Talk:Code of Conduct

Jump to navigation Jump to search

About this board

Suggested amendment: Public logging of bans made by CoC

51
Ladsgroup (talkcontribs)

Given the discussion happened in the past couple of days, here's my proposal to amend to CoC, part of cases:

Any ban made by the code of conduct committee will be logged publicly in Code of Conduct/Cases/Log/2018 with user name, space of ban (phabricator, mediawiki.org, etc.), duration, and if not private reasoning. Unless any of these two conditions apply:

  • The reporter asks that ban not to be public.
  • Or the code of conduct committee decides not to disclose anything as it might disclose private information. For example, logging harassment cases in Wikimedia events can lead to real-world identity of users being disclosed.

It goes without saying that this is not applicable retrospectively.

DBarratt (WMF) (talkcontribs)

Just to clarify, this would apply to bans specifically, not cases themselves?

Siebrand (talkcontribs)

I’m against this. Even with the on wikitech-l suggested option of only making it public while the ban is in effect, if the log is kept on a wiki, the (history of) the ban list can (and will) be used for naming and shaming, long after punishment has been dealt and served.

If any transparency is required, I would counter propose that an aggregate report is published periodically (quarterly/bi-annually, yearly —- depending on the number of expected cases —- max 10 or so per period would be nice), which report on the number of actions per platform and the duration of the actions, where applicable.

DBarratt (WMF) (talkcontribs)

Also, the ban might be worn as a badge of honor, which would encourage further abuse. :(

P858snake (talkcontribs)

Is having a list of bans performed by CoC any different than a wiki having Special:BlockList?

DBarratt (WMF) (talkcontribs)

No, but perhaps that shouldn't exist either. Most websites don't list all of the users who have been banned, and I don't think that is necessarily a bad thing.

Smalyshev (WMF) (talkcontribs)

Most websites are not collaborative projects supported by a huge community.

DBarratt (WMF) (talkcontribs)

That's true, but as an example: GitHub, Stack Overflow, or any an all Open Source projects (including projects much much larger than ours). I can't think of a website/project (other than Wikimedia's) that publicly lists bans. I think the public log is an assumed virtue, when the existence of the list, may in fact, encourage more harassment. I have no evidence to suggest this, my point is only that the existence of such a list may have unintended consequences.

Smalyshev (WMF) (talkcontribs)

Not existence of such list is already having consequences. And bans on the wiki had been public since forever, as far as I know, without any evidence this did encourage anyone for more harassment.

DBarratt (WMF) (talkcontribs)
MusikAnimal (talkcontribs)

At least on enwiki, I don't think many people use Special:BlockList unless they're looking up a block ID to investigate an autoblock. It chiefly consists of drive-by vandals, so it doesn't really serve as a wall of shame or badges of honour, as you'll quickly be buried in there and forgotten. The block log itself obviously is necessary to see prior abuse. At any rate, admins can redact the log entries, but on the English Wikipedia this practice is prohibited.

Ladsgroup (talkcontribs)

Yes, Not even warnings neither any mention of reporter will be logged. Hope that helps.

Krinkle (talkcontribs)

I'd propose one additional clause that requires:

  • In the first month of each year, last year's log is amended one final time to disclose the number of non-public bans by space of ban. This number cannot be opted out of by the reporter, the committee, or anyone else. And this statistic would not contain any start date, duration, user, or reasoning. Just the numbers by space of ban. (Bans affecting multiple spaces could be counted multiple times, or we could decide that "Multiple spaces" be one of the counted spaces.)
  • In the first month of each year, last year's log should be linked to from a public announcement. I have no preference for which platform this post would be on, but it should be decided ahead of time and included in the policy. Perhaps Wikitech-l, or another public wikimedia.org list, or a mediawiki.org newsletter, or a Phabricator Phame post, etc.

Lastly, I would recommend that this clause does apply retroactively, treating all prior cases as non-public. Thus disclosing this year's complete counts by January 2019.

EDIT: I wrote this comment at the same time as @Siebrand wrote his. I did not see his until after I submitted mine. I would support doing only aggregated reports.

Bawolff (talkcontribs)

I would prefer that this applies not just to bans but all sanctions by the committee (e.g. asking someone to apologize doesn't need to be logged, but doing something to someone against their will does. e.g. deleting a phab comment does require logging).

I even wonder if maybe the enforcement role of the committee should be separated from the judgment roles of the committee. Right now I feel like the CoC committee is judge, jury and executioner and lacks any effective oversight due to secrecy. I consider this state of affairs dangerous

I would also like to suggest that for any sanction, the person being sanctioned must be notified of the action, length, rationale and how to appeal. I've heard complaints that in some cases when comments were deleted the commenter was never notified, and only found out by accident (which is supposed to deter them how?). In the recent MZMcbride case, the notification was rather lacking (Assuming the email posted by Mzmcbride is accurate) as it failed to disclose length, how to appeal and the rationale while not totally missing was lacking imo (that's more debatable though).

Smalyshev (WMF) (talkcontribs)

I agree with the publishing of the bans, but I am not sure Wiki (and specifically Mediawiki) is the best place because, as correctly noted, the wiki infrastructure ensures the information is preserved forever. While in theory everything public on the internet is forever, in practice there might be benefits to not carrying old grudges around.

The problem to solve here is not statistics though. It's nice, but it's a different case. What is really needed is when action is in force, I (as a participant of the collaborative platform) can see that it is, and adjust workflows accordingly. It is a collaborative environment, so the ban would affect collaboration. Aggregated report is useless for this case - if I worked with X and their account suddenly shows up as "disabled", it won't help me any that in 2 months time I'd see in aggregated report "number of bans: 7". What I need to know is: a) was X banned? and b) for how long. Most community members, I suspect, would also want c) by whom and d) for what.

I am not sure how it can be possible that the ban won't be public - account being disabled on Phabricator is pretty public, and we'd be kidding ourselves if we ignore the fact that this is what people would assume (especially now). If we do it in non-transparent way, people would just assume more and create a narrative in their heads which might not necessary be even true, but is inevitable. I do not think the fact of the ban in a public collaborative platform can be hidden. Or should be. We might not disclose any details about it beyond the duration (I think this is a must), though I would advise in general case not involving sensitive matters to opt to more disclosure (again, secrecy breeds mistrust, and mistrust makes people construct narratives that we'd want to avoid). But I do not see how hiding bans is practical, or desirable.

MusikAnimal (WMF) (talkcontribs)

I would discourage using wikitech-l to publish ban reports. This mailing list is read by newbies and other developers who just want to stay informed about technical matters. It is not very welcoming to see such drama and it might incite more lengthy and toxic discussion. I think an on-wiki page is fine, or somewhere that's not as "in your face". I should not be forced into seeing this side of the technical community.

Edit -- I meant to post under my volunteer account, for the record!

Smalyshev (WMF) (talkcontribs)

I think using wikitech-l is a consequence of a lack of pre-designated forum for such discussions. When there's a need for discussion and community does not have a Schelling point for it, the next best thing is taken to be it. To avoid it, we need to arrange such space in advance, and specify it when CoC decision is communicated, and maybe also in CoC pages where we establish the rules.

Psychoslave (talkcontribs)

I join Bawolff on the lake of separation of powers. I would add that it would be good to have well defined processes and publicly documented rules stating for which cases what kind of sanction can be applied or not. That doesn't mean absolute transparency on each details of execution of this processes, which is certainly not a required or desirable level of transparency permitting trust to foster, and can itself be a source of legitimate concerns for people safety, as it was pointed. Although I am not sure that this is the most appropriate place to talk about this aspect of the issue.

As there is a collaboration aspect that have been indicated, I think Meta would be a more appropriate place to publish a report, if any.

Regarding the feedback of status of collaborators when they are banned, I don't have a perfect idea so far. My first thought would be something like integrate a feature in Phabricator itself, showing in the user page that "this account as been (temporally) disabled (for 3 eons), for more information consult this page/contact that referent".

Strainu (talkcontribs)

I support the original amendment. I don't care much about the venue for publishing the bans, but it is tremendously important that they are visible at least for the duration of the ban. Not everyone who needs to be aware is in the CoC committee (phabricator.wikimedia.org and wiki admins, event organizers etc)

DBarratt (WMF) (talkcontribs)

This seems like a solution in need of a problem. Would you mind clearly stating what the problem is so we can determine if this is the appropriate solution?

Smalyshev (WMF) (talkcontribs)

The problem is very simple. They way it works now there's no indication to anyone (even Phabricator admins!) what happened when account is disabled. Not who, not why, not for what duration, nothing. It is unacceptable that on a collaborative platform people are just disappearing without any notice and any possible explanation (and since the account is disabled, they can't communicate what happened either). This is not how open collaboration should be done, and not how a collaborative environment should be administered. I understand the privacy concerns, but my firm belief is it is completely possible to produce a transparent and accountable system without revealing any private details. Just hiding everything and not providing any information at all is not a good solution.

MarcoAurelio (talkcontribs)

Yes. CoC santions must all be publicly logged for accountability, transparency and awareness. This doesn't mean that we must require a full disclosure in the most sensitive cases but at least username, type of sanction and duration must be made publicly avalaible. With regards to the case that caused the opening of this thread, we already have two Phabricator administrators that didn't know where the block came from, one of them overturning the ban ignoring that it was a CoC sanction. This is not optimal, as we can lead to involuntary reversals and misunderstandings. It is in the benefit of both the CoC and the users to know all of this. As some have pointed above, wiki blocks are kept logged for the said reasons on a personal log and active blocks and bans listed on an special page. Why the secrecy here? Of course reports should be managed with confidentiality, but actions by the CoC based on said reports should not fully be. I must note that you can find at List of globally banned users a list of users subject to community and WMF-Legal global bans. Those lists ain't aimed at being walls of shame, but help users know why something has happened. I insist, the log doesn't have to have the full details, but at least username, space, duration (and reasoning if that's appropriate) should be published. Thank you.

DBarratt (WMF) (talkcontribs)

I don't believe there is value in transparency for the sake of transparency. Especially when that transparency comes at the cost of someone's privacy. Also, it may have unintended consequences that have not been explored.

Perhaps an acceptable compromise would be to permission the ban/block log to the administrators of the domain? (i.e. Phabricator/Gerrit Administrators). This way, it isn't a public log (that can be used for shame/honor), but it would resolve the problem of administrators not being aware of bans/blocks that they should be aware of.

Strainu (talkcontribs)

How about bans for real-life event? How can other events organizers be aware of a ban if no public log exists?

MarcoAurelio (talkcontribs)

I'd like to insist that I am not asking for the full disclosure of the case details. That'd be crazy. The log can be so brief as to "User X banned from (platform/WMF technical spaces) for (a week/indefinite) by decision of the Code of Conduct Committee [optional: because ...]." This does not IMHO disclose any personal information other than what the user might have disclosed themselves voluntarily (ie: if you use a username which is also your real name [Bad Idea (tm)]), helps understand people why an account suddenly has become deactivated, inactive or is not replying, and improves transparency and accountability of the CoC. Of course cases and investigations should continue to happen in camera, and there might be cases where revealing even that certain user is subject to a CoC santion might be a bad idea. If that situation ever comes I'd certainly support not making that situation fully public, but at least the administrators of the platform where the user got banned should be informed (e.g.: Please note that the COCCom has banned X from Y; please do not overturn) as to avoid involuntary reversals or misunderstandings. Best regards.

DBarratt (WMF) (talkcontribs)

I understand what you are asking for, but I do not understand why a ban/block log needs to be public, when it seems that the problem is that administrators are the ones who need this log. I understand that it would not include personally identifiable information. My argument is that, the block itself ought to remain private. Or are we saying that people who are blocked/banned no longer have that right to keep that information private?

Is there a problem with making this log permissioned to admins?

Smalyshev (WMF) (talkcontribs)

> Or are we saying that people who are blocked/banned no longer have that right to keep that information private?

How can you keep it private if your account is disabled and this fact is publicly visible? If account becomes disabled and then enabled back in a week, it doesn't take Sherlock Holmes to see what happened.

DBarratt (WMF) (talkcontribs)

Perhaps instead of having a log, admins should be trained to contact CoCC before unbanning/unblocking a user?

Smalyshev (WMF) (talkcontribs)

On every disabled account, even one having nothing to do with CoCC, just on the chance it might be another of the ineffable ways of CoCC? That sounds completely backwards - instead of CoCC plainly announcing the action, we have to ask it on every action that happened - is this your action? What about this? What about this? And this? And this? I don't think CoCC members would want this, and IMHO this is not the right way to manage a community. CoCC should not be a black box oracle.

DBarratt (WMF) (talkcontribs)

If that's the case, then what's the current problem we are trying to solve?

Smalyshev (WMF) (talkcontribs)

As I described above, the current setup only shows the account is disabled. It does not show the fact of the ban, the duration, the contact person, the discussion venue etc. Anything that is needed for a proper community management is hidden. The only fact that you consider to be private is not. One could guess, post-factum or after a long discussion on the mailing list and sometimes a comedy of errors with different admins undoing actions of each other, what happened - but that should not be the modus operandi of a healthy community. We shouldn't waste a lot of time and effort of multiple people to figure out what happened. Transparency and user-friendliness should be part of the system functionality, not possible emergent result of hard work by multiple people. That's the problem we are trying to solve.

DBarratt (WMF) (talkcontribs)

Those are valid reasons, and I understand the rational.

As I asked above, Is there a problem with making this log permissioned to admins? That should resolve everyone's concerns I think?

MarcoAurelio (talkcontribs)

It'd be good if you could specify a reason in the "disable account" function. The log of account enables and disables are already only visible to admins. Problem is that I am not sure if Upstream could be convinced this is a priority to introduce in their software, and I am not sure if we can make a local "hack" in our install (or even if that's even adviceable) to do so. Manual logging is the only solution for now. Where to seems to be the problematic part here.

As for other venues, Gerrit for example requires running a CLI command to disable an account (and of course there's no public log either). IRC does have their +b lists. Mailing lists have the moderation function. Wikitech and MediaWiki do have Special:Log/block for blocks and as for other technical spaces I don't know.

There's not a single "administration" for the whole lot of WMF technical spaces managers. Either all of them are informed (thus making the log visible for all of those that can issue restrictions on WMF technical spaces) or we still run into the risk of your left hand not knowing what your right hand did.

Smalyshev (WMF) (talkcontribs)

It would solve half of the problem. It will be still not transparent to 99.99% of the users of the system (who still wouldn't know what happened to the person they were collaborating with), but will allow at least admins to perform their functions efficiently. As a partial solution, it would be better than nothing, but not entirely satisfactory.

Tgr (WMF) (talkcontribs)

Some things to consider:

  • CoC bans could involve at least Phabricator, Github, Gerrit, mediawiki.org, Mailman, IRC. All of those have different admin groups who need to be aware (if nothing else, then to avoid being tricked by the blocked person into removing the block). Some of those places have have public ban lists (IRC, mediawiki.org unless suppressed), some don't have a list but you can see the block in the account status (Phabricator, mediawiki.org when suppressed), some don't have any public block info (Github, Mailman, maybe Gerrit?).
  • Suppose the banned user does not want the ban to be public - they promise to fix their ways and do not want a stain on their record (public logs can be removed but people's memories of them can't). What level of dissemination would be appropriate then? Admins of the affected spaces? Admins of any space? Temporary public record? Permanent public record?
  • If there are public records, is it important that people find it easily? If "collaborators should be aware of what happened" is a real problem to solve then putting a block log on some wiki page no one ever heard of is not really a solution.
Psychoslave (talkcontribs)

It makes me wonder, how publication of this can of logs are impacted by w:GDPR? What would be the relevant people/team to contact to have relevant advises on that matter?

Stjn (talkcontribs)

I think that the idea that logs on sensitive topics (such as any kind of abuse in Wikimedia events) should be private both for victim and abuser is laudable. You can probably have both goals succeed in different ways: a public task with a log, where without any naming there would be somewhat concrete description of CoCC actions, and a private log, visible for moderators on Phabricator and CoCC, where it would be logged with more details (not enough to damage any people, of course).

The problem of bad reputation that is provided by public logging of blocks in Wikimedia projects is real, and consequences of implementing it on a real world abuse would probably be more damaging to both parties, so finding a more sophisticated approach should not be scoffed at.

Strainu (talkcontribs)

Unfortunately this thread seems to focus solely on bans for online activities. For offline activities, there is no "admin" group that can be announced. There are all kind of independent events in universities and other venues that are in no way affiliated with the WMF, many times do not necessarily interact with WMF community liaisons, but are bound by the CoC. None of the alternative solutions proposed above considers these events.

I would ask people opposing a public ban list to consider how their proposals fit with offline events. Thank you.

MarcoAurelio (talkcontribs)

If it's not being done already, maybe the organizers of offline events should be provided by someone at the WMF/COC with a list of banned people privately so they can manage the attendance to the event safely. Best regards.

Strainu (talkcontribs)

In theory this works, sure, but there are 2 issues I can think of right now:

  • Is anyone tracking tech events throughout the world? Would it scale to do that?
  • Would that require event organizers to sign some kind of NDA? If so, that would be a huge turn-off...
Tgr (WMF) (talkcontribs)

Or make IRL bans public but other bans not. I imagine most bans will be about online spaces and not events (as most people are more likely to misbehave online than face to face).

MusikAnimal (talkcontribs)

Sorry to repeat myself (and maybe this is slightly off-topic), but if you noticed we recently had one or two new contributors welcomed on wikitech-l. The long nasty "disabled phab account" thread seems to have died down, and I hope it stays that way. Am I the only one concerned about how bad it appears to newbies? "Welcome to Wikimedia! Hope you like drama!"

Strainu (talkcontribs)

It's a good thing the thread has died down, since the case itself was not worth it. However, this proposal is worth adopting.

ArielGlenn (talkcontribs)

This comment is in regards to the original amendment. Each venue should have a place where bans are listed, for the benefit of other administrators or persons who might be asked to lift the ban. So I would suggest that phabricator have such a place where the action, the target, the actor, the duration and (potentially) the reasoning are made available to phab admins. This does not necessarily need to be public.

There are two things being conflated in this discussion: 1) the need for other venue admins to know what's happening when someone is banned, 2) the desire of the public (contributors to Wikimedia projects including MediaWiki) to know what's going on. I'd rather deal with each separately.

GLavagetto (WMF) (talkcontribs)

I think you nailed the issue; personally I think the CoCC can't really be effective until problem 1 has been solved.

"Problem" 2 is not really a problem as far as I'm concerned - or at least it's something we are not in a hurry to solve.

MGChecker (talkcontribs)

I'm kind of irritated that most people here seem to believe that there is no need to diclose ban information to the general public. How are you supposed to work like that in a community-driven project?

Given the case that a user is blocked on Phabricator, this is visible to the general public and prominently marked whereever the account appears (marked with a dot or greyed out, depending on the form). If I see this happened to a user I worked with, how am I supposed to know what happened? Did this user just quit his MediaWiki work and left? Did he stop working for WMF? Or is he just temporarily gone? If I don't know what has happened, how shall I decide which actions to take, when I need to contact this user?

MusikAnimal (talkcontribs)

I do believe we should disclose bans. A block summary and a log on Phabricator would be most ideal, something more than just saying "disabled". I understand this probably isn't a feasible thing to implement right now, but I assume Phab admins can at least edit the "blurb" or "title" on the user's profile? You could link to the diff of the log entry that lives somewhere on mediawiki.org. This need extends well beyond CoC transparency. For instance, for a long time I collaborated with a particular user on Phabricator, and one day he disappeared. I saw his account was disabled and assumed he did something very wrong. Turns out, sadly, he passed away, which I didn't discover until months later. So we should log all blocks, regardless if they are CoC-related. I think the block summary should be as descriptive as possible, barring privacy concerns. If we need to be vague, "code of conduct violation" (or what have you) would suffice. I'll note that blocks, logs, internal drama, etc., are well-hidden on the wiki and I only hope the same will be true for Phabricator and other technical spaces. I'm all for transparency, but please, pretty please, don't use the highly-visible mailing list to announce or debate these actions, scaring off uninvolved contributors, or distracting them from their work. No one should be forced into this drama.

Tgr (WMF) (talkcontribs)
Any ban made by the code of conduct committee will be logged publicly in Code of Conduct/Cases/Log/2018 with user name, space of ban (phabricator, mediawiki.org, etc.), duration, and if not private reasoning. Unless any of these two conditions apply:
  • The reporter asks that ban not to be public.
  • Or the code of conduct committee decides not to disclose anything as it might disclose private information. For example, logging harassment cases in Wikimedia events can lead to real-world identity of users being disclosed.

Weak support, put I'd prefer to see a third condition there: the banned user asks for the ban not to be public.

Also I think there should be a guideline of how to handle bans in various spaces, as the log is not really useful if it's not exposed when someone tries to interact with a banned user. E.g. it should be linked from the MediaWiki block summary, the Phabricator user profile, the gerrit status message etc. Of course that's not part of the CoC amendment.

Reply to "Suggested amendment: Public logging of bans made by CoC"

Recommendation to modify the appeals process

39
Summary by Ladsgroup

This amendment is implemented.

Qgil-WMF (talkcontribs)

task T199086

Hi, the Technical Collaboration’s Community Health group wants to share some thoughts about the appeal process that we are currently handling. The text describing the appeal process itself is fine. The problematic part is to have an appeals team other than the CoC Committee itself without defining the relationship and governance between both teams.

The core of the problem is that it is not defined who has the ultimate decision on the resolution of an appeal. This is fine when both teams agree on a resolution, but what if they don’t? The options are

  • They have to keep discussing until there is consensus. This would put both teams on equal foot, which is fine but needs to be documented.
  • The Committee has the last word. This means that the Appeals team has an advisory function, which is fine but needs to be documented.
  • The Appeals team has the last word. This might even be the default expectation (?), but it is actually the most problematic one because it means that the Appeals team has more power than the Committee itself.

If we want to go for the third option anyway, then that Appeals body cannot be a team like we have now, formed by Wikimedia Foundation members by design. There were good reasons to make this choice (leaving tough situations to paid professionals, saving some trouble to volunteers), but having a team of WMF employees having more power than the Committee is a setup that we don’t want to have.

Strainu (talkcontribs)

The current text states: "These [appeals] will be considered by the Committee, which may alter the outcome." This suggests to me that the Committee has the last word. I believe this makes perfect sense, since the foundation should only override community-elected structures for legal reasons (in which case the Community Health group doesn't sound like the right group to make a decision anyway).

Bluerasberry (talkcontribs)

Can you link to the pages for each of these two committees or teams? I want to see a page for each, listing who the members are, and stating how anyone comes to be on these teams.

Based on what you say here and my browsing around I cannot quickly come to understand the differences in the nature of these two teams.

Huji (talkcontribs)

Can the auxiliary members of the CoC be the Appeals team? In which case I think option 1 above makes the most sense.

ArielGlenn (talkcontribs)

I'm not excited by having the auxiliary members be the Appeals team. Said as a former auxiliary member, I'd prefer to keep the function strictly as fallback in case of conflict of interest of active CoC committee members.

Qgil-WMF (talkcontribs)

An additional factor to be considered. The Technical Collaboration team doesn't exist as such anymore. The people who form the Community Health group are all active, so if we receive a new appeal we can still handle it. However, we would welcome a decision on our proposal.

Tracked: task T199086

MarcoAurelio (talkcontribs)

I think the Appeals team should have the final word on cases submitted to their consideration. Thank you.

Duesentrieb (talkcontribs)

Giving a non-elected (wmf appointed) team power to veto and repeal decisions my a community committee seems contrary to being a community driven organization. WMF staff should not have "benevolent dictator" powers in social processes.

Qgil-WMF (talkcontribs)

In order to help the discussion, I think two aspects should be considered:

  • Should the Appeals team be nominated by the Wikimedia Foundation or not? (and if not, how is this team nominated)
  • Who should have the last word, the Committee or the Appeals team?

The combination of these points offer four scenarios. A fifth would be that there is no Appeals team.

Ladsgroup (talkcontribs)

So far there are two things that can be seen here:

  • Slight majority believe a WMF-based team should not have power to overrule CoCC remedies. If there is no strong objections towards this by the next 7 days, I will make it clear in the CoC.
  • We don't have a "Community Health group" at WMF anymore. The functionality needs to be given to another body. I don't know WMF internal structure to suggest an alternative body in these cases. I reach out to people for suggestions.
MarcoAurelio (talkcontribs)

I still think that an appeals body should exist and be able to overturn a decision submitted to their consideration. Otherwise, what'd be the point on having one? It'd be bureaucracy for the sake of bureaucracy and a false appearance on the existance of an appeal process. If the problem is that we don't want to grant such power to a WMF Team for whatever reason, then I suggest that the appeals body be formed by community members instead in the same way the COCC is elected. Thank you.

Ladsgroup (talkcontribs)

I don't have any better proposal but making a committee just to check appeals seems too much overhead to me. There are several committees/group/teams we can delegate this responsibility.

Tgr (WMF) (talkcontribs)

The WMF operates most technical spaces, sponsors most development etc. so ultimately it is the WMF's responsibility to ensure technical spaces have a healthy culture. Having it as the decisionmaker of last resort makes sense.

OTOH if most decisions get appealed and some WMF team has to secondguess the CoC committee all the time, that seems like a bad situation. Rather than setting up another committee, I think it might be better to restrict appeals to situations where the committee made some objectively identifiable mistake (and then the WMF team's involvement would be limited to verifying that the mistake indeed happened).

Ladsgroup (talkcontribs)

We have cases review committee in WMF now which is reviewing T&S cases, would that work? I need to ask legal if that's okay with them. The team is also temporary and will be replaced by another team but something will replace it so we won't lose it. The problem that should it have the authority to override or not still stands. @Tgr: I disagree, I think delegating low-level cases to them would cause issues. There are cases that WMF needs to handle like T&S cases but they are on another league compared to cases we have with CoCC.

Ladsgroup (talkcontribs)

I quite like the compromise of these two bodies being peers (the first bullet point). What do you think?

Bawolff (talkcontribs)

what is the point of having an appeals body that is not allowed to hear appeals?

Martin Urbanec (talkcontribs)

It is allowed to hear them, it's however not allowed to overrule them. It's like senate's role in some countries - it is allowed to veto a law, but the lower chamber can overrule it - and force a law it wishes.

Ladsgroup (talkcontribs)

Or a similar case is FDC, they don't have the authority, they can recommand the board on fund distribution but they don't have the authority.

MarcoAurelio (talkcontribs)

I think the term appeal implies the existence of a superior entity with powers to review and confirm or overturn or vacate a previous decision, or remand the case back from where it came from asking them to re-examine. Reading Code_of_Conduct/Cases#Appealing_a_resolution I concurr with Quim that it looks like this might've been the original expectation although limited to the most serious of cases. If this is not what actually happens, I personally find the word "appeal" confusing and problematic because it won't be a real appeal but an advisory council to the CoCC. If the technical community feels the appeals body or process should be only advisory in nature, and not a real appeals body as IRL, I'd suggest to move away from the term "appeal" to remove the impresion that whichever groups fullfils that role can actually change the outcome of a decision by the CoCC, and clearly document that the CoCC has the last word. Thanks.

Martin Urbanec (talkcontribs)

Well, if those two bodies would be peers, then no one would have last word - both would be able to veto each other's decision.

Bawolff (talkcontribs)

And veto each other's vetos!

Ladsgroup (talkcontribs)

In paper yes, in reality I expect both bodies be reasonable enough to handle a disagreement.

MarcoAurelio (talkcontribs)

I feel terminology is important, that's all. If the two bodies are peers or expected to be peers (I am not saying this is good or bad), then no appeals process exist and the process should not be called as such IMHO. We could change it to 'review', 'advisor', 'dialogue' or any other word that more accurately describes the actual process.

> both would be able to veto each other's decision

That reminds me of the old it:Intercessio, but doesn't look like it's happening? I mean, if the CoCC has the last word, as the participants above seems to agree to (not arguing this is good or bad either), where exactly the so-called appeals body have veto powers? They could recommend but the ultimate decision would be yours so that's hardly an appeal, and hardly a veto either.

Regards.

Tgr (WMF) (talkcontribs)

The Universal Code of Conduct (expected to be out this year) will have its own enforcement and appeals mechanism. It's not clear to me what the status of the technical CoC is supposed to be at that point. Will it be replaced by the UCoC wholesale? Will they exist in parallel, with occasionally overlapping authority? Will they share appeal infrastructure?

I imagine these cannot really be answered before the UCoC gets finalized, so this might not be the best time for this amendment.

Qgil-WMF (talkcontribs)

Hi, nobody knows today how these pieces will play together, but a basic principle is that the UCoC exists to provide a baseline. Then specific projects can use / build their own structures on top of this baseline. This means that if the TCoC is compatible with the UCoC and it defines its own appeal process, then whatever UCoC structures won't interfere.

We have a very tangible problem with the TCoC appeal process, as I reported two years ago. On top of this, the Technical Collaboration doesn't exist as such since July 2019, and it is just a coincidence that the five original members of the former Community Health group are still working at the Foundation. The UCoC is in a very early stage, and there haven't been any discussions or decisions about enforcement and appeals yet. For these reasons, I would really welcome an amendment of this Appeals section in the TCoC, at least to reflect the current reality.

Tgr (WMF) (talkcontribs)

I would really welcome an amendment of this Appeals section in the TCoC, at least to reflect the current reality.

That's a fair point. So maybe go with that and state there is no appeals process? In your experience as part of the appeals team, was there ever a need for it? (As in, was there any appeal where the appeals team disagreed or found problems with the original committee's decisionmaking process?)

Qgil-WMF (talkcontribs)

In all this time we have received only two appeal requests. In the first one we disagreed with the appeal request and we agreed with the Committee. The second one is being processed right now.

The lack of appeals (a good sign) has been a reason for me to almost forget about this proposal during this time.

Martin Urbanec (talkcontribs)

I guess we (TCoC) would have the same position & authority as local arbitration committees - it is something T&S definitely needs to count with :).

Bawolff (talkcontribs)

But what is the point of this "appeals ls" process? Its hard to evaluate what to do here if we don't even agree on what role this process is supposed to play.

I strongly agree with Marco that the terminology is very problematic here. In my view it undermines the legitamacy of the CoC as an appeal is a well understood concept and outside of show-trials this is very much not what is meant by the word appeal.

Martin Urbanec (talkcontribs)

The intention is to allow for a review, since the decision is ultimate & binding for everybody (bypassing a CoCC decision is even violation of CoC per se). On the other hand, Arbitration Committees (which is the comparable body on [some] content projects) have no appeal body, and their decision can be only appealed to themselves (which is not really an appeal, right). So, we can also remove the appeals thing for good too?

Ladsgroup (talkcontribs)

If you think this is on par with local ArbComs, You can use their policy here too, for example: https://en.wikipedia.org/wiki/Wikipedia:Arbitration/Policy#Appeal_of_decisions

> Any editor may ask the Committee to reconsider or amend a ruling, which the Committee may accept or decline at its discretion. The Committee may require a minimum time to have elapsed since the enactment of the ruling, or since any prior request for reconsideration, before reviewing it. Remedies may be appealed to, and amended by, Jimbo Wales, unless the case involves Jimbo Wales' own actions.

Ladsgroup (talkcontribs)

Okay, I'd say let's rewrite the appeal section to make it like English Wikipedia's ArbCom:

Replacing:

After being notified of the outcome, the reporter or any people sanctioned may raise objections to the resolution. These will be considered by the Committee, which may alter the outcome. If the outcome is altered, the new outcome will be logged. When the Committee begins enforcing a decision, that is also logged.

Only resolutions (such as bans) that last more than one week may be appealed by people who were sanctioned. Reported victims can always appeal. To appeal a decision of the Committee, the reported offender or reported victim may contact the Technical Collaboration team's Community Health group at techconductappealswikimedia.org and they will review the case. Until an appeal is resolved, the prior resolution remains fully in effect.

To:

After being notified of the outcome, the reporter or any people sanctioned may raise objections to the resolution. These will be considered by the Committee, which may alter the outcome. If the outcome is altered, the new outcome will be logged. When the Committee begins enforcing a decision, that is also logged.

(Simply, removing the second paragraph). Does this sound good?

Jdforrester (WMF) (talkcontribs)

(Edited for readability.)