The Final Solution?

DPReview has been a part of my life for 5 years. It helped a great deal to rekindle my interest in photography and I have spent a lot of time over the years posting technical info there … as best I could.

The place has become a battle ground though in recent years. A few not-so-thinking people like to lay the blame on me and a few others who like to ensure accurate data is posted to the forum. This inevitably brings out a few intolerant people who attack swiftly with great prejudice. Ad hominem arguments are their forte and they like long, drawn out circular discussions that tend to get the baby button pressed many times.
This is the source of the great purge of this week.

Now … someone on the forum (user kenseguro) has suggested a solution that is apparently used by other sites … although I cannot say I’ve ever heard of it.

what many web 2.0 companies do in situations like this is to silo all the offending users into a "offenders" group. So over time, offenders end up in a dense community of highly irritable people. The experience becomes so bad that they eventually leave.. in the meanwhile normal people enjoy an offender-less community. Works like a charm.

Or in other services, offenders become invisible to other users. They can still post, reply etc.. but none of their actions will be seen by other users. (insulation)

All these actions are rationalized by the fact that 1 offending user can take down multiple users and has an extremely negative impact on retention and engagement. So, it makes much more sense to insulate other users from the offenders, or to ban / perm ban the offenders than to solve it through dialog, etc.. After all, it's a website, not a consensus based governance framework.

I know many services who even kill paid accounts with no refunds, no questions asked. (the very successful maple story comes to mind) It's not an uncommon resolution method in the industry. It works, and the hammer (known as the ban hammer) should be used liberally, to maintain a psychologically safe and healthy environment for the users. The disclaimer is there for a good reason.

That’s censorship, pure and simple. The term psychologically safe and healthy environment makes sense in a group therapy session … but in a photography forum? Really?

The fact is, if you don’t want people to argue, then make a rule that any discussion longer than 5 replies on a side will be immediately deleted and the persons both banned permanently. Sounds draconian, but at least you are being dead clear up front. The reasoning would be that discussion leads to arguments and arguments are evil. Of course, the problem is that one person’s discussion is another person’s attack. That’s just how it is, and throwing up arbitrary rules just makes the system seem imbalanced … as the recent spate of bans shows so clearly. Some of the worst offenders managed to “slide” by an appropriate ban.

So the above solution would feel as arbitrary as any other one. DPReview has always had wild fights now and again … and the forum always came back to a decent norm. Until the arrival of the current attack pack and especially the sycophants who like to protect the leader … this was followed quickly by the advent of the baby button (complaint system) … that’s really changed how things work in there and imbalanced the whole environment.

Now the moderators complain that there are too many complaints for them to handle … so what next?
Now here is my suggestion for the final solution: analytics. It is used to track faces in casinos … to guess that fraud has taken place … and for many other purposes. In this case, the algorithms would be almost trivial. Such as:

• Track the number of discussions between two people that hit more than 5 replies on a side. That is a probable argument.
• Track the number of complaints between every pair of people: X complains against Y
• Track groups of arguers … i.e. these three people always band together in the same sub-thread against person X
• Track complaints by groups of people against the same person
• Track the result of complaint audits for people who complain more than twice in a month. Flag complaints as unfounded.

At this point, it is probably becoming clear where I am taking this. With analytics, we can now know who is doing what to whom and how often. All that is needed then is a simple time out system that kicks in automatically for *everyone involved* with no exceptions. All based on rules like:

• The third time an argument between people ensues within a two-week perdiod, they are both banned 3 days.
• The third unfounded complaint results in a 3 day ban for the complainer.
• 10 complaints for the same post results in a 3 day ban (but it is lifted if the ban was unjust.) This covers sexual harassment etc. If the ban is lifted, *all* complainers receive the 3 day ban instead.
• The same subset of people having an unjust ban lifted a second time are all banned 14 days for abuse.
• The same person being legitimately banned three times receives a 28 day ban.
• The same person receiving a second 28 day ban receives a 60 day ban.
• And so on …

This is pretty easy to implement and fairly hard to defeat. And most of all … it is *fair* …