And then...silence

Yes, what I find most interesting about listening to the Stack Exchange podcast is how much attention and conscious effort they put into constructing the community, and in a continuous fashion (i.e. it's never finished).

That said, it seems to me that we haven't reached the limits of the machine part of the human-machine moderation team. Voting up and down (here on HN, on Stack Exchange, Slashdot, etc.) is fine, but pretty rudimentary, for example. I'm wondering if there are any others experimenting with more heavily algorithmic moderation.

This is an incredibly insightful comment. There is simply no way to avoid the hard work of curating the community. That is the way for many valuable things, which is why they are valuable.

That said, there is a way to motivate speakers to speak well, substantially, and with respect: convince them that there's something in it for them if they do. "Reputation points" are only loosely coupled to the real prize here at HN: civil discussion between bona fide hackers about the topics we are interested in. This is very hard to find, and extremely valuable.

Building a community makes sense when the community is the primary focus on the website. But when your primary focus is something else - such as sharing letters from interesting people in history - it's just not worth the effort. I think Shaun made the right choice, and I think it applies to most websites.

Not everything needs to have a 'community'. A depressing majority of the web using public are idiots and having a no-consequence comment area just encourages them to leave their garbage for you to clean up.

All these people here making earnest suggestions about building comunities - you really should think about what went on here. Communities occur when people have something in common they care about, I doubt historic letters was ever going to be one of those. A community of people who collect historic letters? Perhaps. A community of people who just read them from time to time? Don't be so stupid.

To push the metaphor a bit too far, I think it's more about people realizing they just don't want cake that much.

As you say, it's perfectly possible to build a good community if you're willing to put some work into it. People who run community-oriented sites mostly know this by now.

But a lot of sites pick up communities accidentally. Blogs tend to have comments enabled by default. Adding a forum to a site sometimes seems like a simple idea that will make users happy. Social networking features seem like a must have for some kinds of web apps, even if they aren't the main focus. And of course, these accidental communities inevitably decay until the reek of them spreads through the whole site.

I think a lot of people are remembering that they didn't actually want to run communities in the first place. They added community features because it was the path of least resistance, and now that they've seen where it leads, they're turning them off rather than sinking a lot of effort into making them work.

I'm pretty happy with the development, personally. The communities I like are going to stick around, and as for the others, well, there are more than enough places to have flamewars already.

(Disclaimer: I turned off comments on my own blog a couple of months ago. It's made me a happier person, but I may be projecting just a bit.)

You clearly weren't on Usenet in the late 1980s.

I think you mean we soon found ourselves overwhelmed. He still takes discussion from people over email, twitter and facebook linked directly on the site.

Interesting stance you have. Based on your romanticized description:

"The truly free exchange of ideas, thoughts, and opinion."

and what you admit the actuality was:

"I was... one of the aforementioned persons who were... starting flame wars about trivial topics..."

you come to the conclusion that if people don't want to hear/read your flame war, they must not want any communication.

"Soon some of us soon found that we didn't actually want to hear any of it."

The problem is that reputation is relative. You may have high reputation on HackerNews, which could be linked to high reputation on ServerFault, but has absolutely no bearing your reputation on a board for writers and essayists.

To borrow a concept, it's a bit like Whuffie in Doctrow's Down and Out in the Magic Kingdom.

A problem is that reputation-based sites require accounts, which discourages insightful blog commenting by those not ready to commit to another log-in. Bayesian algorithms can defeat most spam while still allowing you to have an account-not-required comment system, but it does nothing to stop trolls.

Although maybe you could train algorithms to detect derogatory nouns and adjectives. And you could still allow reps on individual comments and hide low comments by default, which effectively crowdsources the troll moderation. I believe I've seen sites using Disqus in a way close to this.

Interesting idea, that. I might be less likely to download an app from the app store if the author had disabled comments/ratings, but that seems like the kind of call that the app author should be allowed to make. I recall reading about someone who uploaded a test hello world app to the Android Market and started getting lots of reviews and ratings (both good and bad) from trolls and idiots.

Could they tie Android Market comments to Google+ identities? Would that fix anything?

We had a more subtle approach at a forum for IT pros I used to moderate. We called it the Crazy Cafe. Known trolls wouldn't be banned -- they'd just create a new login and start trolling again, leading to a metagame amongst our community of guessing which of the old trolls the new troll really was. Instead, we'd simply flag the troll as a member of the Crazy Cafe, and their posts would be invisible except to other members of the Crazy Cafe (and to mods, if we so chose). The trolls didn't know they'd been made invisible; all they saw was that no one (save other trolls) responded to their flame bait, and thus they lost interest. It worked shockingly well.

I don't understand why the "disemvoweling" was ironic? That was something that she was already well known for doing on her own blog.

I actually stopped reading boingboing after the Violet Blue thing, because their response to it seemed so fundamentally dishonest.

I somewhat agree, but I don't think it's as dire as your comment makes it sound.

Going by sheer numbers, there are probably a million crappy, spammy forums for every good one.

But, I've found that most topics have at least one or two really good, (mostly) spam free forums where real discussions take place and the signal to noise ratio is really good.

Comments for articles and blog posts do seem to be more hit or miss, though.

Well, that is the worst designed page I have ever seen.

Color and layout wise it's beautiful. Content wise, where on earth is the "contact us" link? "About" link? copyright info? Sign up for beta? Etc... The logo is huge, but I don't care about the logo, I care about the product. And all the screenshots of the product are TINY. And I can't click on them. And why are they checker boarded all over the place, just have images on the left, text on the right. Checker boarding images is terrible practice.

It looks like such a good product and I want to know more but they're hiding everything from me. And that just stresses me out to the point where I hope their product fails. Call me a dick but I never would have guessed visiting a web page would piss me off so much. Why? Because I feel like they're teasing me, showing me something good and keeping me from getting to it. The equivalent of waving food in front of my face when I'm hungry but never giving it to me.

Design like that is just beyond unforgivable. That's the perfect example of what not to do when launching a web app. What am I suppose to do, dig around in the who-is info for their contact info?

The sets of people who can do calculus and people who are trolls are far from disjoint. Until there is strong AI, I don't think that a "content value" filtering mechanism is really possible. Even people have a hard time differentiating irony from stupidity in comments.

If it works well for them that's awesome, but that signup stinks of elitism for me. "You can't do some simple calculus? You must have nothing to add to the discussion."

Google announced recently that it didn't. They thought it would, but it turns out that people are perfectly happy to behave awfully under their real name (or, in some cases, throwaway accounts).

This stopped 90% of the insane comments on my local newspaper (they moved from disqus to facebook).

It also cut the average comments for an article in half or more.

Look at sensationalist TV (eg, Jerry Springer or Jeremy Kyle or similar) to see people who are more than happy to have their real world identity tied to extremes of behaviour.

Facebook comments are incredibly boring and meaningless. There is little trolling or flame wars but there is equally little meaningful content.

I’m not sure why everyone is obsessed with somehow finding a way to enable comments. Many websites are much better without them.

I have not found this to be the case at all.

I just read a story yesterday about a pilot who chopped his head off trying to hand-prop his plane. There were many terrible comments, one from a man laughing at his predicament and nominating the pilot for the Darwin Award and such, and mocking family and friends who were commenting to pay their respects.

The commenter's first, middle, and last name appeared right next to a picture of himself with his kids.

Not really.

Take a look at WSJ comments. Lots of real names, tons of flames.