Warning: I wrote this post for my friends over on friendfeed. It’s a bit geeky and aimed at those who participate there so they can understand what I mean by decentralized moderation and why friendfeed has been building a really great community for discussing a variety of topics and why it is resistant to spam and trolls.
Over on friendfeed an interesting argument broke out. Neal Jansons said “couldn’t invisible comments be abused for search?”
My answer: no.
Why gets to what friendfeed has done that’s its most brilliant invention and one that very few people understand. It’s so good I wonder why they haven’t patented it.
I call it decentralized moderation.
First, what happens to traditional chat rooms and forums and why did they always get crappy after a while? I’ve been studying these for 25 years.
Forums and chats always start out interesting. A small group of people who are having fun and who are on the same wavelength. For instance, when Microsoft released NetMeeting (a pre-cursor to Skype) back in the mid-1990s there was a forum opened.
In the early days we were discovering a new technology together. There weren’t enough people there for the spammers or marketers to get interested. And it was a geeky topic, so the newbies and unwashed masses hadn’t yet come in.
It was the most fun I’ve ever had in a forum. Brian Sullivan was there too. The two of us had the best NetMeeting websites out there.
But soon newbies came in. At first that, too, was fun. We got to look smart and answer their questions. But after answering the same question 50 times things started getting old. We tried to help by writing FAQs and helpfully pointing people to the new search engine called Google so that they could find answers without us having to do the answer each time.
Even then it was still fun because we were having a ball with NetMeeting and we were all discovering cool new people all over the world (I remember talking with people, including a Russian General, for the first time on NetMeeting and sharing that with the group).
But eventually it got old. Part of that was Microsoft stopped development in the product, which made it not only boring, but doomed and I didn’t want to continue supporting a doomed product, but the community also saw more and more spammers and/or people pushing a business agenda that I didn’t find as fun to deal with.
I’ve seen this over and over as newbies coming in change a community and make it less fun to the older founders and also as spammers and commercial/marketers move in.
Regarding Chat I helped Leo Laporte run his chat room in the 1990s when he was on KGO Radio every Saturday night. I remember one evening we got so fed up with trolls and spammers that we blocked an entire country to try to deal with the problem. We had a centralized moderation system with several moderators and we were becoming overwhelmed in real time with dealing with this stuff.
Why did chat start falling apart? Two reasons:
1. As we got more successful a centralized moderation scheme fell apart. We couldn’t scale.
2. There was no way to really keep anonymous assholes out. There wasn’t a strong identity system. We could kick or ban people but they would just come in with another name or fake email address. Why did we block an entire country? Because that blocked an entire range of IP addresses which got rid of the problem, but also caused a new one, because now there was an entire country that was banned.
So, how does friendfeed deal with this problem?
1. Every item in friendfeed has only one owner: the person who wrote it.
2. An owner of an original item can delete comments underneath it.
3. Comment owners can edit or delete their own comments.
4. Everyone can block someone else. More on block in a second, because is like blowing someone away with a nuclear weapon. It creates an interesting social contract between the community on friendfeed.
5. Spammers can be deleted system wide, but also can be dealt with by blocking.
Why is blocking so effective on friendfeed?
Well, when you block you do a few things:
1. It keeps YOU from seeing the person you block.
2. It keeps the person you are blocking from seeing you. This takes away attack surface. They can no longer see your items and can no longer post spam to them (or anything else).
So, why is friendfeed resistant to trolling behavior where chat and older forums weren’t?
Well, for one, there’s social pressure on us all to behave. If I become a jerk there lots of people will block me and I’ll soon be left without anyone to talk with. So, there’s pressure on me (and everyone) to behave in a way that doesn’t get you blocked.
And, because my comments can be deleted if I just am a jerk in one person’s comments, there are two ways that my behavior can be modified.
Why wouldn’t I just delete comments from someone I don’t like?
The community at large shares information about people who do that and items from people who too frequently delete comments get fewer comments. Less engagement means less distribution for your ideas. So that’s a third way the system keeps you behaving.
So, has friendfeed seen any spammers? Yes, a few, but they quickly disappear and never appear again. Why? Because everyone blocks them very quickly, taking away their attack surface. That also warns the owners over at friendfeed that someone is doing something unseemly and they can be globally removed. But the whole system on friendfeed is based on following/followers so it’s easy to see if someone actually has any reputation on friendfeed. Does anyone follow that person? Probably not a spammer. Same thing is over on Twitter, by the way. I can pick the spammers out a mile away because they don’t have anyone of any reputation who follows them, except for auto followers (on friendfeed there’s no such thing as autofollowing yet).
Why can’t this system be applied to blogs? It can be, once everyone is on a common identity system, like Facebook Connect. Until then, though, getting rid of someone who is disruptive is too difficult, which is why blog comments generally are sucking more and more lately in comparison with friendfeed’s comments and also why you need a really complex spam filtering system, like those that Akismet makes (such a system is not required on friendfeed, at least not yet because the spammers haven’t figured out how to get past the decentralized moderation system yet).
Back to newbies, because friendfeed doesn’t have a centralized topic the first thing you’ll need to do, if you want to join in an already existing conversation is do a search. For instance, here’s a search on Quilting. That solves the problem of teaching newbies to search. Second, because advanced users can start private rooms, which newbies can’t see, they can have their fun and also help newbies learn the ropes. That takes care of a lot of the problems of Usenet which was really destroyed, in part, by newbies who came from AOL when AOL added a bridge to Usenet.
Anyway, it’s getting late and I’m not making as much sense as I will in the morning so I’ll post this and see where it goes.
UPDATE: of course this caused an interesting discussion to start over on friendfeed. Behave yourself!