to the level of retaliation. If anything, the CoC provides a bit more structure and a bit more protection to the ones taking action. The existence of a Committee allows admins and maintainers to not be the main actors when i.e. banning someone. The Committee can delegate ugly cases to professional employees at ECT. WMF Legal does protect WMF employees legally, but so it does with Wikimedia volunteers as well. In the case of someone receiving some retaliation for enforcing a CoC resolution, I think the situation with CoC will be better than the current situation without it.--Qgil-WMF (talk) 10:09, 2 September 2015 (UTC)Reply

"Public reprimands" (one of the penalties Wp mirror took issue with) also already exist in Wikimedia communities. Specifically, this is what happens when an admin posts on someone's talk page something like "You have done Xyz which is a violation of our policies. Please do not do so again, or you may be blocked." Mattflaschen-WMF (talk) 20:40, 2 September 2015 (UTC)Reply

Proposing Intro + Principles + Unacceptable behavior

As proposed in #Next steps, let's focus on agreeing Intro + Principles + Unacceptable behavior up to a point where we can make a wider call to wikitech-l and whatever channels we decide. These sections define the scope of this Code of Conduct, and have an intrinsic value.

Please do not discuss these topics or new ones here. Use the existing or new sections instead. Here we are only coordinating the work of proposing these three sections of the draft CoC to wikitech-l.--Qgil-WMF (talk) 10:25, 24 August 2015 (UTC)Reply

I think the current version of these sections is solid. We could leave a bit more time until Thursday evening / Friday morning Pacific to see whether there are still major changes proposed, and if not, then proceed with the call for feedback at wikitech-l. Does this sound sensible?--Qgil-WMF (talk) 07:25, 26 August 2015 (UTC)Reply
The plan and the timing sound good to me. --Fhocutt (WMF) (talk) 19:05, 26 August 2015 (UTC)Reply
Yeah, that seems reasonable. Mattflaschen-WMF (talk) 00:21, 28 August 2015 (UTC)Reply
We have got conduct-discussion wikimedia.org and this is good to go. Mattflaschen-WMF, whenever you want.--Qgil-WMF (talk) 09:42, 1 September 2015 (UTC)Reply
The email address is not yet ready. The privacy settings need to be adjusted. I will send out an announcement to wikitech-l when it can be used. Mattflaschen-WMF (talk) 00:28, 2 September 2015 (UTC)Reply
Email is ready, settings have been adjusted, and wikitech-l has been emailed. --Fhocutt (WMF) (talk) 22:05, 4 September 2015 (UTC)Reply

Level of experience

"We are committed to making participation in Wikimedia technical projects a respectful and harassment-free experience for everyone, regardless of level of experience, gender, gender identity and expression, sex, sexual orientation, disability, physical appearance, body size, race, ethnicity, national origin, age, political affiliation, or religion."

I feel the inclusion of "level of experience" in this list is problematic, if that is implicitly "level of technical experience". The other attributes in this list are identity attributes which have no bearing on someone's contributions, and should never contribute to someone being less respected, welcomed, encouraged, etc, etc that anyone else.

However level of technical experience does influence how respected, welcomed, and even encouraged that a participant will be, especially when at the coal-face of code review.

I can appreciate that our technical environment should be "harassment-free" for everyone, and especially for juniors. (This is especially critical in physical spaces, however that is covered with a different policy.) In the context of technical interaction in online spaces, the term "harassment" can be interpreted much more broadly by participants. Broad definitions of "harassment" are a good thing in the event that the behaviour appears to be discriminatory to identity attributes. They are not good if someone is reaching to find a reason why their technical contributions are not being accepted, and wanting to shift the blame from their inexperience or poor quality onto others.

I have similar concerns about the inclusion of "consistent and unwarranted rejection of patches", as "unwarranted" can also be a situation where code submitter vs reviewers have view different perspective of expectations wrt quality, and ultimately review of patches should be a consensus or stalemate of available +2'er, and this policy is unlikely to be a useful tool in preventing stalemates between +2'ers.

I believe that level of technical experience, if that is the objective here, should be handled separately from identity attributes.

I suspect that "level of experience" is trying to get at either/both of the following, which I agree with:

  1. newness of the contributor (i.e. number of years involved in this community) shouldn't be a discriminatory aspect
  2. power relationship games are completely unacceptable

Sorry I don't have answers for how to rephrase this yet. John Vandenberg (talk) 05:33, 26 August 2015 (UTC)Reply

I think part of what its trying to get at, is making fun of people for lack of skill, is not ok. I think there's also a dual meaning of the word "respect". It can mean being a decent person, but it can also mean how seriously you treat their ideas or how likely you are to defer to their decisions, etc. I believe its important to be nice to everyone. But if someone wrote on wikitech-l proposing to rewrite half of MW, I'm going to treat their email much more seriously if somebody like Tim Starling is writing it, than if someone I don't know is writing it. Bawolff (talk) 05:55, 26 August 2015 (UTC)Reply
"a respectful and harassment-free experience for everyone, regardless of level of experience" means that we commit to avoid that someone shows lack of respect and/or harasses someone else because of their level of experience. If I open a discussion in wikitech-l about rewriting MediaWiki from scratch as a Lua template (!), the community has all the right to ignore me or tell me that I have no idea what I am talking about, but according to this CoC I will still have the right to be respected and not harassed. Same if my knowledge of written English is poor, I write emails as if they were txt mssgs, use all-caps too much, top-quote, use a vague subject in my new email thread, ask every other minute on IRC because I don't know better yet, and many other very realistic cases. These behaviors are annoying and require certain patience and dedication to fix, but no one deserves "Offensive, derogatory, or discriminatory comments" because of this.
For the rest, look at the new version; I think that it has solved your concerns as well.--Qgil-WMF (talk) 07:15, 26 August 2015 (UTC)Reply
There is a genuine choice here about what sort of Community to be. Consider the situation where a novice makes a suggestion that is apparently reasonable but has been tried before and is known not to work, or is unworkable for some other, not obvious, reason. In some communities I have known, it would be acceptable to say "no, you don't know what you're talking about" or "Wrong. End of" and to respond to the obvious followup question with an accusation of trolling. In others, it would be regarded as rude to reply with anything less than "we tried at out in such a date and it didn't work as you can see on such a page" or "that sounds reasonable but the buffer is updated asynchronously and you can't be sure that the data is there although it almost always is, and we have decided to design code for the unlikely cases as well.". In other words, some communities choose to be learning communities, which takes more time, and others choose a model in which experience and expertise buys you the right to be brusque with newbies. Which do we want to be? Rogol Domedonfors (talk) 07:29, 26 August 2015 (UTC)Reply
"an open and welcoming community" (as described in Principles). We already are (or at least want to be) open and welcoming. There are many examples, although we can always improve. In any case, I don't think this discussion would bring any changes to the current proposal.--Qgil-WMF (talk) 07:44, 26 August 2015 (UTC)Reply
We want to be a learning community. Experience is never a justification for rudeness; in fact, it undermines any defence. Ironholds (talk) 18:15, 26 August 2015 (UTC)Reply
One step further: if we want to make it so that "every single human being can freely share in the sum of all knowledge", and not just the human beings whose backgrounds we understand and whose expertise we already respect, we must be a learning community. It takes more work, yes, but it's work worth doing. --Fhocutt (WMF) (talk) 18:56, 26 August 2015 (UTC)Reply
I think we ought to be the kind of community that tries to engage new people and learn together. Besides basic respect, another advantage of the "that sounds reasonable but the buffer is updated asynchronously" approach is that maybe a few months later, they will post to wikitech-l "But what if the asynchronous update triggers an event using this mechanism..." (or whatever), which might turn out to be a really good idea. Bringing new people into the community also brings in new ideas and experience (they might not know anything about wiki software, but what if they know a lot about algorithmic optimization, or UX, or...)? Mattflaschen-WMF (talk) 00:28, 28 August 2015 (UTC)Reply
@John Vandenberg: : both of those, yes, but also what Qgil-WMF mentioned. Knowing the social conventions of an open source project (or how to find them out) is not actually trivial, though we tend to forget it the longer we participate in a given community. Open source has a long and unfortunate tradition of "RTFM, n00b" and this point is intended to address that. Specifically, I'm aware of someone who was turned off MediaWiki development after seeing their early contributions discussed uncharitably on IRC. Learning doesn't happen when people are afraid they'll be mocked or insulted for making mistakes, and many people will just go away when they feel unwelcome. --Fhocutt (WMF) (talk) 19:29, 26 August 2015 (UTC)Reply

Thanks for the responses, which help clarify the intent. Also thanks for removing "consistent and unwarranted rejection of patches".

However the inclusion of level of experience, in this list, bothers me as raises it to the same level as discrimination based on identity attributes. In many countries, including mine, there are discrimination laws that explicitly address the majority of these identity attributes. Those laws are usually full of complicated aspects, and typically only apply to employees, differ per country, and often dont handle online spaces well, so it is wonderful to have a community consensus which unreservedly and strongly rejects all forms of identity based discrimination and harassment, using broad language, in our shared online environment.

To me, adding 'level of experience' in this list trivialises all of the other items in the list, which are all identity based. 'level of experience' (broadly speaking) does deserve being highlighted in the CoC, somewhere else, but it fundamentally is not the same as the others in this list. Including it weakens our commitment against identity based discrimination and harassment, when it should be using stronger language for those. Dealing with contributors with seemingly low competence/familiarity can become quite hard, especially if they are not hearing the sometimes too subtle cues, and we all struggle with how to support newcomers effectively - there is always room for improvement here (IMO not reviewing patches in a timely fashion is not respectful). Of course extreme reactions, such as mocking a contributor, is unacceptable, however "RTFM, n00b" is a completely different type of problem to responses that focus on a person identity. Any complaint lodged relating to identity attributes, irrespective of how trivial it is, should result in strong community rejection, and rapidly escalating enforcement. Non-trivial complaints lodged relating to primarily to 'level of experience' are typically going to be complex to untangle, and will often be unresolvable due to difference in perspectives of the events in question. If isolated, complaints related to 'level of experience' will need to give the accused the benefit of the doubt. Repeated bad behaviour related to 'level of experience' will need more careful handling, and possibly even providing training if the accused is a valuable technical resource. John Vandenberg (talk) 08:22, 27 August 2015 (UTC)Reply

Would it help to have "level of experience" at the end of the list instead of the beginning? Level of experience is a known target of certain types of unacceptable behavior that we don't want to see in our community. Identity is a tricky concept and I don't think it is worth entering into this debate. Some people won't care much about their own political or religious views, but they might feel stronger about their "geekiness" or "non-geekiness". Also, maybe "level of experience" is not the best wording. Someone can be an expert in several fields of humanities and sciences, be a great family and social partner organizing all kinds of stuff, but have little experience in the context of a free software project. "technical skills" might be a better label for that.--Qgil-WMF (talk) 08:56, 27 August 2015 (UTC)Reply
I'd be much happier if our treatment of newcomers / competence levels / skills is clearly separated from the others. IMO it is worthy of its own sentence/paragraph in the Principles section. Putting it at the end doesnt address my concern. If there is any interest in separating it, I'd be happy to put in some effort to try separating it from the others in a way that highlights it as an important problem. If not, so be it.
"technical skills" is IMO better. John Vandenberg (talk) 09:25, 27 August 2015 (UTC)Reply
I think one of the things we're trying to say here is that established contributors do not get special consideration or licence in matters of conduct on account of the claimed value of their contributions. (This is a principle which the English-language Wikipedia continue to struggle with.) Is that indeed what we are agreed on? If so, is there is suitable form of words? Something like "The behavioural standards expressed in this code apply equally to all members of the community irrespective of status". Rogol Domedonfors (talk) 06:47, 29 August 2015 (UTC)Reply
I have tried to address your feedback in this edit.--Qgil-WMF (talk) 14:05, 31 August 2015 (UTC)Reply
That is pretty much what I believe the consensus here to be. Rogol Domedonfors (talk) 16:52, 31 August 2015 (UTC)Reply
@Qgil-WMF: @Rogol Domedonfors: I completely agree with the principle behind "The behavioural standards expressed in this code apply equally to all members of the community irrespective of status" (not saying we necessarily need that exact sentence in it, but I support that idea). However, I don't think that's all it should mean. It should also including the ethos of (these are ideas, not proposals for the draft) "Treat newcomers in a welcoming way" and "Try to make interactions with them educational where possible, rather than devolve into RTFM, or just 'Won't Work'", and "Don't harass them because they're new". Mattflaschen-WMF (talk) 23:15, 31 August 2015 (UTC)Reply
Now? (I thought I know everything about wikitext discussions, but I'm getting lost with bullets, colons, and indentations...)  :) --Qgil-WMF (talk) 09:34, 1 September 2015 (UTC)Reply
Yeah, the current text for this looks fine. Mattflaschen-WMF (talk) 23:40, 1 September 2015 (UTC)Reply
Yes, I like the phrasing in the draft now. --Fhocutt (WMF) (talk) 17:47, 1 September 2015 (UTC)Reply

Much appreciated everyone who helped fix this. The current wording is great! John Vandenberg (talk) 06:06, 5 September 2015 (UTC)Reply

Who is going to approve this/how?

It's not clear to me how we expect this draft to be made into policy. Are we going to vote on this? If does eventually become policy, what would be the process for amending it? --Krenair (talkcontribs) 21:00, 29 August 2015 (UTC)Reply

It was suggested above under #Governance that it be formally adopted by Resolution of the WMF Board of Trustees. Rogol Domedonfors (talk) 21:37, 29 August 2015 (UTC)Reply
I think who approves and how depends on the final draft, and this is why I'm focusing on pushing the draft further as proposed in #Next_steps. I would personally prefer to try first a process of discussion and consensus.--Qgil-WMF (talk) 14:09, 31 August 2015 (UTC)Reply
I agree with Qgil-WMF that consensus here on this talk page is the best way. A Board resolution is unnecessary. See my last comment at Talk:Code_of_conduct_for_technical_spaces/Draft#Proposed_additions ("This does not require a Board resolution, [...]") for why. Mattflaschen-WMF (talk) 23:44, 31 August 2015 (UTC)Reply

Authority/power of committee?

It's not clear how a committee is going to work unless they are either actually given the power to issue bans etc. themselves, or can convince local technical space admins to give effect to their decision in every single controversial case (of which there will hopefully be few/none, but that can't be assumed) --Krenair (talkcontribs) 21:04, 29 August 2015 (UTC)Reply

Local technical space admins are subject to this CoC just like anybody else. There is a specific sentence mentioning them in the Principles section. They should apply the decisions of the CoC or have a good argument for an appeal.--Qgil-WMF (talk) 14:16, 31 August 2015 (UTC)Reply
There is also specific text ("The committee can also bring in people from the spaces involved (e.g. if it happens on both MediaWiki.org and IRC, bring in involved admins and IRC contacts).") about bringing in local space admins when appropriate. In such cases, the local space admins have a direct opportunity to participate in discussion, and thus better understand the basis for the committee's decision. Mattflaschen-WMF (talk) 23:50, 31 August 2015 (UTC)Reply

IdeaLab/Community discussion on harassment reporting

Has this discussion take full account of the parallel meta:Grants:IdeaLab/Community discussion on harassment reporting? Rogol Domedonfors (talk) 11:29, 30 August 2015 (UTC)Reply

Do you see specific points that we could take in our CoC draft?--Qgil-WMF (talk) 14:24, 31 August 2015 (UTC)Reply
It contains a lot of useful material. The section What types of harassment have been reported on Wikipedia is revealing (and depressing). Not all are explicitly covered here. That's just one example. Rogol Domedonfors (talk) 16:50, 31 August 2015 (UTC)Reply
Looking at "Community claims" and "What types of harassment have been reported on Wikipedia", I think the cases mentioned there would be covered by our CoC, even if they are not necessarily spelled out with the same words here. If someone finds an actual gap, please report it.--Qgil-WMF (talk) 09:40, 1 September 2015 (UTC)Reply
I don't know if that specific document has been used. However, Grants:Friendly space expectations has been, so we are taking some lessons from their experience. As far as the specific list of things that happened on Wikipedia, this is partly why we have "Examples include but are not limited to". Mattflaschen-WMF (talk) 00:07, 1 September 2015 (UTC)Reply

Unacceptable behavior

In the beginning of the "unacceptable behavior" section, the phrase "unacceptable behavior" (which I just changed to "Behavior that is unacceptable") linked to the Terms of Use. What exactly are we trying to communicate with the link? That we consider all the behavior mentioned to be harassment and therefore prohibited by the terms? If so, we should make that statement explicit (and check that it's legally okay too, although I'm not sure the WMF legal team could give us an opinion).—Neil P. Quinn (talk) 22:06, 18 August 2015 (UTC)Reply

That was taken from meta:Grants:Friendly Space Expectations. My reading was that it is a reminder that harassment is already legally prohibited for Wikimedia-hosted projects, rather than a statement on the specific behavior listed here. If it's confusing, I'm fine with removing it. --Fhocutt (WMF) (talk) 01:03, 19 August 2015 (UTC)Reply
The link is inappropriate, in my opinion. I already removed it once. --MZMcBride (talk) 02:36, 19 August 2015 (UTC)Reply
I moved the link to the "See also" section as that seemed reasonable. --MZMcBride (talk) 02:44, 19 August 2015 (UTC)Reply
I think that's a good place for it, especially as this applies to spaces outside of WMF resources. --Fhocutt (WMF) (talk) 19:21, 19 August 2015 (UTC)Reply
Clarify "outside of WMF resources"? --Nemo 07:16, 21 August 2015 (UTC)Reply
The technical IRC channels, for example; the WMF doesn't "own" those. Ironholds (talk) 14:36, 21 August 2015 (UTC)Reply
Indeed, which also means it can't legislate over them. Hence the statement "this applies to spaces outside of WMF resources" appears to be an oxymoron. Nemo 16:45, 10 September 2015 (UTC)Reply
Not in the slightest; sure, you can't kickban someone from IRC for creating a tremendously harassing experience for contributors (well, not necessarily, although I'd like to think the channel admins would do something about it) but you can absolutely factor that into whether they get to wander around other spaces when they've demonstrated that kind of behaviour. Ironholds (talk) 20:31, 11 September 2015 (UTC)Reply

Another membership proposal

I'd like to propose a self-perpetuating committee, boot-strapped by the Engineering Community Team (ECT).

  1. We open up self-nominations (if you think someone would be a great candidate, you can ask them to self-nominate). These self-nominees can make a statement up front.
    As part of self-nominating, each nominee must agree to comply with the full code of conduct, including the Reporting and Enforcement procedures. Mattflaschen-WMF (talk) 23:55, 31 August 2015 (UTC)Reply
  2. People would then comment on nominations, both publicly and by emailing ECT (as they prefer).
  3. ECT consults with Community Advocacy on the candidates. The past behavior of the candidates should be in the spirit of the Code of Conduct.
  4. The ECT will then select five candidates, considering their involvement in the community, background, and experience that would inform their Committee work (if any; this is not a hard requirement).
  5. Every 6 months, the Committee will vote on a new Committee to replace it (Nominations are opened one month before this month; re-nominating yourself is allowed). Majority approval (3/5) of the overall slate is required, but unanimity is encouraged if possible. Re-selections are possible, but no one can be on the Committee for more than 12 consecutive months, unless the Committee unanimously agrees that an exception to this is necessary because suitable new candidates are not available.
  6. ECT can remove Committee members, but this should be only be exercised in case of severe problems.

It is required that the Committee can never be 100% WMF staff, and it's encouraged to choose various affiliations.

I think the main benefit here is that the Committee will have some knowledge about the community, and problems we face, so they can use that to help inform their decisions on the next Committee. Mattflaschen-WMF (talk) 04:23, 21 August 2015 (UTC)Reply

I find this proposal a starting point more interesting than candidates + votes, a model that I don't think suits well for this committee. Some comments.
  • Self-nominations or just open nominations? I think it is everybody's interest to see as many potential names as possible. Some people might discover that they are considered good committee candidates only after someone else sends an open invitation to a mailing list (happened to me once, ended up in the GNOME Foundation board even if I had no preconceived plans for it).
  • I don't see why ECT / CA should have such privileged role. We shouldn't receive private positionings about candidates (anonymous comments on-wiki work fine), we shouldn't decide (the community should, and the own candidates nominated should). If we have, say six candidates but only four of them look like having wide support, I'd rather start with these four, allowing them to fine tune the team whenever it's time to refill/renew.
    • Given that we're accepting private email feedback for the CoC itself, why can't we accept it for candidates? This proposal gives ECT and CA a role in the first committee because right now, they probably have the best knowledge of conduct issues in the tech community. Once the Committee is up and running, probably the Committee will have the best knowledge of that (hence why the Committee elects its successors). Mattflaschen-WMF (talk) 20:31, 2 September 2015 (UTC)Reply
      • Yes, precisely. There is no reason to require someone to post about their negative experiences with a candidate on-wiki, where their story and reactions will be picked over and discussed at length, in order for that information to be taken into account by those choosing. I know that Community Advocacy knows a lot more about various bad actors than I ever want to, and a lot more than is appropriate to ever post on-wiki. I value that knowledge and want there to be a way for it to be taken into account when the committee is chosen. I think that asking ECT to use their trusted position to help bootstrap the committee is very reasonable. --Fhocutt (WMF) (talk) 03:06, 3 September 2015 (UTC)Reply
        • Mmmok. So private feedback about candidates will be provided. Still, we are mixing private channels here. Private reports about CoC violations will go to the committee. Private feedback about candidates will go... where exactly? The committee yes/no? ECT yes/no? I'm reluctant to include CA or anybody else, just to avoid them an automatic new responsibility (they have enough) and to assure to the senders of private feedback a small cosy audience. If ECT feels like sharing/asking more information to CA or whoever, we can, keeping the privacy barriers as needed.--Qgil-WMF (talk) 10:01, 3 September 2015 (UTC)Reply
          • @Qgil-WMF: Private feedback on candidates should go to whoever chooses the committee. If we go with some version of the above, ECT will choose the first committee (and thus get the first round of private feedback). The committee itself will choose its successor committees, so it should get the private feedback for those rounds. Mattflaschen-WMF (talk) 03:50, 9 September 2015 (UTC)Reply
  • Reopening the membership every six months sounds unnecessarily demanding. Becoming a good CoC committee member probably takes longer than that even if you already are a good candidate.
  • Why forcing someone to leave after 12 months? There are no strong reasons to remove a member if they want to stay and everybody is happy about them. I don't even think we would have enough good candidates for such refresh rate.
    • We could consider tweaking the timeline. Maybe elections every 12 months, max duration of 2 years. There are a couple reasons for this idea: Avoiding burnout, and making sure the committee continues to represent the whole community without being too identified with individual people. Mattflaschen-WMF (talk) 20:31, 2 September 2015 (UTC)Reply
    • Burnout is a serious concern for positions like these, and people tend to stay in them out of a feeling of obligation even as they become less effective and more burnt out. Having a mechanism that forces members to reconsider and take a break is a good thing, although I'm open to different implementations. --Fhocutt (WMF) (talk) 03:06, 3 September 2015 (UTC)Reply
      • What about bootstrap + 12 months + changing at least one member every six months, and leaving to the discretion of the committee to change more positions in every refresh? Looks like a good compromise between continuity and change. Other than the principle of at least one change every 6 months, I don't think the CoC needs to get into details such as 3/5 votes for approving new members. The committee should be entitled to decide the best way to renew itself.--Qgil-WMF (talk) 10:07, 3 September 2015 (UTC)Reply
        • @Qgil-WMF: Can you elaborate on your proposal? Does "12 months" mean elections every 12 months? If so, how would a member change every six months? Are you suggesting staggered terms? I recommend we do specify the rules for the committee electing its successor. This is a substantive issue. I assume you don't want 2/5 or 1/5 to choose the next committee. So the question is whether it should be 3/5, 4/5, or 5/5. 4/5 and 5/5 are supermajorities. The well-known downside is that supermajorities allow a minority of the committee/parliament to control all forward progress. E.g. if the committee required elections to be unanimous (5/5), any one committee member could block the election unless their friend was assured a seat in the next committee. Mattflaschen-WMF (talk) 03:57, 9 September 2015 (UTC)Reply
          • I mean 12 months in the first term, then always 6 months -- see #Bootstrapping the committee. I still think "trust the committee" is a better principle than trying to define numerical rules in advance. A committee trusted to enforce the CoC should be trusted to handle its own functioning. If a specific committee cannot be trusted on the latter it probably won't be trusted on the former, and we will have a bigger problem than supermajorities anyway.--Qgil-WMF (talk) 17:39, 9 September 2015 (UTC)Reply
We have the model of the Architecture committee, which is similar to this case (experience, trust, complex issues...) The CoC committee should be transparent in their initial formation and their renewals, but other than that I think a model based on empowering the committee to maintain themselves will work. If the committee misbehaves there will be crisis and loss of trust, and if that happens then we will surely discuss how to move forward.--Qgil-WMF (talk) 09:52, 2 September 2015 (UTC)Reply
  • On the capacity of ECT to remove members, I don't think this is a good idea. If a committee member deserves to be removed, the case should be clear enough for the committee fellows to take action. If ECT wants to remove a member when the rest of the committee is opposing, that means that we all have a deep problem that no rule will solve.--Qgil-WMF (talk) 10:33, 3 September 2015 (UTC)Reply
  • On the permanent seat for an ECT member (something not mentioned in this proposal but discussed elsewhere without a clear conclusion) I still firmly think that ECT should have no privilege. If one of us wants to go for it, then self-nomination and selection should be the process like for anybody else. And if an ECT member is in the committee but for whatever reason it is thought that it is better to leave, the process to refresh seats every six months could be used for that.--Qgil-WMF (talk) 10:33, 3 September 2015 (UTC)Reply
    • The Engineering Community team members agree that ECT should not have a permanent seat or power to veto committee members.--Qgil-WMF (talk) 01:02, 9 September 2015 (UTC)Reply
      • Okay. I do think there needs to be a way to remove members, for very serious problems. However, I agree that in most cases, it should be sufficient for the committee to do so. I've added this (requiring a supermajority because unlike regular elections, removing someone mid-term isn't something that should be required often). If this isn't (e.g. the committee is turning a blind eye to one of their members), a procedure could be developed then. Mattflaschen-WMF (talk) 04:16, 9 September 2015 (UTC)Reply

Next steps

We need a bit of organization. Even when the discussion is being shared basically by just a few people, it is becoming harder to follow and get involved productively (even more for volunteers with little time available). The goal of this exercise is to draft a proposal of an enforceable Code of Conduct to be approved by the Wikimedia technical community. We can slice the drafting in modules, allowing for a lighter process of discussion and consensus:

  1. Intro + Principles + Unacceptable behavior define the scope of this CoC, and have an intrinsic value. If those of us active in this Talk page now focus on these points, we should be able to reach a level of agreement good enough to send a proposal to wikitech-l.
  2. A new section "Committee" (that I will create now unless someone is faster) comprising all the details about this committee will need some more iterations. We can work on it while the community has time to discuss the points above. Hopefully by the time that the points above are stable and clear, we could share our proposal about the committee to wikitech-l.
  3. With the previous points on its way for agreement, Reporting, Appealing, and Enforcement should be relatively simpler to deal with, as they are mostly a matter of process. Once we have a draft we are happy with, we would share it with wikitech-l, adding that this is the last chunk of the CoC and we are seeking general consensus for the approval of the entire document.
  4. The CoC would become official as soon as it is approved. Then we would proceed to the creation of the committee.

The trick here is that the most active contributors to this discussion would focus on the topics that need more urgent approval, leaving he rest for later. Sporadic contributors can comment anywhere, of course.

There have been some comments about a self-selection (for a lack of a better word?) of people participating in this draft, in this discussion page, and in the related threads in wikitech-l. In order to help capturing a wider range of opinions, we could offer an email alias i.e. coc-discussion wikimedia.org that would relay emails to, say, three identified people that will commit to treat this feedback with confidentiality and channel the contributions received to the public discussion. Three is big enough to be reliable (a single person might miss an email, be away, etc) and small enough to maintain personal trust.--Qgil-WMF (talk) 08:36, 21 August 2015 (UTC)Reply

I really like this proposal. One implication it has, though, is for the approval process; the problem with a consensus-driven approach is that it is very hard to identify the trend of agreement or disagreement when confidential feedback or disapproval or approval is being factored in. What are your plans for dealing with that? Ironholds (talk) 14:34, 21 August 2015 (UTC)Reply
I like this staged approach, although I share Ironholds' concerns about how to make consensus (or its lack) apparent. I do think that offering the opportunity for anonymized input is important in this discussion. --Fhocutt (WMF) (talk) 00:53, 22 August 2015 (UTC)Reply
I think anonymized feedback can still be taken into account in reaching consensus, as long as it is posted to the page (in anonymous form). Mattflaschen-WMF (talk) 22:53, 23 August 2015 (UTC)Reply

I also like this plan—it isolates what seems to be the hardest part of the discussion: the composition of the Committee. Ironholds: I think there are two routes we could take to deal with confidential feedback:

  1. Have the trusted parties channel it into the discussion, but have it impact the discussion by (hopefully) affecting the positions of the people taking part directly.
  2. Allow people to send their positions anonymously to the trusted parties, and have those parties paraphrase it, confirm that the senders haven't !voted publicly in the discussion and have some basic level of involvement in the Wikimedia technical community (I don't think this should be high—for example, a preexisting Phab account should be sufficient, even if it hasn't commented), and add it to the discussion on the same level as an identified comment.

I lean towards 2, personally.—Neil P. Quinn (talk) 06:39, 22 August 2015 (UTC)Reply

I like 1 better. I appreciate that some people may be reluctant to publicly state their positions, but I think its important to make sure that their is no question about the community support for the proposal (In order to ensure the entire community feels bound by it), and part of that is being able to verify that the vote isn't being controlled by a sock puppet army. Bawolff (talk) 07:23, 23 August 2015 (UTC)Reply
@Bawolff: I did think about that, so in option 2 I mentioned that the trusted parties receiving the anonymous feedback would take some basic steps to guard against sockpuppets (essentially the same steps that anybody would take take when checking public comments): checking that the sender has a Phab/mediawiki.org account that predates the discussion and that they haven't already commented. Do you think these wouldn't be enough?—Neil P. Quinn (talk) 18:09, 23 August 2015 (UTC)Reply
Its more I would like to err on the side of caution. Its a complicated issue, I'm not 100% sure where I stand. Bawolff (talk) 19:27, 23 August 2015 (UTC)Reply
It's not trivial to cross-check someone's email address with their MediaWiki.org or Phabricator account. First, some accounts don't even have email addresses associated (but the anonymous feedback might still come from an email). Even if there is an email associated, you would need to use Special:EmailUser to send them a message to confirm the previous email came from them (otherwise, you don't know their email address and don't know if the from address of the original email was forged). Mattflaschen-WMF (talk) 23:45, 23 August 2015 (UTC)Reply
OK, then let's start applying this sequence. About anonymous feedback, I think we should go for 1 as long as consensus is build through discussion (and not votes). If we decide to go for votes, then the scenario changes, but there should be enough time for anonymous feedback in any case. Good ideas are good ideas, regardless of where they come from. If they are influencing, so be it. The fact that such ideas come from a techie sockpuppet, someone alien to Wikimedia tech, etc, it doesn't matter. Considering that users can post here anonymously already (with an IP or a newly created account), users of the private channel would probably be people willing to make a point without having followed the entire discussion or without willing to pick a fight publicly.--Qgil-WMF (talk) 09:51, 24 August 2015 (UTC)Reply
Who should be the three people publicly identified as recipients of coc-discussion wikimedia.org? Once we have the names Matt or I can create the request in Phabricator. We should offer this email address when sending our next email to wikitech-l (see #Proposing Intro + Principles + Unacceptable behavior so I hope we can go through this step quick.--Qgil-WMF (talk) 08:14, 26 August 2015 (UTC)Reply
I am willing to be one of those people. Also, I suggest "conduct-discussion wikimedia.org", as abbreviations can be confusing. --Fhocutt (WMF) (talk) 19:09, 26 August 2015 (UTC)Reply
No more volunteers so far? Then I will join as well. Two people are enough to get this started. I have requested the conduct-discussion wikimedia.org email alias to WMF Office IT and I have also CCed Mattflaschen-WMF, so he knows when this email address (the last blocker for the announcement) is ready.--Qgil-WMF (talk) 07:53, 27 August 2015 (UTC)Reply
The email address is not yet ready. The privacy settings need to be adjusted. I will send out an announcement to wikitech-l when it can be used. Mattflaschen-WMF (talk) 00:28, 2 September 2015 (UTC)Reply
The email address is now ready. Kalliope from CA has agreed to be a third person receiving the conduct-discussion emails. --Fhocutt (WMF) (talk) 22:04, 4 September 2015 (UTC)