The Taming of Comment Trolls | Hacker News Reader

The Taming of Comment Trolls

wired.com

27 pointsdanw20 comments

> It was invented by Disqus, a company whose discussion software handles the threads at 90,000 blogs worldwide (including mine). In this paradigm, if a comment gets a lot of negative ratings, it goes invisible. No one can see it—except, crucially, the person who posted it. "So the troll just thinks that everyone has learned to ignore him, and he gets discouraged and goes away," chuckles Disqus cofounder Daniel Ha.

This is just a lack of fact checking. This technique hsa been in use for many years now.

I've used it on my discussion groups since 2001/2002, I can't remember the exact year. Joel of Joel on software describes a similar technique that he uses.

I'm not sure when it was introduced but I understand that reddit does this as well. It's a pretty clever idea, but it really hasn't helped much there so far.

Reddit isn't very strict with its moderation, unlike HN. My understanding is that reddit attempts to only block out-and-out spam / crapflooding / malicious links. There are ways to circumvent it or at least check to see if you've been banned as well, so it's not a perfect defense against a tenacious troll.

Good point. There is definitely a distinction of intent that is reflected in the tenor of the two sites.

In the most recent WIRED (magazine), the errata section acknowledges that the technique wasn't invented by Disqus. Not sure how it slipped past the initial fact check, but they figured it out eventually.

You do have to wonder how that didn't get corrected in the very mutable online version of the story, though...

Too bad it cites as some sort of authority Teresa Nielsen Hayden, who in my view only tends to disemvowel those who disagree with her or question her in some way.

And no, I have never been disemvowelled by her or anyone else, for that matter -- but I did stop going to boingboing after her reign over there began.

I think her moderation is fabulous.

On a forum I help run, we tried the method of hiding a person comments from everybody but themselves (we only did it once for a particularly troublesome member). It was completely effective. After a couple of weeks you could tell from their comments that they really felt left out, but didn't realize what was going on. They kept trying anything for attention and eventually went away.

Couldn't they log out, check the thread, see their posts omitted, and become as inflamed as if they're banned? Or was the entire forum private?

That does require suspicion though. So I guess the less people know of this method, the better.

They could. One workaround is to record the IP address a post came from (generally a good idea) and keep showing it to that IP, too.

They can still ask a friend to look, but the goal here is not to be perfect but to make any kind of improvement.

This whole article made sense to me until the last couple grafs, when I thought about "censorship" and first amendment lawsuits about whitehouse.gov and asked myself, "why is it good for us to be able to comment on whitehouse.gov"? Don't we already have enough forums for this discussion?

I don't see why whitehouse.gov shouldn't be allowed to delete comments within a specific defined set of published rules. If you're in a courthouse, and start hurling obscenities at the judge, or do anything "out of order", you can actually get thrown in jail. There are no free speech issues there, so why can't a similar (although hopefully less draconian) set of rules apply online?

This seems pretty effective. Giving the trolls what they want (or the illusion thereof). But insulating them from people who don't want to have to hear it.

I'm surprised that hacker news didn't get mentioned (okay, I'm glad they didn't mention it).

IMHO, this moderation system is the best (or maybe a close tie with slashdot)...trolly posts don't disappear, they just get lighter and lighter until they all-but disappear. All it takes after that is a highlight and you can read it.

I like that feature a lot, but I'm sometimes tempted to do CTR-A, just to see what others have "censored". Must...resist...

HN doesn't play the same game, so it's not fair to compare with it. It's very deliberately trying to stay small, or at least smallish, which solves a lot of the problems that plague other communities. It's also actively trying to limit the submissions to certain criteria, which helps too.

I think the main problem I have with this article and the entire way of thinking is the incredibly poor definition of a troll.

From their descriptions of a troll's comments, they make it seem like trolls are easy to spot and moderate when real trolls are often anything but.

The main focus of a troll is an attempt to game other human beings. That doesn't have anything to do with swearing or abusive language and it sure as hell isn't easy to spot.

Frankly, from the techniques described in the article, I think the only one that has a hope in hell of working against real trolls is the slashdot one. All the others are just casting a net over foul language and blatant stupidity.

It's mostly clever way of doing manual moderation. No matter what you try, it all comes down to a guy reading another guy's post and moderating it. Nothing has really changed for 20 years.

You might be interesting by this article on Edward Tufte's forum, a lot of good principles based on years of experience:

http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0...

All this needs to make it perfect is a bot that occasionally replies to the troll.

Something like:

Troll - M$ is teh suxxor

Bot - no you [anti-compliment], you suxxor moar!!!

Reply on news.ycombinator.com