Toward a philosophy of Chat

Question

TL;DR: The Problem

This keeps happening in chat:

1. Surprise at flags on vulgar messages. Language that would invariably get your comments deleted on the main site occasionally gets flagged and deleted in chat. Confusion reigns.
2. Controversial topics leading to bickering and name-calling. Folks bring up polarizing topics in rooms dedicated to other topics and filled with people who joined for the same. Anger and resentment ensues.
3. Moderators step in to try and resolve #1 or #2. Who elected them to solve our problems? Indignation follows.

These are inherently social problems; the first step toward addressing them is better communication. Therefore, I propose that we begin by re-writing the chat FAQ in an effort to clarify WHAT chat is for and HOW it can be used constructively.

I'll begin by laying out my own observations regarding chat...

Background

For five years now, Chat has been the red-headed stepchild of Stack Overflow and Stack Exchange: not a particularly good fit for a Q&A site, but still an integral part of the experience for many.

During this time, we've had some... Divergent evolution, so to speak: the perception of chat among the minds of those who use it has taken on different forms, and the perception of chat among those who don't regularly use it is something else again. This has led to problems.

Jeff envisioned chat as a "third place":

I think a web-based real time chat system like Campfire could offer that informal public gathering third place -- a space for people who love the topic to meet, discuss, and collaborate in a different way. It would foster community, and be complementary to both strict Q&A;, and meta-discussion.

This is essentially chat's charter, both the reason for its existence and the hope for what it would become. Over the years, folks have used a lot of different metaphors to describe chat: the watercooler (informal work conversations), the tavern (socialization after work), etc. These metaphors work, to a degree... But they also leak when stretched too far: you can't have 20 people all gathered around one jug of water, and rarely does anyone talk to everyone crowded into their local inn, much less expect them to listen and respond constructively. Yet, these scenarios are common in chat. And these forms of social interaction are the source of the problems described above, which cannot easily be resolved with fanciful comparisons to physical gathering-places.

I hold that the problems which arise in chat are rarely those of topic or language, although that is often how they appear; rather, the persistent problem in chat is participants who hold expectations for how they or others should behave which don't match the reality of either the system or the larger culture in which it resides. One does not repeatedly charge into a brick wall thinking it will move aside; one does so believing that the wall does not exist.

The actual nature of chat on Stack Overflow / Stack Exchange

These are incontestable facts:

1. Chat is not IRC: conversations are permanent by default, public by default, and linked to your main-site user account by default.
2. Chat is user-moderated: flaggers, privileged users and elected moderators all participate in deciding what is allowed and what is not.
3. Chat is moderated: there are no "anything goes" chat rooms; all moderation tools and privileges operate in all rooms.

These facts imply a few things that are not always obvious:

4. Chat is not transient: what you say tonight while drunk still exists, in public and attached to your account, tomorrow.
5. Chat is not separate from Q&A: you have to participate on the main sites to establish an account that even allows you to chat; you earn the chat privilege on main, and can lose it there as well.
6. You can't control who reads what you write in chat: transcripts are completely public, open to anyone, even people who aren't visible in chat, even people who aren't members of the main Q&A site. Your boss, your wife, your worst enemy, that annoying kid from 6th grade... They could all be reading your chat logs at their leisure, potentially even years after the fact.

Taken as a whole, these factors make chat well-suited as an auxiliary to the asking and answering that happens on the main Q&A sites... But poorly suited to many other roles, even roles that other chat systems commonly serve! I believe it is essential that we communicate these factors within the guidance given to chat participants, and do what we can to dissuade them from relying on chat for purposes they will find it ill-suited for.

Communicating the nature of chat

I said at the start that my primary goal here is to re-write the guidance that is given to folks using chat; indeed, several of my co-workers are already hard at work on this. As such guidance must necessarily be grounded in a shared understanding of the nature and purpose of the system itself, and that starts here: I've laid out my observations above; now, what have I overlooked?

Note that there are several outstanding technical issues with both flagging and room membership... If you're interested in these issues, please participate in the relevant feature-requests.

Answer 1

I guess I'll expand my comment into an answer, because there's a lot I didn't say.

On chat, I see a lot of rooms with "regulars" - a small group of folks that have been in the room since its conception, and continue to play major roles in the dynamics of conversations there. They often have personalities that get along well together, and so they agree on certain things - including the room's informal "policies".

Over time - weeks, months, years - the standards of behavior gradually align with this group's expectations. They're the major players in the dramas of chat, and so they set a standard. New users see that these regulars are the main folks here, and so they might change their behaviors accordingly. Without outside intervention, you very quickly build up what's called "room culture".

The phrase "room culture" is generally used defensively. When a moderator or other flag-handler deletes a chat message, suspends a user, or otherwise acts in a way that the regulars don't like, they can be attacked for it. Worse still are cases when the moderator isn't a moderator on the site. They'll then be called an outsider, someone who just doesn't understand the "room culture".

I think that at least part of the problem here stems from this idea, and I've seen certain recent issues on chat parallel this pretty closely. It all comes down to an attitude issue - and I know that that sounds harsh, but it is the truth. The attitude that some folks are above the overarching chat policies and can make their own rules because they're established is one that can cause problems.

What I'm not saying:

• I'm not implying that room culture is always a terrible thing. Different sites have different personalities, and so by extension different rooms have different personalities.
• By extension, I'm not saying that the idea of regulars is a bad thing. Having folks who know the main site well and are available to help the new contributors is always great. I've walked into many sites where I have no clue what the standards are, and between observing the main site and meta, and asking the experienced users in chat, I've learned about the site much more.

What I am saying:

• Over-the-top, offensive, and otherwise inappropriate room culture is not okay and should not be tolerated.
• Exclusive cliques, while uncommon, are not good, because they can give new users poor experiences on the site.

I don't have a fix for this issue at the moment. Trying to eradicate this is like trying to eradicate bigotry - it won't work. But we need to have measures in place to prevent these users and these attitudes from harming chat and Stack Exchange.

Answer 2

Chat isn't moderated.

I hate to put it this bluntly, but I have to challenge your assumption. Yes, there are moderators, and yes, there are flags, but these tools don't do much to solve real problems in chat. They don't help keep things in check.

Let's look at the tools we actually have:

• [Anyone] Our voices - Look, talking with people is great, but once a situation has started, it's awfully difficult to wind it down. This is equally true on IRC and on Stack Exchange chat, and of the dozens of times I've been tasked with cooling things down, they rarely succeed.

  Even if they do succeed, there's a very long period of tension that evolves as a result, and it leaves a stain on the local culture of the room.

• [Anyone] Casting flags - This is actually detrimental to an evolving situation. It prompts all active 10K users and moderators to poke in and see what's up, which, to the users in the room, only feels like prying from outside. Any attempt by many people to adjudicate a room's proceedings from the outside is an affront to the members of the room, and in truth, this is a response that actually makes sense.

• [10K, Mod] Approving flags - This deletes the message but doesn't solve the underlying problem, making it effectively totally useless in the long term.

• [Owner, Mod] Room timeouts - "Cool, our room's gone. Where should we go next? Hey, how about we make a new room to express our bitterness that some mod just ruined our fun?"

  Nobody's behavior is changed by room timeouts.

• [Owner, Mod] Kick/mute - See suspensions.

• [Mod] Room freezing - See room timeouts.

• [Mod] User suspension - This is a big one, and one could say it's the last resort - the ultimate chat moderation tool.

  Unfortunately, suspensions are useless.

  Here's the thing: automatic suspensions go in increments of 30 minutes. Suspensions by a moderator rarely last more than a couple hours to a few days. This doesn't solve the problem at all. It's a thwack on the wrists to users who are used to thwackin'.

  Moderators for IRC recognize a simple truth about chat: once a user has clearly identified that they are a problem user, it is very difficult to change them into a constructive one. Obviously, it's a generalization that has its exceptions, but it still informs a stricter moderation policy. On IRC, you get close to zero - often actually zero - warnings before you are permanently banned for disruptive behavior. Why? Because the mods know you'll be back, and they know temporary bans rarely work.

  Your post makes this point even more stark: on IRC, things are transient and vanish. On Stack Exchange chat, messages are permanent - and yet, we let people talk so freely without fear of real repercussion.

I realize what I'm saying goes against much of Stack Exchange's philosophies. From A Day in the Penalty Box, it's been clear that Stack Exchange's primary aim is reform, and not removal. I agree with this goal for Q&A - people have a given focus, and if you can guide them back to that focus, then both they and the site are much better off. Often, all it takes is a gentle nudge.

Chat doesn't work the same way. Right now, we're seeing the consequences chat's distinct lack of moderation, simply because the tools we've been given don't do the job they're supposed to.

If I were in control of making these decisions, I would implement a relatively stringent chat suspension and ban policy - something that allows users to actually moderate chat. Users who are problem users need to be kicked out for long periods of time - perhaps, in the worst cases, forever - to get the point across. I hate to put this all quite so bluntly, but it's the simple truth of chat.

Answer 3

Chat in the physical world

To me, chat is like the group discussion that congregates spontaneously outside the trade show conference room… or a social gathering which has tapered off into post-occasion hangouts. You don't generally have someone external "in charge" of the conversation, and the party host doesn't usually come running over to tell you every time you're doing it wrong.

That's the fallacy of "Moderators will watch over chat..." (when they're around)

"Real" chat isn't centrally moderated

Small social groups don't operate under oligarchic control. Chat isn't like the carefully-orchestrated behavior of an SE Q&A "publication". The norms of a chat room are determined in large part by the participants, and those social cues have to come from the group.

Of course, many of those social mechanisms break down in a virtual environment. Folks either miss the subtle cues and either fail to change course or go elsewhere… or they willfully become spoilers emboldened by their anonymous persona. Moderators can handle the spoiler case when things go wildly out of control, but calling in the authorities every time some blowhard gets too loud or someone starts veering off topic doesn't help them not get there in the first place.

"Light annoyance" isn't generally well-expressed in a chat environment, and it is easily misinterpreted and blown out of proportion. Bringing that disapproving glance into the public conversation can cause a disruptive backlash or trigger a piling-on effect that isn't typically worth raising. That's where the core system can help.

Crowd-based Chat Norming

Not every conversation-spoiling activity is cause for a capitol action. Sometimes all you need is that virtual librarian who whispers "shhhh..." when you're getting a bit too loud, or the well-meaning compadre who says "not cool, dude" when you've gone a bit over the top.

The chat system I built many moons ago had an integrated, lightweight "poke" feature. It wasn't meant to be punitive. Occasional pokes and ribbing are part of any social dynamic, and most can be laughed off (and aged away) as a friendly gibe.

But it is very effective.

Occasional pokes generate friendly (very friendly) private reminders to "be nice" and stay on topic. Repeated poking can escalate to increasing warnings/guidance until ignoring them is done so at their peril.

(I hesitate to get into a further implementation conversation here; it will only derail the point of this post)

Not everyone has an intrinsic "right" to be in a conversation. Using a crowd-based chat control, someone repeatedly disrupting the group dynamic can lead to brief timeouts; timeouts lead to blocking; blocking leads to bans. It's all very transparent (by the UI), but that doesn't preclude moderators from jumping in for the occasion "exception handling" (repeated warnings from the group dynamic can eventually escalate to a Moderator, and overtly hostile actions need a swift response).

Obviously the UI has to handle abuses, and the UX of poking a user (and the tone of what those actions entail) have to be meticulously considered.

Why does this work?

Even social discord can be a healthy way of building cohesiveness if expressed properly. Social systems harness the power of crowds to create self-regulating systems. But those means have to be expressed properly, and they have to work from the bottom up to create socially-pleasant macro-behaviors. You don't call in the singular authority when things start to veer off course.

I love all the talk about improving chat guidance. But I've had difficulty getting across that over-dependence on Moderators and "read our FAQ" for everything we didn't get around to building into the system is a problem at all. That's why I'm not delving too far into implementation issues of empowering the community more broadly… but taking this opportunity to raise it as point of broad consideration here.

Answer 4

One problem with vulgarity in chat is that the rules that officially apply, the no profanity standard of the SE network, aren't actually enforced. A lot of profanity is tolerated in chat, but if someone flags it, it is possible that it'll get removed. There is a disconnect between the official rules and the actually enforced rules, and that results in almost random decisions as every moderator or user handles those flags according to their own view.

A certain amount of cursing is tolerated in many rooms, among them also the moderator-only Teacher's Lounge. I don't think this is a problem, and I think this roughly mirrors the real world where the same language would also be often acceptable in some less formal, but not entirely private situations.

Having a tool to deal with inappropriate language that doesn't also suspend the user would be useful as well. It just escalates the situation and pisses off the users when they're suspended for misjudging the language a bit.

I personally find the profanity part to be a huge distractions from the actual issues of chat. So I would like for the official rules to be adjusted to the reality of chat to remove this distraction, while still not allowing just anything in chat.

I also don't think you can separate the social problems in chat from the technical issues. The flag system tend to escalate the drama in chat by drawing a lot of attention to each minor issue. Simply fixing this part might help with all the other issues as well. Flags are seen by many users as mostly misused, a validated flag isn't seen as confirmation that a message was problematic, but as the action of some trigger-happy outsiders.

Another issue is that chat moderation can be much harder than moderating the main site. It's in real time, you can't just take all the time you want to read up on the history of an event. There is also pretty much no guidance at all from SE on how to use the various chat moderation tools. I personally strongly dislike using the room timeout, but many other moderators use it as the first tool. Suspensions are mostly trivial, and rarely used outside of serious offenses.

Answer 5

Just wanted to add my two cents addressing the first problem you mention, and a counterpoint to those suggesting comment-style moderation or lists of offensive words -- too long for a comment:

Surprise at flags on vulgar messages. Language that would invariably get your comments deleted on the main site occasionally gets flagged and deleted in chat. Confusion reigns.

We definitely need to have a solid definition of what should happen in those instances.

^{I use some language below that some consider obscene, for examples. Please don't take that as my support for or opposition against any specific words - they're just random examples that turned up reasonably often in chat logs.}

As said in the comments, this is a really hard problem. The definition of what counts as 'vulgar' varies widely depending on which country you're in, your socioeconomic status, religion, heck (!), even which generation you're in. It's a very subjective topic.

This same argument has been made for the main site. There is one significant difference: chat, by its very nature, tends to include a lot more colloquial language, some of which may be considered vulgar by some people but not others. In chat, we often write how we speak. On main sites, we take more time to write answers and comments, usually with a very narrow focus on the question at hand. I don't think it's reasonable to expect the same level of care from chat, and expect people to avoid colloquialisms.

Some people have proposed a list of language considered vulgar, so everyone can follow it as a rule. Who decides the list? What's the cutoff? How will the list be updated as language changes? What's the benefit in banning specific words but not euphemisms and slightly-censored variants with the same meaning? "shit", "sh!t", "sh*t" -- should they all be banned? Is "ass"/"arse" considered vulgar? Some might consider "hell" or "goddammit" to be extreme violations (blasphemous), while others don't care at all.

As a starting point, I would suggest that words themselves should not be banned outright but rather considered in context, especially the intent. There's a world of difference between, say, "win10 is being a bitch" (actual quote from chat), "she is a hot bitch" (another quote from chat), and "emo fat bitch" (another quote from chat ... wow.). The first is colloquial language that carries no offensiveness apart from (minor? subjective!) vulgar language -- and IMO the language itself should not be banned. The second and third can be considered offensive for the subject matter, and because they are targeted towards other people. I would consider attacks against other people offensive regardless of language used -- whether you call someone "an idiot" or "a fucking idiot", it's still equally offensive.

Yes, this brings us back to subjective territory. But I do not see a good way around that. At best we can define specific topics as unacceptable (though that again runs into issues with blanket-banning).

Answer 6

A few modest proposals.

Firstly, there's really two sets of issues here: the errant user and the errant room. And a few linked problems.

I'm starting to feel that twenty reputation is not enough when we have folks wander in from unrelated sites. An option to either restrict a room to folk who have 20 rep on the parent site of the room or a +m style higher rep threshold when there's trouble would deal with trouble users on non trouble rooms.

Secondly as a former room owner, and a current mod who's active on my chat, there's no clear guidance on what the appropriate actions for things are.

I'd like to suggest two things for general moderation and a few for how rooms are handled:

1. We need a theory of moderation for chat. Basically an official document saying "Hey, as a room owner, we expect these of you, and we trust you with these powers". Soft skills are hard problems, and having something that says "hey, here's some official guidelines" would be awesome.

2. Have the equivalent of mod messages and the progressive suspension system for chat. The current "manual suspension by hours" system is tricky, and not quite handy.

There's also the issue of "foreign" mods. My views have changed - while my users are not that enthusiastic that it had to happen, they actually did a good job. That said, having something tell me "Hey. Someone from another site had to suspend someone on a room associated with your site" would be nice. Typically my users let me know but its not quite efficient. That way I can get back to the mods and have feedback, or have a talk with the user myself.

Rooms are a whole different issue. We have some rooms that are completely problematic and are just complete trouble. We have a few that are a work in progress. We have a few known troublemakers.

All this would tie into something I think would handle rooms better.

I'd like to suggest letting mods annotate rooms. Automatically add room freezes, and let us stick messages like "hey, I noticed that X has a tendancy for making statements that are meant to inflame other users" or even "Y is on his last warning. Feel free to suspend if he insults Z". In short, let mods communicate out of band, and have a audit trail for mod actions through that. This might also help handle the 'site culture' and 'tone' issues people complain about, and as a mod only feature, lets a mod be responsible for these things.

Answer 7

I want to focus my answer on an aspect of chat that is important to larger rooms, but has very little useful power: Room Owners. In Emrakul's answer, it's made clear that the room owner has very little power - suspending a room or kicking a problem user. This is like the 15K privilege. It's a milestone. But...meh.

In larger rooms, Room Owners are often selected via some process. Some rooms have it spelled out, some do not. In most cases though, the room owners are selected because they are trusted members of the chat room community. They are expected to be able to keep the room from exploding into a mess of flags that draw in 10K users, moderators or Community Managers. But, they are given two large hammers and often little back up that doesn't show up with even larger hammers. The position is given because the community respects the people and their behavior, but now that they are in the position, there isn't a whole lot more they can do to help the room.

I think Room Owners need more support. Moderators across the network have the Teacher's Lounge and that allows us to interact with the community team on a regular basis. It's helpful for many reasons:

• Moderators can talk with one another about issues on their sites
• Moderators can talk with the community team regarding growing issues on their sites
• Moderators can learn from other moderators

Room owners don't have this type of support. They are isolated to their room and have the support of the other owners of that room. I think this needs to change.

Room owners of large rooms need a place to talk with one another and moderators. The idea would be the same as TL. Room owners can see what is and isn't working in other chat rooms. They also have a more direct line of communication to the site moderators. If an issue starts to fester in a specific room, the room owners can request moderator help. The other direction is also possible. If a room is generating flags, a moderator can summon the room owners and ask what's going on.

I do not have a threshold for which room owners this would work for. I do believe this would be most beneficial for rooms that get a lot of traffic though. One off or short lived rooms probably aren't the target of many issues.

Answer 8

Flags:

I have seen my share of flags, and been the - let's call it the "flaggee" - several times, and on a couple of occasions, I was the "flagger". Here's my take:

There are several kinds of flags:

1. Spite flags: "You annoyed me, so I'm flagging your chat comments to get back at you". So rare as to be irrelevant.

2. Legitimate flags on unquestionably offensive material: Almost no one could deny the validity of such flags - someone is saying something that is obviously vulgar (beyond mere naughty words), bigoted, insulting, etc. Surprisingly or not, these are also relatively rare.

3. Flags on things deemed "acceptable" in that room, but which are actually offensive to a large number of sensible people: These are among the most common, and the most problematic. When someone says something that the room's regulars deem par for the course, and it is flagged, any attempt to handle the situation is bound to cause a furor.

   In these cases, the argument is invariably something along the lines of "that's normal for us, so it shouldn't be flagged. Everyone knows that I/he/she/we is/are just kidding. If you don't like it, you should go somewhere else." This reveals that the problem is two-fold:

   • Person(s) make offensive comments. Usually, the person(s) in question represent a tiny minority of the room's participants. This would make it easy to handle, if not for...

   • A large portion of the other participants in said room then defend the person(s) who got out of line and said offensive things. The flag draws attention, people (regular users and mods) come in from other rooms to see what is happening, and the room turns on these newcomers with guns blazing. They are trolls, they are exacerbating the problem, etc, etc. Most of all, they "just don't understand that that's the way this room is, and it really isn't a problem".

   Room culture is fine, and cliques in chat are unavoidable, but both become untenable when the circle closes on itself.

4. "I don't like what you said, so I flagged it" flags: I've raised one or two of these, and had some raised on me - my "home chat room" was recently closed for days following one such incident. I don't know how these should be handled. In our case, we're adapting our room culture to avoid similar incidents (and the topics that lead to them) in the future. I sincerely hope it works. On a personal level, though, the lesson I took away from the incident in which I raised one or two of these is "reach for the 'ignore user' button before the 'flag' button".

Suggestions:

1. Guidelines/Revised FAQ: I think the most important and common sense guideline is already in place, but could use some reinforcing: "No bigotry of any kind, even in jest." Recent problems on one SE would have been avoided by a more thorough acceptance of this "Be Nice" maxim. Instead, mods who entered the room to remind people to "Be Nice" were accused by the room's regulars (including mods and room owners) of trolling, and afterwards, some of the mods on the associated site continue to insist the flags were the result of grudges another SE held against all mods.

   A less rigid restatement regarding vulgar language could also be helpful, although the question of what constitutes vulgar language is highly subjective. Automatically censoring the most obvious naughty words might help, but such scripts are always tricky to implement.

   As mentioned above, sometimes a room is very quick to turn on "flag chasers" - people who are in other rooms, see a flag, and come to check it out. The FAQ should mention that when a flag is raised, this is an inevitable result. It isn't an attack on, or invasion of, your space. It is how many people react in the chat format, and chat rooms aren't yours - they are ours.

   In a similar vein, the FAQ should explicitly state that the expected chat conduct standards are universal, not unique to each room. Don't base your behavior on what the people in your room have come to expect and accept; base your behavior on how a new user would react to what you're saying. Stop with the crap about "This is how our room is, if you don't like it, go somewhere else", and ask yourself "Would a significant portion of the population take serious issue with what I'm saying?"

2. Reworking the flag system: As on the main site, your reliability (i.e., the validity of flags you've raised in the past) should determine whether or not you can flag in the future. Also, it might help to have chat flag reasons, as we have for flagging Q&A on the main site. "Bigotry"/"Insulting/Disrespectful"/"Vulgarity"/"Other: Fill in the blank (required)", etc. This could be used to determine how to handle the flag based on why it was raised.

3. Improved handling of problem users: I was a problem user, and I speak from experience. A warning will sometimes (often? usually?) be enough to drastically improve a user's behavior in chat. I went from "You keep causing problems and you're flirting with a lengthy suspension" to "You're a genuine asset to the room", in the words of a CM (slightly paraphrased, in the first quote), in a matter of weeks. This can and does happen. You just need to let the problem user decide whether to accept the warning and stop being a jerk, or continue being a jerk and take a lengthy vacation from chat.

4. Quick action in a situation that is getting out of hand: When a conversation turns into a heated argument, and the heated argument turns into a bunch of chimps flinging poop at each other, mods should freeze the room, delete everything said from the start of the bickering onward (to stop the tit-for-tat starring and sniping that inevitably follow, and often motivate the participants to return fire), and issue a general warning to everyone, not singling anyone out publicly, to calm down and drop the subject.

   And one of the worst things a mod can do when a situation has already devolved into furious bickering and flags is to say "Stop flagging that". It comes across as saying "If you're offended by that, your feelings are wrong and inappropriate, and your opinion and offense don't matter". If multiple people have already said "That's offensive" by raising a flag, the comment is clearly offensive, whether or not you share in that offense. No one likes having their feelings dismissed out of hand, and they will often react badly to such dismissal.

5. Crack down on recidivism: Once a user has been given a warning, ignored it, been briefly suspended, and returned, that user should know they're on probation. The first suspension was probably 30 minutes; the next one should be a week. Week ends, problem behavior continues, suspend the user for a month. Still no improvement after a month? See you in a year, or perhaps, never again. I am in total agreement with Emrakul here. I'm not just blowing smoke - I just got done with a week-long suspension, so I'm saying "If I get seriously out of line again, boot me for a month".

6. Objective handling of problem users: It is inevitable that good mods will develop a strong rapport with regular users; it is inevitable that if you like someone a lot, you'll be tempted to give them the benefit of the doubt and handle them gently. When a regular user is also a problem user in chat, perhaps their case should be handled by other mods, who aren't familiar with them and can see their behavior with eyes wide open.

Answer 9

I think one aspect that has been overlooked is the divided nature of the existing chat systems. Currently, there are three separate chat environments - one each for MSE, SO, and the rest of SE.

I think that divided nature has helped foster some otherwise niche behaviour that wouldn't have been able to take root had their been greater visibility. I bring this up because new guidance is helpful, but guidance requires governance in order to become habit.

One particular challenge that the divided nature of chat creates is the number of trusted users or moderators who can keep an eye on activities and help discourage unwanted behaviour.

Put into numbers¹, we have:

• n diamond moderators (SE employees) on chat.meta across 15+ rooms
• 18 diamond moderators (SO mods) on chat.so across 30-40'ish rooms
• 400+ diamond moderators on chat.se across 80-90'ish rooms

And while I'm not claiming that chat.se doesn't have some of the problems that rooms on the other chat sites have experienced, those problems tend to be more quickly resolved simply based upon the sheer number of moderators available to help fix things.

So one thing I would suggest considering is unifying all of the chat systems. It's akin to how we handle navigation across SE sites - whether I'm going to SO or Programmers or Engineering, it doesn't matter as it's all part of the Stack Exchange system.

Unifying chat would help resolve some of the issues that the divided chat systems inadvertently foster.

And, if unification is not a feasible option for whatever reason, perhaps allow users to affiliate their chat profile with the main site profile of their choice. I don't think this would be as effective, as it will require 10k+ and moderators to be signed into both systems, but it would at least allow our core experienced users to help provide governance.

¹ _{Numbers were based upon a quick check of number of rooms at each chat site as well taking a swag at the number of moderators across SE}

² _{Please also note that chat room numbers fluctuate a bit. When I first pulled counts for this the numbers were different than where they are now. Also note that miscounted the number of SE sites and my initial guess on SE mods was low.}

Answer 10

The flagging system is kind of terrible

Just my two cents' worth. Note that I'm coming from mainly chatting on SE; MSE and SO chats are probably different because there are drastically less moderators.

When a chat flag is raised in a SE chatroom, hordes of 10k users and moderators come to check out the problem, when often the moderator(s) who are already in the room can manage the problem on their own. It's not so much "invading" (though I have jokingly referred to it as such), but it is a drastic overreaction in most cases. This has been brought up a million times already, so what's one more?

• Chat flags should have a delay in broadcasting to the entire network, so that moderators/10k users currently in the room can deal with it first, to prevent unnecessary swarming of the room by other mods/10k users
• Only if a chat flag has not been attended to after a certain period of time (1 minute? 5 minutes?) should it be broadcasted to network moderators and 10k users

An image says a thousand words:

Answer 11

I think clearly defining the purpose and nature of chat Stackwide is an excellent plan that will help moderators as well as community understand what is and isn't appropriate, thus avoiding some future drama.

As we are considering this, it's critical to take into consideration that chat transcripts are a permanent record.

As Matt Ellen and others have said, we can't forget what is still in the record. However, we do forget when we are chatting that everything we say is recorded for posterity. Our thoughts and interactions are ephemeral, but the evidence is permanent.

The reason I bring this up is that we have community members who will search the transcript looking for things to flag. It is pointless to flag something that someone wrote three weeks ago and have them auto-suspended for 30 minutes and yet it is possible for that to happen. It's also possible to be suspended from chat when you aren't even there. This is disruptive and not helpful in the context of what flags are supposed to do.

On the other hand, there are things that should probably be removed from the transcript. Right now, a chatter could use a mod flag to request that a comment be removed, but this never happens.

Here is my suggestion:

Offensive/spam flags only need to be broadcast/escalated if the flagged message is immediate within say, a one hour or 200 message span or something. The reason that flags are broadcast is to get eyes on them so any unfolding situation can be dealt with swiftly. If it's older than an hour or so, it's extremely unlikely that it is sufficiently urgent to wake the entire neighborhood. Otherwise, flags are queued for moderator/room owner review. Approving flags on anything old removes the chat message from the transcript but does not issue an automatic suspension.

This has a two-fold impact:

1. Behavioral correction can be applied immediately and in context.
2. Spiteful counterflagging is mitigated because searching the transcript to target a particular user is less dramatic.

The number of flags/approved flags on chat users is monitored and chat suspensions are issued as necessary (I read this in one of the answers and thought it was a good idea, can't find it right now) like we do on the main site.

Also, there is a lot of focus on what words should be allowed in chat. I disagree with this plan. The words are irrelevant; it is the intent that is flaggable. Therefore, I propose that there should be more than "offensive/spam". I would like to flag statements as "rude/unwelcoming" and "picking/extending a fight" as well, which would help everyone understand the context a little better when they come tromping into my chat room in the middle of the night. At least it would make it more understandable when I see a flag on "How is your garage coming along?" -- that's not offensive, unless I happen to know that the person who said it is trolling.

It would also help to have an official policy about cleaning the starboard. I clean mine of single stars regularly, but I also try to remove stars on inflammatory statements that are designed to continue the drama rather than quell it (though I usually wait until a day or so has passed).

Answer 12

Okay, so we're pretending this is about all StackExchange chat rooms

Hi, I'm Tristan. I'm a room owner over in the Python chatroom. We affectionately call the community "SOPython."

The room owners in SOPython put in a non-trivial amount of work into preserving our community and I want to let you know that as I'll be speaking from my experiences (and specifically not speaking on the behalf of the experience in other rooms that I don't frequent).

Changes to chat pertain a great deal to my interests and I appreciate the effort and emotional investment that people are spending on discussing it.

Anyway, here's a bunch of words from a stanger on the internet:

Creating a room is free. Building a community culture is expensive.

Do you think that a particularly room culture is toxic? Do you think a room culture is a negative thing? Make a new room.

I'm going to tell you a tedious story: Near my first apartment, roughly a decade ago, there was this crappy bar. It was overpriced, it smelled bad, it was dirty (not in a charming way), and it wasn't much fun to go there. But went there I did, because where else would I go to hang out?

I recently drove by and saw that it was replaced. Why? Probably because a new bar opened a few doors down the block and it was more popular. People preferred it and some of the regulars switched. Or the place attracted new regulars.

I'm not saying it's going to be easy to start a new room. It takes a lot of effort to attract users and keep enough around to establish a community of regulars, but it's far more productive than demanding that an existing community change to fit a myriad of changing, conflicting, and vague desires.

A room culture is good, necessary because it helps keep the community a desirable place to be, even when moderators aren't actively watching. Sometimes there isn't a moderator in the SOPython room. Does the chat descend into XBox-Live-style insanity? Nope.

Please do not go down the endless path of defining text content as offensive.

We don't need a moderation charter or a list of no-no words. It would be exhausting and difficult to compile; the end-result would be something daunting to enforce and the community would turn into a dried up husk of its former self.

Further, demanding censorship or trying to protect others from reading words they don't like is incredibly toxic. Offense is taken, not given. If we made a list of naughty, double-ungood, don't-say-these, words for StackOverflow, would I allowed to say "oh god, this regex is a rat's nest?" What percentage of users can state it's offense and demand I self-censor? When does it end ("trigger-warning: regex")?

Does that mean that we tolerate bigotry/sexism in SOPython? Hell no. I'm pretty sure that I was made a room owner because I have a low tolerance for bullshit. That and I gave Jon Clements £20.

Site moderators are elected by popularity, room owners by other room owners. This system can/does work.

Programming languages and related subcultures differ in their behaviors and communication. Site moderators are being asked to preserve the space that all of the subcultures share. Like a very popular janitor in a school cafeteria. He/she cleans up messes, but also is there to break up fights, stop bulling, and, when things get out of hand, can lay down the rules (as they are defined by his/her employer).

Room owners are like that cool math teacher that let us eat in their classrooms during lunch periods (just me?). We'll clean up some garbage and help things not "get too out of hand" with the room occupants, but ultimately, we don't have the same tools for cleaning up as the janitors.

We have a SOPython teacher's lounge. This is where we smoke cigarettes and drink due to the stress you people cause us give each other feedback and discuss how to productively perform our roles or deal with problem students.

In some cases, such as with SOPython, those teachers are also janitors (either this metaphor fell apart on me, or I really should have respected my school staff more).

It is very easy to get chat privileges. StackOverflow accounts are free and points can't be burned for heat.

Rewriting the documentation for chat isn't going to change behaviors for problematic new users. That user isn't reading the FAQ. He/she probably has never read an FAQ.

Regulars know when they're breaking the rules. I think most don't really care when they do, to be honest.

The value I've derived from StackOverflow is the experience that I've had, which is something that can't be removed/moderated (unless the moderation tools go in a new, frightening direction). As always, I could be totally wrong here, but any policing action that would lead to losing internet points or account removal wouldn't be that effective for long-term users.

I bring this up because StackOverflow's model only works when there's a number of people providing a large amount of help for free. Similarly, no chatroom regulars == no chatroom.

---

What would I like to see change?

Have chat not be public, not be permanent.

I believe that we as humans express ourselves best when free from judgment or fear that some outside party will read it later.

That doesn't mean that you're free to say whatever you want -- the way you make others feel and the reputation that you create in a room do not go away, although the exact content of your messages could.

Better moderation tools.

It would be great to be able to have a system in which a majority of the currently present room owners could vote to suspend a user account from being able to join the room (for N hours). Accounts are free, but the asymmetric time-spend between account creation and moderation would make this work.

Maybe this is a discussion for another time, but having a feedback loop with SE staff/moderators could help in this regard.

Remove the 10k flag broadcast

A driver does something stupid and crashes on the highway. Other drivers stare at the crash and then themselves crash. This continues until there's a group of people not related to the crash staring at it, getting riled up. This is the current system of 10k flag broadcasts.

Answer 13

IMO any discussion about the future of the chat first and foremost needs to answer one questions:

Should chatrooms be limited to discussions about the subject of the site they belong to?

If the answer is yes, then the taverns, lounges and other "un-professional" rooms need to be shut down, and (given that subjects are hard to define around their edges) limits up to which deviations are accepted need to be agreed upon.

If, however, chatrooms are to be a third place where "water cooler discussions" about sport, beer, music, and other stuff is allowed as well, then meta will either have to fight its way through a cultural debate and at least attempt to agree on a list of cultural norms that need to be followed by all participants no matter what their cultural background, or it will have to define different places where different cultural norms are allowed, essentially resort to telling everybody who freaks out over some subject to go somewhere else. Or maybe it has to agree on some of both.

Because, let's not kid ourselves. This posting from Shog9 was triggered by the recent drama about Stackoverflow's Lounge room on MSO. This isn't the first time that the room's regulars got into conflict with meta's regulars. Why? Well, basically, more or less anything can be discussed in that room – usually, but not necessarily, interspersed with programming, most of which centers around C++ (and Haskell). If, however, the subject "sex" not only comes up in the room, but also accidentally spills over into MSO, a few dozen users there more or less erupt in rage over the fact that people dare discussing this on SO. Then the room's regulars look at them in bafflement, trying to figure out why the topic is supposed to be a problem. Hilarities ensue, and usually it goes downhill quickly from there.

Whatever is your side in this debate, I suggest you lean back a bit and try to gain a wider perspective. This is an international site, after all, and the company running it is never tired of proudly pointing out that its users come from all over the world. Surprisingly, though, people have very different cultures around the world, usually tend to cling to the ones they grew up with, consider it the best and most logical one, and (initially) take theirs with them wherever they go. However, some users (not a few of them Americans) have not really yet realized that this site they are using is an international site, have little clue as to how bewilderingly different the people from other parts of the world can be (both from them and from each other), have no idea how to deal with all those differences raining down on them, and tend to interpret the site's basic rule "Be tolerant" exclusively as "...towards my POV".

In some cultures, the word "fuck" is (regularly used, but) seen as not appropriate in any more or less official setting. In others, it isn't even used as the strongest swearword. In some cultures, alcohol is a strict taboo, in others, employees hang out the chat's "tavern" during their workday or announce their departure as "beer-o-clock". In some, people happily exchange recipes for roasted pork, in others, pork is not seen as fit for human consumption. In one culture the whole nation celebrates their female football (ha! soccer!) team having won the world cup, in others it's unthinkable that women even just visit the stadium as a spectator. In one, sex should not be discussed at work, but the purchase of assault rifles is freely debated, in others they laugh at the puritans disallowing mentioning the name of the programming language that's used in the joke on the site's 404 page, but cringe when people discuss buying guns.

Now, what is really hard to grasp for people who have never lived in different culture even for a few months: All the cultures that are different from my own do not have it wrong. They are just different. Even worse: Their members do not cling to the wrong ideals out of sheer pride or lack of education – and despite them seeing you showing them the right way. No, they really assume their ideals to be better, and wonder why you refuse to adopt them, given that the advantages are so apparent. However, when looked at from a 3rd-person POV, mine is not better than yours, nor is yours better than mine – they are just different.

Given this wild mix of ideals, moral frameworks, and sensitivities, how are you going to determine what is allowed on the website and what not? Shall we allow discussions about alcohol or rather not? What about pork? Shall we make Americans cringe by allowing sex to be discussed? Or Germans by not declaring "grammar nazi" as verboten? Shall we make Europeans suffer through a bunch of Americans freely discussing the advantages and disadvantages of different types of assault rifles? What about women and sport?

If you keep allowing non-programming topics to be discussed, then you will have to tackle this problem and find some consensus about which topics are allowed and which are not. In doing so, you will have to deal with a bunch of people being very cross, no matter what you decide for and against, and with some people leaving your website and being lost. And you might have to constantly defend, discuss, and re-adjust this consensus.

Of course, you could just as well avoid all this, do some hand waving, mumble "something decent behavior something", delete all the nasty discussions as soon as "the opposing party" looks the other way, and thus keep having them until kingdom come. You are the meta crowd, you have proven in the past that you basically own this website, and you have an impressive track record of stomping out opposing views. So, however, had once the ruling elite in some Usenet groups.

But please at the very least stop insulting my intellect by saying things like "this subject has no place on a 'professional' website", when every time I ask for a definition of "this topic" it turns out that it is something you cannot describe in any other way than as "one of those which I don't like to be discussed here". (And don't get me started at the funny notion that "room culture" has no place here. If you believe this, then I suggest you read a book on the subject in order to learn what culture actually is.¹)

Note: This could have had an impressive list of links to meta postings over the years, the basic rules, the FAQ, and whatnot. However, I have gone through these discussions a bit too often, have put effort into backing up my arguments – only to find out that these have been deleted when I want to point them out, rather than keep repeating the same stuff over and over year after year. Consequently, it not only took me days to finally pull myself together and post this as one more attempt, but I also lack the motivation to put any more effort into this.

---

_{¹Just two sentences: Every spontaneously assembled group of chimps immediately develops its own culture the moment it's formed. Their second-closest relatives certainly cannot not have culture. (And do not bother putting your chimp jokes into comments. I'm a gorilla. I look down on them.)}

Answer 14

One thing that I've observed is that site moderators are not enough to moderate chat. The SE platform tries to alleviate by having 10k users also handle chat flags, but I think this doesn't really solve the problem of chat being under-moderated.

I suggest that a new class of users be created: chat moderators. People who are somehow chosen (elections? appointed by site mods?) to be moderators in chat only. Give them the same full range of powers that diamond moderators currently have in chat, but don't make them moderators on the main/meta site.

Also, do a better job surfacing how people are supposed to behave in chat. When someone first shows up in chat, have a popup that summarizes the rules.

When someone becomes a 10k user, only grant them the ability to clear chat flags if they agree to documentation on how to moderate chat. This documentation should be created, but largely should be modeled on the existing moderator's agreement and the chat FAQ. If the 10k user doesn't read and agree to this documentation, don't grant them the ability to moderate in chat. Also grant moderators the ability to strip people of this privilege if they abuse them by declining flags in groups to prevent moderation from being enforced in their communities.

When someone becomes a moderator or a chat moderator, similarly they should be required to agree to enforce this new chat moderation documentation before being granted moderation powers.

Answer 15

I have mentioned it before in various contexts, but whatever solution ends up being put in place needs to somehow take into account the scale of the problem. Because Size Does Matter:

Security only has 100k views a day, so we have a different set of problems to SO, for example:

In Security chat, the numbers of users who have improved far outweighs the number of permatrolls. Generally we are light touch, and we guide people onto the 'path of niceness' - and this includes our regulars. If we see foul language we edit and warn, and recurrence earns a suspension.. Initially brief, but extending, and for those who continue to be offensive we will happily suspend for as long as needed.

So while a solution that solves problems on a massive site is definitely needed, be aware that transparent self-policing can actually work very well on a smaller site, and can be monitored by CM's if required (although that may feel like an audit, it can be sufficient)

Answer 16

1. What is appropriate conduct for this site?

---

What is considered appropriate behaviour for the site is clear.

From our help centre:

What Kind of Behaviour is Expected of Users?

This includes a link to Be Nice

1. Rudeness and belittling language are not okay. Your tone should match the way you'd talk in person with someone you respect and whom you want to respect you. If you don't have time to say something politely, just leave it for someone who does.

2. Be welcoming, be patient, and assume good intentions. Don't expect new users to know all the rules — they don't. And be patient while they learn. If you're here for help, make it as easy as possible for others to help you. Everyone here is volunteering, and no one responds well to demands for help.

3. Don't be a jerk. These are just a few examples. If you see them, flag them:

• 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").
• 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.)
• Inappropriate language or attention. Avoid vulgar¹ terms and anything sexually suggestive. Also, this is not a dating site.
• Harassment and bullying. If you see a hostile interaction, flag it. If it keeps up, disengage — we'll handle it. If something needs staff attention, you can use the contact us link at the bottom of every page.

2. What is appropriate conduct for this site's chat rooms?

---

From the Chat FAQ

Do I have to be nice? Yes. We expect community members to treat each other with respect … even when they don't deserve it.

This site is collectively moderated by the community through participatory flagging. If your chat messages are repeatedly being flagged by your peers you might find yourself muted, banned, or – in extreme cases – your The Stack Exchange Network reputation may be negatively affected.

If you see an inappropriate, spam, or offensive chat message, please use the hover menu at the far right of each chat message to flag it. This will draw it to the attention of the community moderators.

Stack Exchange Network Terms of Service

By using or accessing the Services, you agree to become bound by all the terms and conditions of this Agreement. If you do not agree to all the terms and conditions of this Agreement, do not use the Services.

So there is no justification for crying unfair when you break the rules of the site.

However this does lead to the next question.

---

2. Are the rules satisfactory (the question would imply no) so how do we change them?

---

The rules are satisfactory, it's just the site users need to be reminded of them. It's just bad habits have formed and people need to be reminded that this is a public place and the behaviour of it's regular users is a reflection upon the site. For better or worse, this is a fact. With higher rep comes more privileges, but coupled with this is more responsibility. Whether we like it or not that is how the main system of this site has been built. As a general rule people will respect answers on a given topic from users with higher rep.

So the rules need to be tightened and re-explained and it needs to be expressly stated that these rules apply to everyone, even the founders of Stack Exchange (just mentioned to reinforce no one is above it). Hopefully the change to the rules will help circumvent any potential loop hole arguments that have been poked through the current rule set.

---

3. How do we cope with significant community division?

---

We need to learn tolerance (and yes it can be practiced and learnt). There is too much conflict here, and arguments over side issues. Most of the people here are intelligent and we have well developed frontal lobes for a reason: to reason.

As for what is offensive; the whole nature of offensive material is it usually targets a minority group or a group of less empowered individuals. Whether I can or cannot relate to someone else's life experience, it's obvious some things are just not ok to joke about on this site (given the expected code of conduct). To feel the offence of another, one has to have empathy. If you don't have empathy, you need to learn logically what's right.

No amount of shouting that it's not offensive will make it ok. Particularly when the offending parties are usually not in the target group of the insult. Now when a group of people who are not within the this target group are also offended, you can be pretty sure it is way over the line.

The site has many tools in place to cope with frustration:

particularly this feature in chat

And what if we can't agree to disagree?

The site cannot continue with such a strong community rift and hope to sustain a productive and welcoming place.

This is the difficult part. If any user cannot accept the terms and conditions of the site, the user risks a ban from the site and the site risks losing users. By taking these steps people have been given more than fair warning.

There have been many users banned from chat, suspended from the site and they are frequently angry and belligerent. When people are in this state where they are at war with a community, as a group or as an individual, it can be almost, if not impossible to reason with people in this state. There is nothing the community can do about this, as it is within each individuals psyche to realise that their behaviour, rebellion, or rallying is counter productive.

For a background to this I refer to this post Request for community attention on a moderator's behavior.

---

¹Definitions of vulgar - have not included archaic definitions:

marked by lack of taste, culture, delicacy, manners, etc

Collins Dictionary

1. a vulgar joke, comment, action etc has a sexual meaning that is considered to be rude or offensive
2. someone who is vulgar is rude, unpleasant, and offensive
3. lacking the ability to judge what is attractive, suitable etc

Macmillan Dictionary

1. Lacking sophistication or good taste
2. Making explicit and offensive reference to sex or bodily functions; coarse and rude

Oxford Dictionaries

vulgar adjective (RUDE) › ​rude and ​likely to ​upset or ​anger ​people, ​especially by referring to ​sex and the ​body in an ​unpleasant way:

Cambridge Dictionary (did not include reference to upper society)

Definition of VULGAR for Kids

1. having or showing poor taste or manners : coarse
2. offensive in language or subject matter

Merriam-Webster Dictionary

Please feel free to ask for clarification and definition of any ambiguity in definitions

Answer 17

Let's start with a simple claim, just to make it clear, since I saw so many flames during these days claimin that staff should stay out of "room culture" since the actually room dweller are capable to handle the management themselves.

The chat is owned by the StackExchange staff

All of us, top community contributors users, ancient users with thousands rep points or newbie users that created an account yesterday are just that. Users, guests on a platform that someone is kindly offering us. Rooms aren't ours, and we are entitled nothing on them because they are just something the staff is kindly giving us. Sorry if this seems harsh to say, I really don't like to need to make this point but I fear some users are forgetting that, that they are just users of a system so I think that this premise had to be made.

That said, let's return to the problem we have. The "Be Nice" policy. While it is already quoted by other answers, I think that one more repetition won't do any bad.

1. Rudeness and belittling language are not okay. Your tone should match the way you'd talk in person with someone you respect and whom you want to respect you. If you don't have time to say something politely, just leave it for someone who does.

2. Be welcoming, be patient, and assume good intentions. Don't expect new users to know all the rules — they don't. And be patient while they learn. If you're here for help, make it as easy as possible for others to help you. Everyone here is volunteering, and no one responds well to demands for help.

3. Don't be a jerk. These are just a few examples. If you see them, flag them:

   • 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").
   • 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.)
   • Inappropriate language or attention. Avoid vulgar¹ terms and anything sexually suggestive. Also, this is not a dating site.
   • Harassment and bullying. If you see a hostile interaction, flag it. If it keeps up, disengage — we'll handle it. If something needs staff attention, you can use the contact us link at the bottom of every page.

I see many people often arguing in the chat rooms, thinking that the problem is just "bad, uncouth words" said in a chat room and since we are all grown up we are expected to live with that. Well, to the ones thinking that I would like to remember Stack allows for users as low as 13 years old to create an account. Is that the example you want to give them? I don't think so, would you? Do we really need to be constantly speaking like that to show that we are mature? Are we really demonstrating our maturity that way? And even than, would it really be fine to make others feel uneasy just because they should be the ones adapting? Please, why should we come to that?

Also, forgive me, but I don't think that "bad words" are the main problem we have. Just a few days before this whole process of reworking the chat rules was made public, I was exchanging some mails with the SE community manager Jon, expressing my worries at the reactions one flag I raised triggered in a room.

Surprise at flags on vulgar messages.

This is what Shog said. Sadly, in my case

Anger at a flag on racist / gender discrimination message

would have been more appropriate.

From my view point we have two problems.

• Inappropriate language, bad words and vulgarity. As I said before, I often receive a "we are all grown up, our room culture allows that" answer here. I personally avoid flagging these, but I want to remember everyone that nowhere is written that the be nice policy isn't enforced on chat, and you would be violating point 3.3. I won't be the one calling you out for that most of the times if you use moderation, but know that you will put yourself at risk. Chat transcript are permanent, readable to everyone and you can't know when someone will "go on a field hunt" and flag your message. It is lame to flag you one year after? Yes, it probably is, but sadly, you was the one that created the opportunity. And that is just disregarding the "Do you really need to make others feel uneasy/Do you really think is just their problem" argument I made before.

• Bigotry, race, gender, sexual orientation, religion discrimination or jokes. Now, this is something I don't even want to think to tolerate, something I don't want SE to tolerate. And something I pray people out there to not tolerate or turn their back to.

I would love to be able to think that all our problems are just about "Inappropriate words", but sadly while I browse the various boards, keeping an eye on them after the incidents, I see a lot of discrimination there. My whole discussion with the SE community management was started by some Homophobic/gender discriminating jokes I flagged, and my perplexity about the reaction the room had. Surprise and a bit of anger against that "unneeded flag". I won't post the details here, I don't want another flag war to start, but believe me when I said that I was quite perplexed when I was writing to Jon.

Lucky, I also don't think that the majority of the community is fine with such behaviors, and the reactions posts about the recent issues in the various chat room (incidents like the freezing of the Sci-Fi main room or the Lounge one) seem to confirm so.

I am again asking the staff, as I did before, to start making thing clear to the people misbehaving. This post from Shog is a perfect start. We don't need to tolerate such behaviors, and there is no room "culture" thing that can justify that. That "culture" isn't magically allowing users to discriminate.

As someone may already know, I had helped with an online game community management in the past. One of the staff members there had one saying.

This is our community. If you don't like the rules, the door is there.

It is a bit extreme yes, and since then I have seen and learned a little more of this thing they call friendship. Once, a wise one said:

The magic of friendship doesn't just exist in Equestria. It's everywhere. You can seek it out, or you can forever be alone. The choice is yours.

I am asking you if we really need rules to be better. If the only way this community we share can find harmony is by having some Cerberus - or a floating head - watching over us. I hope not.

I have tried a strange experiment on the chat some time ago, asking people how they would go to stop griefing in online games communities. In the end, we agreed that changing the mechanics of the game can only fix a griefing instance, avoiding it for a while but will never prevent griefing altogether, because the griefers will just find another hole in the system. We are no different.

We are the ones that need to change, for we are the Community.
This is my plea, let's try to be just that. A community of friends working together.

Rules can help. But the decision to enforce them must begin inside us.

Answer 18

One of the biggest issues was that we were trying to moderate based on a set of rules that didn't exist, and user exploited the loopholes as a defence to any inappropriate behaviour.

To place that into an analogy:

Suppose I drove a car without a licence: I'm 15 after all. The police chase me down, arrest me, and charge me for Driving without a license under the law. But what happened if there was no such law? Then I can get away, arguing that I didn't know I had committed a crime, and that I shouldn't be charged under something that arguably doesn't exist.

Chat Moderators are the Police Officers that enforce no such existing law. This is why so many counter-arguments citing room culture have had considerable voice in recent moderation incidents, because there was no high rule to enforce that.

---

Take a look at room culture:

The concept of room culture is just a concept, and nothing else. It's a term that should be used to denote the personality, and style of the regulars present in a chatroom. Room Culture is even something that has settled the tone, and the 'rules' of a room. That should be fine: the idea isn't bad. However...

The issue that comes with this though is there was nothing to make sure that the culture was 'constitutional'. We didn't have a clear set of rules and guidelines to govern culture, and moderators didn't have the authority to enforce it, because there was no clear rule that we could use to cite.

So will a reform of the Chat FAQ fix this?

Hopefully it will. Personally, I would want something like this to be addressed: Room culture surrounding topics should be encouraged, and there needs to be reminders that there is a grand set of rules so that moderators can keep things under control, effectively.

Answer 19

tl;dr Yes, chat is a problem. No, you can't solve it with words in general. Maybe you can solve it with words to mods.

---

I disagree with the suggestion that chat is not like a pub. Certainly, individual rooms are not pubs. Individual rooms are like tables in a pub. The chat system as a whole is one big pub.

The thing with pubs is, if someone walks past a table and there is a conversation going on and they don't like it, they might kick up a fuss. They probably won't, because a) they want to get back to their own table, b) drunk people are known to escalate to violence rather than reasoned discourse and c) the underlying social structure that keeps them in their groups could be tested.

Chat however, lowers some barriers and raises other. In online discussion, people don't have to be drunk to become (verbally) violent. In online discussion, you can be sat at more than one table at once. In online discussion, because of the multiple table situation, you can sit at a table you wouldn't normally sit at without your regular tables knowing (so the social structures don't get tested by you making a fuss at another table).

Another problem with chat is that the bouncers who would kick out troublesome people are sat at the tables, and are friendly with the people at the tables.

Of course, preventing moderators from chatting is a bad idea, because they are often the people who provide the help to passers by, and they are leaders of the community. If they don't participate in chat, they miss out on part of their community.

So this leads to bouncers not from the table coming over and getting up in the face of the table regulars and the bouncers of the table. Obviously this will escalate and cause hurt feelings.

I think that demonstrates the problem with chat and that pubs try to mitigate this through fear of pain, fear of ostracising and having bouncers separate to the people they are policing. It is worth noting that this still doesn't stop violent outbursts in pubs.

Of course you also say chat isn't IRC. Sure, technologically they are different, but a chat room is a chat room. If it looks like a hammer and sounds like a hammer...

If you don't want people to use chat as they would a chat room, then it's got to not be a chat room.

Chat has the added complication, over IRC, that it stores everything ever said. This means that people can't easily forget things. The water can't pass under the bridge. People's memories are fallible. The transcript is not.

This means that chat has the same problem that forums have. When things flare up on forums, people remember who did it and the forum software doesn't lose the fallout.

If you can't forget you can't move on. It's a lot more difficult to pretend that someone who was an arsehole is a different person when their arseholery is still on display.

Yes, you can delete things, and this does happen, but if a mod isn't called in then nothing goes away. On top of that, if a flag is raised on something in the transcript then all hell will break lose again.

IRC and forums are a progression from pubs, and so is SE chat. Certainly they could be used for reasoned, useful discussion, but mostly pubs don't offer that atmosphere. Mostly pubs are for groups of friends to have fun in pseudo-private.

So if the chat members can't police themselves, and telling them what chat is really all about won't make a fig of difference, maybe we moderators need to buck the trend.

Perhaps we need to be more reserved in what we allow ourselves to participate in, so that when we need to bring the hammer down, it's not such a surprise. We can't be accused of favouritism. When a mod comes from another room then you get the same treatment from them as you would a mod who is a regular in the room.

In chat, as on the main site, we can't have it all. We want to participate in chat and make friends with the regulars, but that creates a conflict of interest, which sucks, because I was a regular user, but now I'm blue and I can't have fun with you guys anymore. But maybe that's what we need to do to keep chat civil.

Answer 20

Stack Exchange generally promotes an idea of "focus on the post, not the poster," and for good reason. However, whereas the main Q&A is trying to build a library of high-quality answers (posts) that answer questions (also posts), the chat is intentionally trying to develop the community, which is made up of users.

In other words, "focus on the post" really doesn't map well to chat, where they really should be doing exactly the opposite: Focus on the poster user, not the post message.

The chat FAQ doesn't really put a lot of emphasis on that though, compared to what it puts on what to do when you see "…an inappropriate, spam, or offensive chat message…," or what happens when "…your chat messages are repeatedly being flagged…." This isn't helped by the current flagging system which only lets you flag messages; about the only tool available to room regulars to handle users is muting (aka "hide posts").

There are definitely situations where a blatantly inappropriate post gets flagged and appropriately handled, but many of the problematic users I've seen have basically turned "well, I'm not technically saying anything offensive" into an artform. Even when the signs that they, as a user, were refusing to play nice with others are fairly obvious early on, they seem to be allowed to…fester…for quite a while before anything actually comes to a head.

I for one would much rather have a room full of generally well-behaved people who may occasionally say something inappropriate in the ignorant heat of casual conversation (but remedy their behaviour when they realize, and actually feel bad about it afterwards) than I would a room full of antagonists who are unfailingly careful in their word-choice so as to avoid flags. But focussing on the message instead of the user risks punishing the former with one hand while attracting the latter with the other.

So FAQ-wise, I really think there needs to be a heavier emphasis on how rooms need to cultivate users who actually are willing to work with the community (both the room community itself, as well as Stack Exchange in general), whether or not they ever actually say anything that's technically offensive and/or off-topic. This would especially include the need to control — or prune out — bad apples before they become even more problematic.

Ideally, this would also come with an upgrade in the moderation tools available to room regulars — not necessarily just room owners — to handle problematic users in-house without needing to bring in a swarm of 10k outsiders who barely understand what's going on.

This is all of course ridiculously subjective, but I don't think there's really any way around that: These are inherently social problems, and social problems are inherently ridiculously subjective.

Answer 21

A short proposal, which I think could make a big difference.

On main and meta sites, we have the Be Nice policy. This is a good thing, because it sets out concrete standards for what is and isn't OK. In chat, the closest thing we have is a short summary saying "treat each other with respect".

When I step in to resolve issues in chat, I often find myself linking to the main-site Be Nice policy, instead of the chat summary. I do this because that policy does also apply in chat. However, because it's not explicitly stated anywhere that these are the concrete rules, they don't get followed.

Hence, I think we should copy the main-site Be Nice policy across to chat somehow, potentially as part of the new FAQ proposed here, or perhaps something else - would it perhaps be a good idea to implement an entire help center for chat?

Answer 22

Let's stop associating chat with the idea of "third place". It is a terrible foundation.

Here is what Wikipedia says a "third place" is:

In community building, the third place (or third space) is the social surroundings separate from the two usual social environments of home ("first place") and the workplace ("second place"). Examples of third places would be environments such as cafes, clubs or parks.

Expectations regarding acceptable behavior are going to differ depending on the place. Going by this theory of places, the only place where a professional demeanor is imposed from above is in the workplace. You don't have to act professionally at home. As a customer in a cafe, or in a club, the expectations are those appropriate for being in that setting. And if we talk about taverns, then the behavior which is acceptable there is much broader than what is required by professional standards.

The privilege page for chat, which is probably what introduces the concept to most users puts front and center the notion that chat is a third place, and explains:

All sites have a real time web chat component, or as we like to call it, the "third place".

The third place is a term used in the concept of community building to refer to social surroundings separate from the two usual social environments of home and the workplace. In his influential book The Great Good Place, Ray Oldenburg argues that third places are important for civil society, democracy, civic engagement, and establishing feelings of a sense of place.

Most needed are those ‘third places’ which lend a public balance to the increased privatization of home life. Third places are nothing more than informal public gathering places. The phrase ‘third places’ derives from considering our homes to be the ‘first’ places in our lives, and our work places the ‘second.’

By saying that chat is a third place, we are giving license to chat users to imagine whatever standard of behavior they want. The standard of behavior could be that of a church, or that of a tavern. But again and again, the third place is contrasted with that one place where professional behavior is expected, the workplace.

But wait... here is what the FAQ currently says:

Do have fun, but please keep it professional and always be respectful of your fellow community members.

(Emphasis added.)

It seems to me if we have to act professionally in chat, then it is in fact an extension of the workplace, and not a "third place" at all. Is it any surprise that users do not agree to what behavior is okay in chat?

"Keep it professional" should be the foundation of the behavior deemed okay in chat.

Answer 23

I'm an user that rarely use chat rooms of any kinds. However, I somethings walk into some SE chat rooms to see what topics are hot and now trending. I know that SE provides a very high quality chat rooms for people to discuss with each other easier and for moving long comments into chat rooms.

Although undoubtedly I believe that such chat rooms are a must for Q&A networks, recently I'm getting noticed that approximately in all chat rooms (not only those of SE), there are some kind of permanent folks who always are online and talking about trivia and potentially of no importance and value.

I also strongly believe that human moderation is not adequate at least in this case where potentially thousands are chatting. Also in another point of view, flagging mechanism is in my opinion useless, because, when a group of folks are chatting with each other about unrelated topics, it's unlikely that one of those flags another's message.

In my opinion, you should make use of automatic mechanisms to automatically find suspicious discussions and a rise an alarm to moderators not those that are chatting, then moderators can verify the evidences and proceed with appropriate actions.

Having a glance at this approach, it apparently doesn't seem to be that far-fetched. IMO, deploying a light-weight backgound process monitoring conversations could lead us to good news. Nevertheless, I do know that there might be many false-positive alarms that take valuable time of our moderators, but having some machine learning techniques could improve that system.

_{I'm not a native English speaker, so excuse me for my bad English.}

Answer 24

Maybe this topic i.e. your interest is problems which happen within chat (i.e. flames) but here's a slightly different problem.

---

Problem

On some sites chat is hardly used at all.

People who want to chat can't (easily) attract other people into a chat room, therefore chat is useless, therefore chat doesn't serve its purpose -- whose purpose is, I guess is

• To leave the Q+A site free of chat
• To enable some let's-get-to-know-each-other community building

There are three use cases:

1. Chat about a question
2. Chat about an answer
3. Chat about a comment

Sometimes I see a comment (or I'm tempted to make a comment) and think, "yes but that's a bit conversational and not exactly what comments are supposed to be for on the Q+A site".

After that thought there's then no easy way to branch the potential conversation (or invitation to a conversation) into chat, therefore chat is suppressed (or posted to the Q+A site). Alternatively,

• Moderators can migrate all the comments into a chat room (which is a blunt instrument).
• Users can create a new chat room and invite other users into it (which is cumbersome).

---

Suggestion

Maybe it would be a nice new feature to let moderators, and/or the user who posts a comment; flag any comment as chatty -- in which case:

• The chatty comment is posted in or moved to a chat room.
• The chatty comment retains a link to the Q+A topic or the comment which triggered it
• Whoever the chatty comment is addressed to sees it in their inbox (and can answer it in the chat room)
• The original Q+A topic and/or comment might have a discreet link to the associated chat
• Perhaps a choice of a new room per chat, or (for smaller sites) a dedicated room for all such chat, to get the chatters in one room together

Answer 25

Point of view of an outsider:

It appears that the name "chat" isn't appropriate. What the OP describes tells me that the chat is not a chat in the sense that everyone imagines when reading the word. I'm not saying that simply renaming it will solve any problems, but keep calling it chat and that's what it will behave like.

Answer 26

Just a brief suggestion on how to address the issue of room culture in terms of drawing a line and generalising the point:

With respect to developing an individual culture, chat rooms enjoy the same freedom and limitations as Stack Exchange sites.

From this we could derive for example:

• If your comment would be flagged as offensive on any Stack Exchange site (taking into account the context, if necessary), it does not belong in chat either.

• If a discussion digresses to a topic that would be inconceivable to be the focus of a Stack Exchange site (even if it gained enough support on Area 51), it should be put back on track.

Answer 27

I know this is old, but I don't see anything having been changed, so I'm replying.

I love SE chat. I like how easy it is to use, link "oneboxing" and image URL auto-embedding, etc. It's so damn usable, I put it up as the standard for internet chat engines.

I like that it's permanent. It allows me to wait months before responding to something and the OP is still pinged. We can pickup on conversations left unfinished years ago. In rooms where deeper topics are discussed, this becomes the perfect blending of immediate conversation and letters to old friends.

I'd think by now any internet user has learned that permanent records can be devastating for future plans and interactions. Well, count to three before you hit enter or something. Your words are your responsibility and yours alone. Truly offensive things will be deleted, and that's all that really needs to be said and done.

The issue of moderators in chatrooms creating more upset than preventing and resolving is, in my experience, entirely due to moderatorship being given to all rooms. Moderators in Room A see flags for messages in Room B, despite any connection they may or may not have to Room B. I just roll with it, but in situations where you know you're fine (because you visit that room all the time) it can be quite frustrating to have your argument interceded by someone you've never met. I have even seen these non-visitor moderators corrected by regular visitor moderators, and both concede along the lines, "well, run your room the way you want." If only SE limited moderatorship to rooms you actually frequent or otherwise have connection to, then this common problem would go away.

Answer 28

I found a really interesting article about preventing flame wars, which has a multiple-perspective model of communication.

(the page is really badly designed sadly)

In short, it posits that the sender's intention, the recipient's perception, and a neutral third party's perspective should be taken into account, in order to determine whether something constitutes flaming or not.

A table describing the multiple perspective model.

The paper seems to dismiss the widely-held view that the problem of flame wars is exacerbated by the fact that online communication lacks the contextual cues provided by face-to-face interaction.

I wonder if using this multi-perspective model would help in determining whether flaming has occurred, and therefore enable a more robust policy. For example, if user A is chatting with user B, and user A thinks that user B has said something racist, they can flag the comment; other users will then flag the comment as well (they are the 'neutral third parties' - insofar as anyone can ever be described as neutral).

Presumably comments being flagged affects your reputation on the site? You could also have upvote, downvote, and flag on chat posts (and comments generally) and when someone clicks the flag button, they can specify what content guideline they believe the comment violated.

Once your reputation drops below a certain level, you can't chat or comment anyway, I presume.

Answer 29

One of the things that generally help people adjust to rude comments is knowing that the person is rude to everyone not just them

One of the things that makes a constructive criticism easier to digest is when it comes from a credible person

Now, any language clearly against policy I think still needs flagging/banning just to show that it a moderated site and some behavior is simply not acceptable. But, in my experience, most people don't go there. And definitely not initially, perhaps only after escalation.

What I would suggest is a simple, slightly different system than on the Q&A site:

You can upvote or downvote comments

This will add or subtract 1 from your room rating

Ratings go from -10 to 25 to allow people to step up their game and prevent somebody building up enough points to not worry about an occasional drop

Then color code, something like:

• -10 red
• -2-9 orange
• -1 to 24 standard
• 25 green

That way if a person tagged red or orange offends you, you have a better chance to ignore the troll before it escalates. If a person tagged green asks you to tone it down, you may listen.

Since a scorpion likely won't change his nature in a new room, when joining a new room your starting room rep should be something like one half your average room rep

Answer 30

Here's something to throw into the pot. If, as seems likely, we have to come up with a spelled-out list of guidelines for stuff not to say, why not distinguish between two grades of such things? Namely, things that would offend many/most on Stack Exchange, vs. things that did offend at least one person. (That might actually be about all we need to spell out, in fact.)

If this distinction is made, both sorts of messages should generally be removed when the flag is validated, but the consequences for the flaggee need not be identical. Someone who persists in targeting easily-offended users and saying borderline things should suffer the usual consequences (escalating suspensions), but someone who accidentally offends in unusual ways a few times need not be suspended, just warned that they offended someone.

This requires a judgement call to distinguish the two, but I don't think that's either avoidable or impractical; the only real challenge would seem to be keeping good track of a particular user's behavior, but just logging the number of dubious-flag-triggered warnings (and comparing their flaggers, for diamonds) would go far toward enabling that.

And, of course, there are still some things that have no reason to be offending anyone; a flagger shouldn't be blindly heeded if they're taking offense at ridiculously harmless stuff.

TL/DR: There's a difference between taking offense at something, taking offense at "nothing", and taking offense at nothing at all. Let's be sure to keep those differences in mind.