So, in keeping with This Meta.SO post, I figure I might as well mention the obvious:

Chat moderation tools are terrible

Let me describe a real event that illustrates this. I was in a room where there's banter. People sometimes take friendly jabs at each other and someone crossed a line. Someone mentioned it and the poster admitted they had a crossed it too. But the 2 minute limit had been reached and the concrete was hard. This left us in a lurch. Normally messages a Room Owner dislikes just get moved to another room (like Trash or a custom trash chat room), but that doesn't delete them. The only options left were

1. Flag it and have the user take the chat ban on the chin for 30 mins
2. Mod flag it (mods don't primarily moderate chat so this is often slow for a rolling conversation)
3. Hope the lurking mod in the room was paying attention

I'm a bit baffled by this, because chat is very much like Twitter in that it's a little bit of a "stream of consciousness", but lacks any ability to delete messages without a ban.

There's three things I'd like to see added here if we're going to add more accountability

1. Give us the same options for chat that comments have. Right now we have one giant flag for everything, and if you sustain that flag, the message is deleted and the user is banned for 30 mins per flag. Sometimes things just need deletion without a ban. We fixed this for comments, let's fix it here
2. Allow Room Owners to delete chat in their rooms. If we want internal policing, that must come with expanded powers for people we're going to hold to the fire. Internal policing can and should be the first line of defense here. If a RO sees things going off the rails, all they can do is temporarily kick people. Deleting messages would add more teeth and moderators can still see the deleted messages
3. Allow Room Owners to issue room bans for up to 60 mins. Can't behave? You can be thrown out without a moderator needing to do anything.

This sounds perfectly reasonable and I support this stance. Stack Overflow should not stand for abuse nor should it tolerate it.

Thank you for this.

Questions? Observations? Anything else?

Not once has a community manager asked chat users how we're doing during the 5 years I've been here, whether or not we're happy, if we're given the tools we want or need to effectively moderate. We get tools we ask for after several years maybe.

As an example: I've asked for some tools in order to be more welcoming. This was ignored for a year and then we were told there is no problem 3 years later by a community manager. ^{(I'll gladly provide more examples if you'd like)}.

I am certain that when racism, bad culture or other problems happen you are rarely aware of it.

One anonymous community manager* has been kind enough to show up from time to time to give us advice or tell us when we do things you don't like. Their advice has been helpful.

Let's have an open dialog instead?

If there are cultural issues you'd like to address with a certain room - I recommend we sit down and have an honest discussion about it.

I think Stack Overflow needs to communicate a lot more clearly to its community using the chat service.

Culture is built, not dictated

I'm going on a limb here and going to discuss the JavaScript room and the recent unfortunate interaction that resulted in this post.

To be clear - though it appears that the community at large agrees with the JavaScript room's ability to govern its on-topicness - the JavaScript room itself wants to welcome any efforts to improve its culture.

We've had members speak up against things that bother them, we've always been very open to discussion. During those few days no one from Stack Exchange talked to us. The only communication we've had on your side is on social media** and Shog's comment on an answer attacking us. (Well, unless you count the comment a community manager left on my answer and later deleted)

We've always optimized towards "let's not get Stack Overflow staff involved" because we mostly like you humans and we don't think wasting your time on a small subset of the site is worth it. We don't want to alienate you, frustrate you nor make the situation adversarial.

I would very much like to deescalate the situation as much as possible. We want inclusivity, we want to be more welcoming. That is not at odds with allowing "adult" off-topic topics like drugs and we're all up for discussion which we're doing internally but would happily do externally as well as explained in my postmortem.

* it's always Shog.

** We have decided to not respond on twitter to not escalate the situation.

I'm appreciative of this move. I've seen at least one room go beyond acceptable behavior and be closed down and, at the same time, I've heard of rooms that regularly have very constructive discussions about subjects that would normally lead to very problematic behavior.

While I'm not always in chat at all times, I tend to agree that when I see flags, they tend to involve the same rooms or users - some of the latter who have dozens of short (30-60 minute) chat suspensions. This seems a strong indicator that at least some of the problem lies in our (the moderators) response and (possibly unwillingness) to give these users the longer suspensions they deserve.

How difficult would it be to automatically increase the suspension length based on all suspensions over the last n time with time being greater than hours, which it is right now?

This means that repeated offenses in a week/month/year would be automatically recognized and the suspension would more directly correlate with continued poor behavior without the intervention of a moderator (in the case of 10k flag message deletion). It's worth noting that removing a chat suspension is pretty simple in cases of poor flag handling.

These users shouldn't be allowed to continue chatting but, often, the moderators may miss these events entirely as the only historical record of it is on the user's chat profile or on an admin page that many mods are likely unaware of that wasn't really intended for general use.

I really don't like putting more work on your plates but I think these things would help:

• improved auto-suspension escalation for repeat offenses
• improved moderation tools/chat flag history
• improved chat FAQ page (it's sorely out of date - I'd bet some mods would be happy to help draft it :cough:)
• better guidance for moderators on how to respond to repeat offenses from the same person/group of people

Thanks, again!

Well - I guess its about time. I don't really have that much newer info than I did the last time we talked about this three years ago.

The chat flag system is still a little odd; there's no real "out of band" way to keep track of problem sites and users.

One thing I think is an issue is how there's no real "framework" for what suspensions should be. Contrast the mod message system on main sites with how suspensions work on chat - with arbitrary numbers of hours settable.

In a sense, outside the rooms we're regulars in we don't really have context. A random moderator might not know a certain user has a tendency to kinda be on the borderline or even offend regularly. A mod who is a regular would throw the book at him. A non-regular one would need to go through the user's suspension record and decide.

And while it's not quite how things have worked lately - we kind of need a backstop. We do realise that our CMs are busy, but sometimes folks may need help from someone with a little more authority to sort it out. Sadly, effective chat moderation is instant or close enough, and it would be nice to have someone to ask folks to knock it off, or even hang out.

If there's one thing I've learnt being a RO and mod on root access - effective moderation starts with the community. For the most part we know what we need to do. Bad chat moderation usually ends in fire.

Also having been appraised of the situation - I do realise that it's a potential PR issue - but if a CM's going to talk about a chatroom on twitter, could we have someone pop by and talk to the room please? (If it has happened, well, awesome, but it should be a policy. "I've talked to the folk involved, and ______" would be such a nice thing to see). Communication and setting expectations is very important here.

Chat self-moderation starts with the individual user. We've been talking about ways to make moderation easier (and that's desperately needed!), about ways to empower room owners more, about automatic suspensions with teeth... but the individual user, the author of a chat message, lacks one important ability: after two minutes, whatever you said is permanent without moderator intervention.

We need to allow users to say "that thing I said was pretty stupid" (or didn't come out right, or was the scotch talking, or whatever), and let them clean it up themselves. We allow users to delete their own comments at any time, even if it might make other comments obsolete or puzzling; if they can do it with comments, why not in chat?

In comments Shog raised concerns about abuse; apparently there have been cases where people have done creepy things in chat and then tried to cover up the evidence. I think we can find a workable place between "locked in at 2 minutes" and "can delete everything and hide the bodies". Perhaps (just as a starting proposal to be refined), we limit people to five chat deletions per day and only allow deletion of messages within the last day. This allows the user who realizes (but not immediately) that he messed up to fix it, without opening the door to widespread deletions.

We could also raise a moderator flag if somebody deletes some threshold number (or percentage) of recent chat messages, like the flag we get for a sufficient volume of comment deletions. If somebody deletes the allowed five messages a day, every day, we probably want to notice that. This auto-flag isn't so that the deletions can be reversed; neither chat deletions nor self-deleted comments can be reversed by moderators. It's to let moderators know that there might be a concerning pattern to look into.

We could also log post-two-minute deletions in the chat user's history, alongside the flags, to make it easier for mods to see what users are deleting hours after the initial post. This would make it easier to review attempted coverups of creepy behavior.

Please tell us more directly what you want from chat behavior. Communication through concrete examples would make things so much simpler. Using examples doesn't mean giving an exhaustive list, it means giving people something to extrapolate from.

If a room is going in a bad direction, come into the room and tell people that there is an issue and explain what the issue is.

I would like to see concrete examples of behavior that have crossed the line. I understand what the rough intention behind the "be nice" policy is, but I don't know where the actual line is. Real examples would help a lot when moderating chat. If you want to keep the bad examples to moderator eyes only, that's fine.

I've only ever spent very little time in chat, so maybe I'm missing some important context, but I feel strongly enough about this that I think it deserves to be an answer, though it has been mentioned in the comments.

I'm distinctly getting the impression that this will be a one-strike, no-warnings policy. This seems fundamentally unfair. Certainly a chat room "going bad" is a gradual process, and there ought to be several red flags along that process where it would be natural to give mods a warning of "get your things together, or we're going to have to delete you". In this way, a room which might be beginning to display bad tendencies could be guided back towards good behavior, and the necessity to annihilate could be avoided. Of course, rooms which refuse to comply will have to receive their due consequences.

Other people have mentioned strengthening chat room moderation tools. Again, I really don't have enough experience in chat to attest to any lack thereof, but enabling moderators to do their job better would certainly go hand in hand with encouraging moderators to do their job better, rather than preemptively pulling the plug with little to no warning.

Aside from this, I'm definitely in favor of deleting chat rooms that refuse to comply with the Code of Conduct; I just feel we need to take care to strike the right balance and not be needlessly punitive where the necessary action might not be so drastic.

I appreciate this motive about being aggressive towards the chat users who are regularly breaking "Be Nice" policy and being problematic for constructive discussions in chat rooms. I also support action against chat rooms which attract flags and controversies on a regular basis. This also makes chat a useful place after main and meta sites. (Some users in the past said in their answers that chat is not much useful).

It is said that the rooms where code of conduct is violated will be permanently shut down.

Now what is offensive in chat is not as clear as it is on sites. It is depending purely on the luck of the flagger and fate of the flag on which moderator it reaches.

There is no better and uniform guidance to the moderators.

• When to take the policy literally?

One of the points in the Code of conduct is:

Name-calling. Focus on the post, not the person. That includes terms that feel personal even when they're applied to posts (like "lazy", "ignorant", or "whiny").

Name calling. It specifically black lists the word "ignorant" which may feel personal. Of course, this should not be taken literally when there are funny and hilarious conversations. This is about pretty serious conversations when one of the users leave the conversations without any word. I have read some conversations like that where there were repeated use of the word. I asked the moderator and they replied that the word should not be taken literally but with respect to context of conversation and topic. They didn't remove those messages saying they (mod) were not referring to the opposite person. It is fine to say "ignorant" within the context of that specific site and said that site specific rule applies more than general policy of Stack Exchange. So, I had nothing to say more. Left silently.

• When to take the flags with respect to context?

Bigotry of any kind. Language likely to offend or alienate individuals or groups based on race, gender, sexual orientation, religion, etc. will not be tolerated. At all. (Those are just a few examples; when in doubt, just don't.)

Not before the recent Stack Overflow blog post, I had another conversation where there were some personal comments about a user. They are not literally rude if we see individual chat messages but definitely problematic if read complete conversation. I flagged one such message without a second thought. The flag didn't survive a moment. I didn't flag any more messages because I know the result. Those messages survived. I had nothing to say to the user who typed those messages except "Be Nice". (Where in the chat is aggression except in your words and my thoughts? Those messages received stars instead of flags :/). Left silently.

Some Room Owners do not care about moderation

Some Room Owners ignore such conversations even took into their notice. Moderators don't visit chat often to check the history and conversations. Some messages look controversial but sometimes they say the reason of Room Culture. I also Room Culture should be taken into consideration while taking actions but how many times? Sometimes, it is repetitive but the action taken is minimal. Deletion or moving of the messages happen. Users continue such behavior again. Some only create rooms to talk about a separate and topic of the site. They don't care about moderation. This is also a reason that many messages are not flagged and brought to notice of a moderator.

I believe there will be many instances like that where there is no active participation of Room Owners and moderators in the chat. These conversations happen when there is absence of moderators. Chat flags are not effective all the time. Room Owners have limited tools to action against such conversations. So, to be honest, self moderation is not going on in many chat rooms where mods are not around. It is only going in active rooms where there are always 2 ROs and a moderator to check what is happening around in the room. I also believe that these incidents happen in some of those 97 % of the rooms. It may not occur regularly but this is not occurring once or twice in two months. If the team has taken the decision of 3 % of rooms, they should look at unnoticed 1% ¹ in the rest of 97 % also.

Lack of awareness of responsibility of a room owner

Some users are not aware of the responsibilities of a Room Owner. They do not know that they should also moderate the chat room in the absence of moderator and when there are some problematic messages in the room. This is also another reason for messages are not being flagged.

So, here are my requests:

• Uniform and accurate guidelines to the moderators on when to take actions on users.

• More tools to RO to take actions on problematic users. Removal of Room Owners if they they have not visited the chat for a long time. Like the system selects a new owner based on activity, there should be some process to remove them too.

• Community Managers and moderators should check flags and conversations of the rooms even with less activity.

• Update of Chat FAQ. The chat FAQ is still old. The ancient version of mobile chat updated to new version but the FAQ are still old and outdated. They need to be updated.

_{1. Percentages derived from Nog's brain from his experience on chat.stackexchange.com}

The concept of "normal user", "room owner" and "moderator" needs to be expanded into more granular abilities: (imagine file system per-user rights, it's a very similar concept)

• Room owners who can change other room owners' and users' abilities (so you can have some users with limited extra abilities and who can't change other users' abilities)
• Delete own message older than 2 minutes (up to 48 hours/older) — (automatically granted if you have > 10k rep site-wide, like being able to see chat flags, but may be explicitly allowed by a room owner)
• Delete any user's recent message in the room (up to 48 hours of age) — there's a case to be made about not so frequently visited rooms, but we'd have to collect more data from such rooms' regulars to figure out their needs
• Delete any user's message in the room

Deleting messages should be possible in ranges, same as with moving them.

Flags should only propagate outside the room or site after 3-5 minutes if no action is taken by users and if there are no online moderators who can see it in mentions and come into the room to figure it out.

I rarely use chat and I have no opinion on how it functions, but I do wonder about the language used here. Is it really necessary to enforce self-moderation aggressively? Could one not enforce it carefully or diligently or even cheerily?

I understand the word 'aggressively' is meant to convey the seriousness of SE's intent, but must that really be brought so ... aggressively? It might be a cultural thing: in American shows nowadays there's a tendency towards aggressive language, even in comedies people are all the time killing it or crushing or destroying someone. It gets the laughs.

But SE is not going for easy laughs here, just making a statement and giving it some extra charge by referencing violence. I notice it and it bothers me. "We're more actively enforcing self-moderation in chat" gets the message across just as well.

The problem I have with this, and I suspect part of the reason the rules have to be really vague, is that there are (at least) two completely different kinds of chatrooms.

1. "Third place" chatrooms. Long-lived rooms where people can go to hang out and banter. These places work best when they are welcoming, and it makes perfect sense to be a stickler about the "Be Nice" (and related) policies in here.
2. "Take it outside" chatrooms. These happen because two or more parties got engaged in a discussion/disagreement in question/answer comments, and some sensible person hit the "move it to chat" button. A large part of what triggers this is discussions getting acrimonious. Now obviously there are boundaries (eg: flat out slurs or name-calling), but otherwise the Be Nice rules need to be a little looser here to give the parties involved a chance to work out their differences and cool off in a less public place. As a mod, if I didn't have the ability to take such discussions "outside", my only recourse would be mass-deleting comments and locking posts so the deleted points don't re-emerge.

In short, I don't think the rules of behavior can, or should be identical for proper posts, Third Place chatrooms, and Post Argument chatrooms.

I realize reading over the question that likely what isn't being said is that some "Third Place" chatrooms were regularly getting content far outside the standards I outlined as appropriate for them under point 1 above. But part of my point here is that the need for things to be looser in Post Argument chatrooms is confusing things here. Perhaps if the software made some kind of distinction, it would be easier for us humans to do that somewhat objectively as well? There are some things that really ought to be functionally different for Post Argument chats too (eg: they really shouldn't get deleted as quickly, because then people who come late and want to join the discussion end up doing it back in the post comments. Arg!!)

I just need to drop my view cents here. I already wrote a comment:

Ah yep I can confirm that you close rooms in that case, unfortunately you suggest that it would be possible to reopen chats. However you didn't give a final statement for more then 6 month, that you will leave the room closed. That is how you lost my trust in community management. It is valid and correct to close rooms for good reasons, but don't imply that there are chances to change that if you already decided to leave a room closed.

I really want to give some more background for my comment. I was one of the room owners of “Android Era with Kotlin and Java”. In that room happened some shit which is not excusable. The majority of the room owners where informed that the room will be closed some minutes before the room was closed. It was implied that the room will be closed and if we would add some good reasons the room may be opened again. We never got a final answer; now it does not matter at all. In the end I am sure that the room was closed by missing moderation tools.

I'm sure that the room had been saved for good if there would be tools to enforce "room rules". Like no gifs, per room configurable ban of users which where simply not welcome in a specific room. A more intuitive explanation of the timeout feature. A official bot API would be nice for many reasons (e.g. for fun or documentation lookup).

In the end chats can be a good thing with the right people or can go to the dogs with the wrong users.

My brain requires rules to be black-and-white in order to have any hope of being retained. However, the rules of life (especially surrounding interpersonal contact) are often subjective, so in these cases I try to find a comparative simile that I can instead remember.

I've been using chat regularly for only a couple weeks, and I found myself forgetting that I was still on Stack Overflow, which I now attribute to the different "vibe" of the live 1:1 venue (as opposed to a forum where "paying attention to what everyone says", is the whole point).

At one point I realized I had posted a piece of personal data that should not have been shared. The member I shared the information with wasn't the issue -- the problem was that within 2 minutes ^†, it was now a permanent record in that chatroom (and potentially permanently Google-able) and thus becomes an indelible mark on both me and on the site we all work so hard to keep "a step above" other forum sites.

I contacted a mod to remove the item and while waiting I looked through my own history and was shocked at myself — politics were only part of the off-topic and/or inappropriate things I has brought up, intermingled with programming talk.

Anyhow in the end, a couple patient mods heeded my request to delete the room entirely, but the process made me realize that the "behaviour rule" actually is very black-and-white (for me, anyhow):

We must carry ourselves the way we would when in a Workplace.

(i.e. behave the same way we would in an office environment, or perhaps, a school.)

To me, this means:

• Small talk is not only allowed, it's encouraged. It builds comradery and makes the day more enjoyable. Humans are social creatures.
• By all means discuss "what you did over the weekend" or the upcoming episode of a favourite show.
• Joking, laughing, maybe an occasional silly post isn't a big deal — as long as we remember that these rooms are open and that anyone maybe within "earshot" currently, or in the future.
• The important thing in a workplace (or our chatrooms) is to remember the primary purpose of the room. We're all here to do a job, of sorts (or you could compare to post secondary school would be an alternate comparative example.)

Just like a workplace or school, if we want to venture outside of what's appropriate, we're perfectly able to -- elsewhere. Make plans to "meet" somewhere after "work" for the electronic equivalent of a pint of beer. That's a more appropriate place to "unwind and let loose", get raunchy, misbehave, and then recover in time to return to "work" the following day, reassured that "work" and "play" have been kept safely distanced, thereby protecting yourself and your "co-workers" from embarrassment or worse.

I'm generally not great with similes and I've pushed this one so I'm not sure if I'm properly communicating my point, but am I on the right track?

Related links for tips on how to carry one's self, etc:

• Stack Exchange: The Workplace
• Stack Exchange: Interpersonal Skills (Beta)

^† _{After posting a chat message, we unfortunately we only have 2 minutes to "take back" something we say if we end up regretting it. I believe this is not enough time for us to: Walk away from computer → realize what we said/did/posted → run back to the computer → find the message → edit/delete it. I have a meta question proposing an increase (or alternate rules like "room owner can always delete" for this reason.}

There are rooms that regularly and consistently violate this Meta post. For ages, nothing has been done to either those rooms or the posters there who get flagged.

Is this new policy going to be a universal rule?

Or only when the feelings of people who think like SE management politically get offended?

Forgive me if this has already been said: I can't justify reading 23 thousand words to find out.

But, doesn't this just give the trolls another weapon to cause trouble?

"Hey, Joe! Let's see if we can write a couple of bots to get an SE chat room deleted!"

While we're talking about chat flags, we also need a way to prevent misuse of the flags for simply advertising a message to every 10k user on the network.