The Self-Moderating Forum, or, Who Moderates The Moderators

There is no concept of voting down. It does not exist.

I think you're mischaracterizing "dissent" with "being an asshole to everyone". There is a difference. You can disagree without being disagreeable, and nobody objects to that.

What does get reliably flagged, and justifiably so, is people being assholes.

It always shocked me on Stack Overflow how quickly the community would strike down overt cruelty and meanness. It's almost enough to make you believe in people.

Almost.

It did make you believe in Software Engineers, didn't it? =)

I dunno. Maybe I hang out in different places, but I see a lot of people objecting to someone who disagrees politely with the community consensus. Though perhaps I'm not understanding what you mean by "disagreeable"? People can give dangerously bad information, or just say horrible things, in a very polite and friendly manner at times. Or they can say something accurate and useful in a polite and friendly way, and still get jumped on if it's an unpopular idea in the community.

One of the things I like moderators for is being the voice of authority to say "No, you can't spread bad information, even if you're being polite" and "No, you can't jump all over those people for saying that, even if you disagree." Works both ways.

Well, there are related bad forum behaviors that can certainly happen even if you're not technically not being disagreeable or an overt asshole:

• Griefing: you bring the same thing up over and over all the time, in every possible thread. You hound people.
  
  
• Trolling: every post you make is designed to stir up controversy for the sake of drama.

In both cases I would expect flags. And a history of getting flagged for persisting far beyond what any normal human would do.

One thing I've been frustrated with at Stack Exchange is that "building trust" feels like serious work. I didn't get involved at all for several years until I realized that I was always going to SO first to find answers to programming questions.

I registered and started attempting to contribute but I found that I was essentially surfing through unanswered questions trying to find things that I was even capable of answering. I then read that one of the ways to build trust was to ask questions so I tried to find some topic that I had a question about that had not already been asked and answered and found that I couldn't find any.

And so the whole thing quickly devolved into an exercise in opportunity cost analysis. I could theoretically have "worked" to build my reputation at SE/SO but there were a lot of other things that I wanted to do more.

At the end of the day I walked away from being "active" there because I didn't even have the right to do some pretty basic things like upvoting answers. It took me a while before I was even allowed to comment.

Now SE has way way more users than the typical small forum is ever going to have but I think that if you don't do something to combat the sense that a group of "elite" users are able to enjoy and benefit from the site and everybody else is kind of frozen out until they can raise their trust level then I think you'll have a problem.

This forum, BF, is a perfect example of how why should be pretty egalitarian. Everybody was able to make a fresh start here regardless of their (in some cases) diminished "trust level" at the previous hangout. A few people pulled the same shenanigans and quickly got banned, but most people just immediately started building a cool community.

I like what I see so far but I think the proof will be in the pudding. You want it to be easier for people to build awesome communities but not create annoying barriers to entry.

This is a good point. If this system were in place here on BF and you went to the Sanctum and posted that you didn't think gun control was the real problem in mass shootings you would probably be instantly permabanned. You would then have to appeal to a moderator, and if your post wasn't left hidden or deleted you'd probably keep getting banned and having to appeal, creating a bunch of extra work for a mod or leaving you with low "trust" even if you weren't trolling. This is also assuming there was a moderator around to unban you, which pretty much negates the point of the system.

Opinion systems (Forums) are fundamentally different than science and fact systems (Stack Exchange). For example, downvoting makes no sense on a forum, because how can sharing your opinion about Wolverine and the X-Men possibly be "wrong" in any rational, objective sense?

This also flows through to tracking trust. Forums are conversations. What kind of conversation can you have with someone who only talks, and rarely listens? That's the absurdity of post count as a visible metric on a forum. The people who talk the most are somehow more experienced or more trusted, simply because they talk more than everyone else? I don't think so.

I believe that when it comes to forums, Reading is Fundamental. The people you should trust aren't necessarily those who posted the most. The people who show up every day, who read a lot of posts day in and day out -- even if they never post a thing -- are the kinds of users who will understand that forum's culture and what should be there.

So even if you're a hard-core lurker and post absolutely nothing, never flag or reply or like a post, you'll generate some trust over time.

This isn't true at all. There was lots of discussion on that subject in that thread, and nobody got pissy. You can have a contrary opinion to the forum at large and not be hated.

Wumpus, I'm not trying to be a wet blanket, but do you honestly believe that the reason communities like QT3 became toxic or had moderation problems was due to a failure in technology? I disagree with that entire premise, and it therefore seems like your idea is a solution in search of a problem.

I was a moderator for a while on a forum for a site I was a staffer on, and the admin, in his infinite wisdom, decided to change the moderation policy to appoint whoever had the most posts at the beginning of the month as moderator. That did not end well. Nor did the site, but that's another matter entirely...

The root of the problem was unwillingness to let the community participate in moderation. But that forum software has two options:

- You are a moderator and a God. Your word is law.
- You are a user on the forum. You can do nothing but post.

I'm proposing that better software would add a few options in between there that let the community also participate in its own moderation, to some degree.

This can of course be disabled, but it'll be on by default. Software defaults do in fact influence people's behavior and decisions.

I guess before we created Stack Overflow, you could say that Q&A worked at answers.com and expertsexchange.com and yahoo answers... that forums worked well enough for programmers.. that we were creating "a solution in search of a problem".

Never mind that every single person I ask hates forum software. Never, not once, have I asked someone about forum software and had them say "oh yeah, I love forums. They're great. Wish we had more of 'em." When startups ask me questions, I tell them to ask their own damn community what they think of their service, rather than me, and if they have no community around their service to ask.. well there's your problem right there. The next thing they always say is "ah, you're right! What forum software should we use?" Then I start mentally playing sad trombone, because every forum option out there is 50 shades of terribad.

I actually like the forum software here. =) It's probably the other reason I migrated more and more over here compared to qt3.

It sounds like the community would soon be defined by whichever faction clicked the most, on aggregate. Not only griefers, but smaller factions or subfactions would be disenfranchised, although at different speeds.

Actually, it really would be a hivemind, wouldn't it? Defined by a specific set of agreements and disagreements. And the more closely an individual adhered to that set, the more "trust" he'd have.

It might be just the thing for a fan forum or a one-sided political forum, but I just don't see it for a forum with any diversity.

Same sort of thing happened on a Porcupine Tree discussion list I'm on - the owner of the list appointed mods based on how much they know/like PT. One of those guys created a really toxic tone on the list by basically treating everyone like they were stupid.

This and Vanilla are the two least-worst options. But still pretty bad. And written in PHP, which is synonymous with "bad". Neither is open source either. (Well, Vanilla has a crippled open source version, but that's ... lame.)

In a broad general sense, yes, things happen in the world because people get out of bed and do stuff. Those who are completely passive don't tend to affect much; Wikipedia is 'controlled' by those who are willing to participate and edit. If you scroll up you can read how passive readers gain trust, though. Please do.

Only in the sense that laws (flagging) and guidelines for civility create a "hivemind." Like "we have a hivemind around not being assholes to each other!" So yeah, there is intolerance for consistently being an asshole to everyone.

And anyway, "hivemind" is an odd thing to bring up, since forums are by definition groups of people with shared interests, like say http://www.stormfront.org/forum/ -- do you really want to hang out there and discuss Plessy v. Ferguson with these folks?

I guess the idea here is that the assigned moderators will get to confirm or veto the community moderation through appeals, and thus the "trust" will be given to the community members that apply their flagging in line with the wishes and style of the assigned moderators. Right?

Okay, now that I can understand.

Well that was a bit of a jokey exaggeration based on the dogpiling I remember someone got for saying something similar. The point is that even in mostly nice communities you get factions, and any time the conversation got heated a trust point system would basically let a larger or more vocal faction quash any contrary opinion. It would also probably involve a whole bunch of flagging and counter-flagging and lots of extra work for mods.

You've made a hidden assumption that the reason why there was unwillingness to let the community participate in the moderation was because that forum software only had two options. I think your assumption is wrong.

Furthermore, I don't believe that the quality of discussions or the culture of a group of people will be improved by offloading work from a human moderator onto an automated algorithm. That's simply not the problem.

This is a classic bad analogy fallacy and I see it from startup founders all the time. Just because you had a good idea once doesn't mean all your ideas are good.

But it is in fact the problem; out of the box, forums have zero native immune system. The community has no ability to protect itself from trolls and bad actors and griefers, without a formal moderator who a) knows what they're doing b) is there and willing to do it c) is able to do it.

Humor me and click this; what's one of the first things you read? http://apple.stackexchange.com/about -- yep "It's not a discussion forum. There's no chit-chat." Forums are the yin to Stack Exchange's yang. Stack Exchange is not-discussion. A system of science and facts, not opinions.

Having piloted an X-Wing for the last 4 years, I have a pretty damn deep understanding of how a Tie Fighter, and the rest of the Empire, works. This isn't some alternate reality venture far outside my problem domain. It is intimately related.

All hail the emperor.

The problem is that even accepting this at face value and saying it is entirely correct, I fail to see how your idea really fixes things because it still requires a formal moderator who is a, b, and c. On top of that, despite your good intentions you also give griefers and trolls a way to literally take over a forum if they game the system when there is actually poor moderation. This comes as a consequence of forums not being fact-based objective systems that are easily offloaded onto algorithms.

I'm still really not understanding what the problem you are trying to solve is. Trolls and griefers were absolutely not what caused the biggest problems on QT3.

Let me ask you this: Are you trying to create software which will support community creation? Or are you just trying to create software to let a company quickly stand up a feature request / technical support forum? They are two different kinds of things, but I think you are conflating them. Or I just don't understand which one you're talking about.

Heh, you just Godwinned your own request for opinions.

Meanwhile, if a simple shared interest were enough to prevent the kinds of divisions and cliques I'm talking about there would be many fewer people here on Brokenforum, for instance.

If you have a forum that is nothing but griefers and trolls -- and this does happen, because certain forums are explicitly "anything goes" -- then the behavior norms will be bad. I don't see how it's avoidable. This is the Escape from New York scenario.

But assuming that griefers and trolls are a minority, the majority will correctly flag them as consistent jerks and they'll eventually be shown the door.

Well, I spent hours -- literally hours -- reading every post in the epic Q23-to-BF diaspora topic from beginning to end. Did you? What I got out of it was that inconsistent removal of trolls and griefers was absolutely the problem. Either Tom threw up his hands and said "fuck it, enjoy Lord of the Flies", or was arbitrarily removing some great posters along with an incomprehensibly chosen slice of the worst trolls and griefers, while doing nothing to address many other known trolls and griefers that somehow enjoyed protection from Tom for reasons that are not clear.

In other words, the community had a better idea of who the bad actors in their community were, and who really needed to be shown the door. Certainly a far better idea than the lone moderator did.

What pissed people off was having no voice in decisions that really mattered, namely, getting rid of the worst of the worst that were poisoning the well for everyone.

Again, this is not about "silencing dissent". Not even a little. Civilized people can disagree without being disagreeable. This is about evicting the rare sociopaths from your community. Research has shown time and time again that allowing even one bad apple has profoundly negative implications for the group:

http://www.codinghorror.com/blog/2008/07/dealing-with-bad-apples.html
http://www.codinghorror.com/blog/2009/02/the-bad-apple-group-poison.html

I did. While it was happening. Imagine that! Someone read the same thing as you and didn't arrive at the same conclusion!

Each person in the community had a different idea of who the "bad actors" in their community were, but there was no community consensus, outside of a select few. There was fighting, but it was part of what made QT3 what it was. As evidenced by the period during which Tom left and the community was basically healthy for several months, without a "slightly more complicated than reddit" flagging system.

The liveliness and hilarity of QT3 would not have existed if some of the more confrontational personalities had been algorithmically purged years before. And that's just one example of how the culture of a community (sometimes with many varied sub-cultures within it) ebbs and flows and changes over time, and often needs sensitive human guidance to allow minority sub-communities to grow and thrive.

If I seem annoyed, it's because it appears to me that you came into this thread not with the intention of honestly discussing an idea, but already sure of your own correctness. Is it because you've already started building it and don't want to think about changing directions? I'm glad you're passionate about your next project, but if you wanted honest feedback, you should at least appear receptive to it.

I have a lot of confidence in Jeff to come up with cool software because he definitely has an impressive record of success.

I do think that the "hivemind" issue warrants more discussion though.

When I was a conservative, I was routinely savaged at QT3 because of my opinions. In many cases (though unfortunately not all) I was really trying hard to have a legitimate discussion with people about the issues. This did not matter.

My opinions were very widely disliked, even reviled, by many and I got a lot of flack and snark. There was no real way for anybody to moderate me though. I was never once the subject of any moderation at Qt3 but I am completely convinced that if the trusted users had the ability to attempt to moderate me they would have done so very mercilessly.

That, to me, is the definition of a hive mind. While there were some people that really tried to make me understand their points, there were plenty of others that seemed to think that anybody holding the opinions I held was essentially beyond help (it's a very common belief on both sides of the political spectrum and in many other areas of debate)

The fact that nobody could downvote me or do anything to get rid of me meant that I kept hanging around and started to gradually realize that the more reasonable posters had some serious and compelling arguments. I estimate that the influence of Qt3 was about 50% of the reason I stopped being a conservative and became a left of center moderate.

If there is any sense in which a group of people can lower somebody's trustedness in a community due solely to the unpopularity of their opinions, that community will rapidly become an echo chamber. I think the ignore feature is a far superior solution to the problem of "people you don't like" than any sort of flagging or downvoting.

I recall being blasted any number of times for my opinion even when I was not trolling in any way. I really have little doubt that I would have been downvoted and called a troll if it had been an option.

One of the main things I took away from my conversion from conservative to left-center was that one of the biggest problems we face in civil society is getting people to have a little more patience with those that disagree with us.

Any forum software that wants to streamline these issues will need to carefully consider the possible impact of hivemind behavior.

Those are two great posts. I also like the comment in the first one about the guy who insisted that dBase was the best database.

I still don't think that workplace or Q&A site rules are well-suited to a discussion forum, though.

Exactly what I was describing.

What I've also seen in forums is the exact opposite, that being acceptable targets. I've seen users pull shit that would get them banned in a second if they were not part of the in group. Fewer people in a liberal forum would flag a liberal bashing the lone conservative on the forum, even if it did break half a dozen rules.

I also think having some type of way to respond to the flags is good, preferably somewhere separate from the post. Meta comments are needed, but they also need to be put somewhere else. It's kinda like why we have a politics forum on here and QT3, it's a place to put the inevitable discussion that does not distract from the actual discussion. Like a discussion page for a wiki. Maybe flagging it would take you there.

I know you've probably already discussed this with them, but have you asked the MeFi mods what tools they'd like? Given the amount of moderation they do, and the number of times they've complained about the tools not being what they need for large or fast moving discussions, they'd probably have input. They have also had to push against the hivemind a fair number of times.

What you describe means there can be no system of civil law, because everything is subjective. I disagree. I find your beliefs fascinating, but ad hominem attacks, viciousness, and overt rudeness are regularly and consistently flagged by the Stack Exchange community.

You prefer a Lord of the Flies norm, with no rules. Noted. :)

I do not recall Tom characterizing that period of zero moderation as "everything went great!" But honestly, I didn't ever go to Politics & Religion, so I guess I didn't see much of the worst trauma. (I did notice after the great purge that P&R was almost empty.)

Ideally, the moderators and the community work together, a system of checks and balances, harnessing the community's scale and "eyes everywhere" but also having an appeals fallback, with moderators to expedite things or rush in if need be.

Perhaps. But I also know that good ideas tend to be a little scary. That's what makes them good. "If it's [an original idea], you will have to ram it down their throats." per http://en.wikiquote.org/wiki/Howard_H._Aiken

Personally, I prefer to trust my fellow community members, not assume they are a "hivemind" or out to get me and suppress my opinions.

Flag you for what? Were you being an asshole? Were you making vicious personal attacks? Did you post anyone's personal information? Did you grief folks, hound them and reply to every post they made with the same arguments? Did you curse like a sailor getting a beating?

If the answer to the above is "no", then why would you have to worry about being flagged? I've made mistakes. I've occasionally been an asshole sometimes when I should not have been. But not all the time. That's the difference.

Exactly, this works both ways, because the entire community participates. If anything it is the anti-hivemind.

Meta is a category. Reaching the flag threshold hides the post and generates a PM; discussion can happen in that PM.

Metafilter? We invited Josh Millard (cortex) so he's around. But they are doing their own thing for the most part.

Heh, yeah, that was the big irony, when Tom left, and then...nothing happened. The forums kept on trucking for the most part, and we had great posters come back. Stusser was still there doing the obvious spammer bans. I recall some mild idiocy with Dawn Falcon, but that only really went away because all the busy P&R people left. I think he got bored and left, or did he finally get banned by Tom? DF's threadshittery continued after Tom's return.

Starlight is his new nom de plume guerre merde.

*Cough* I guess it's out.

Most of the features sound good, and lot of the basic ones are what makes this place nice.

Yeah, thanks for posting that, I was just here to follow up.

OK so our thing is live now. Take a look!

http://www.discourse.org

Thanks for all the feedback on this and the other topics. Remember that unlike virtually every other forum software in the world, Discourse is 100% free and open source, now and forever. And doesn't suck. So feel free to use it, contribute, or ignore it completely.

Carry on.

But are there likes? And titles? And badges?

Important questions.

yeah wumpus, those are important questions! it looks pretty snazzy. the browser support worries me somewhat.. sure, most savvy/nerdy users will have their browsers up-to-date, but i kinda wonder if saying "we only work with these 0-day browsers" (exaggeration: slight) limits the audience of people who would potentially pick it up for use with their %demographic% of users? i'm sure there are very sound technical reasons behind that and also a lot of "you know what? fuck it, they need to update their browser!" behind that. either way, it doesn't affect -me-. good luck with it, despite the shit i gave ya when you started the thread.