It happens repeatedly that a small, esoteric bulletin board is constructed which attracts a small, tightly-knit group of contributors. They share a common goal, a common interest, a common set of values, and the community thrives.

Then it starts to attract attention.

Others whose interests are slightly different find the community and realise that it has a wealth of knowledge. It's mine of information, with highly regarded contributors giving freely of their time and experience.

The newcomers start to ask questions, often slightly, occasionally very, off-topic, and look to the participants for assistance or advice. Frequently they start to contribute their own thoughts, even though they themselves are less experienced, and less knowledgeable.

And the forum starts to drift. The contributions are no longer of such high quality, and the newer contributors start to cover old ground, repeating material that has long been absorbed by those whose presence attracted them.

Then the forum finds that the majority of the contributions are from newcomers, a mix of elementary questions asked by those who haven't done their homework, and contributions that are of questionable value compared with the earlier content.

The older hands start to complain about how it's all going to hell in a hand-cart, and begin to drift away to find, or found, a place to discuss in peace their ideas, and to share their experience.

And so the cycle repeats.

A common method to try to delay the decline is moderation. A few trusted people are imbued with the authority to reject contributions, both original and replies, and they try to keep the discussions on-topic and rich in content. If they succeed, the site becomes even more popular, and now even more moderators are required, and then who watches the watchers?

The difficulty of a small set of moderators is quickly realised, and so attention turns to moderation by the mob. Contributors themselves can up-vote or down-vote articles, potentially with the author gaining some sort of credit when people think they've done well.

But now there are two concepts that become conflated: Quality and Popularity.

And what about the "Me Too" item. It adds no content, except that someone agrees. Surely that is actually of some value.

The "Me Too" is easy to deal with in theory. Ignore it, and up-vote the original, thus supporting the sentiment without rewarding the contributor of the content-free reply. It's much more difficult, technically and emotionally, to reward an author for saying something well when you disagree with it.

So what's to be done? It's simply not going to happen that people up-vote purely on quality. Make it easy to down-vote, and people will happily bury even the best thought out and reasoned contributions if they don't agree with it. Certainly that's easier than constructing a well thought out and reasoned reply. Hit the down-vote button and move on - no time to waste.

We could provide two votes - one for quality, one for agreement, but people won't bother with that either.

Hacker News strikes a good balance. There are several measures: You can't down-vote until you yourself have gained sufficient "Karma", down-voting floors out at -8, and a few other technical subtleties, but even that's not perfect. It's still too easy to down-vote a content-full contribution simply because you disagree, and down-votes teach the contributor nothing.

There's no reason given for the down-vote, so there's no right of reply. Further, the guidelines specifically say not to complain about being down-voted (for good reason) and if you flag an item as inappropriate, don't also add a comment saying you've done so.

But the community might disagree with your reasons, if only you gave them. The contributor, and the community, can't learn why simply from the accumulating down-votes, and allowing the community to learn seems critical to the long-term survival.

On-line communities are still new, and the anonymity seems a precious commodity, but without some sort of responsibility,and no way to learn from community adjudged "mistakes," the progression of small, high-quality declining to popular but mediocre seems inevitable. The best efforts so far seem to rely on the small, secretive group of guardians, guiding and culling as necessary, working for the collective good. But that won't scale. We need some way to allow the best contributors to drift into power, and to help them be recognised as such.

Moderation is everything, so let's get the best moderators we can. Comments and suggestions welcome over at Hacker News: http://news.ycombinator.com/item?id=41373541