I think that it is a bit premature to speak of an 'always' when we haven't had more than a few years of experience with any of this. It may take some time to get it right, and as long as the lessons from the failed communities stay learned there should be some visible progress.
HN is an excellent example of progress compared to those that came before. It almost certainly won't last forever but whether it succeeds or fails it has already raised the bar for what is possible considerably.
"Eternal september" isn't really new. We've had nearly 20 years experimenting this, we've seen usenet declining, slashdot, digg, reddit and whatever other net community you want to. Even HN is on the downward slope nowadays, I'd say.
What's interesting in this analysis is the second drawing. It looks like one of the best "well kept garden" out there, Perlmonks.org ; it never devolved thanks to its users and moderators dedication, but it's probably losing steam and users to stackoverflow and al., I think because of its 1999 look and feel.
I agree with the premise, but believe the author cites the wrong reasons.
What happens to online communities over time is the same that happens in any social organization. A few, more aggressive and hardcore members gain the power to snuff out dissenting viewpoints and opinions, thus overwriting any chance for diversity or originality.
To make matters worse, those that use downvoting systems wind up alienating new users that can't seem to figure out the "game." Those without voting systems use humor, belittlement, and castigation to do the same.
In the end, the scene is controlled by those holding a majority opinion, and any objectionable opinion is pushed out. This happens in any community, whether it's your local church, a football forum, or a place like HN. Where they fail in terms of diversity, they succeed for the group that holds the power.
I remain unconvinced. Yes, the arguments presented here are persuasive, but we're talking social phenomena here - nothing is black and white, there are no unbreakable rules.
So use the analysis to design a system that works in spite of these pressures. Perhaps you can start by being less restrictive of the membership and the topics, and then become more restrictive once you've gained users. Perhaps those who are contributing directly to the core subject can be rewarded in some way, and see an impact in the "real world," such as more exposure in relevant communities, more followers for their blog or on twitter, or similar.
Break the rules. At least, break the rules that need to be broken.
This piece finds rules that need to be broken. Find a way.
Agreed. The purpose of most communities is to grow. Certainly Digg and Reddit wanted growth; it's their business. Communities don't have to operate that way. Surely they can be designed for optimal quality; or as you said, modified so that any nonconstructive growth is curbed after a threshold.
There he talks about rules that matter, and rules that don't, and I think that partially encompasses my attitude.
But only partially. In addition I'd add: when it makes things better.
When people break the rules here they usually make things worse. Duplicated items split discussion. Multiple items about the same story with no extra information dilute the "New" and "Front" pages. People who leave witty replies that teach us nothing reduce the information density.
I expect there are ways of "breakingthe rules" here on HN that make things better. Generally people frown on "meta" items and discussion. I break the "rules" on that, so do others, and sometimes we get down-voted, and sometimes not.
It depends.
Does it make things better? That's my usual guideline, and that's why I've taken the time to reply to you, even though your question could be taken as a content and information free jibe. Maybe it wasn't intended that way, so I'll assume you were intending to be helpful, and provide my thoughts for your consideration.
The reason I asked was because a while ago you had posted about flagging people who you saw as breaking the rules, and I was wondering how you reconciled that with your very valid point here that the only way to keep a community vibrant is to break the rules. Your answer here makes sense and is pretty much how I think of it; for example, I knowingly break the rule of "don't bring it up if you're downvoted" knowing that my popularity will take a small hit.
The challenge though is that what rules "matter" and "better/worse" are subjective -- and when groupthink sets in, attempts to change the rules are opposed even if many people think they would result in a better community.
The big difference between metafilter and e.g. Hacker News/reddit is that the moderators at metafilter are very active and often delete posts and comments which are of too low quality, and they have a very active discussion with the community about the moderation (see http://metatalk.metafilter.com/).
In addition to this, signing up (required for posting) costs $5 and this is a significant barrier of entry for creating spam accounts and for users that just don't care enough about quality.
The MeFi moderation team is simply the best at their job that I've ever seen. Considering the shit they have to deal with (which gets way more heated on a daily basis than anything I've seen on Hacker News), the fact that they're able to consistently establish the center and keep MetaFilter welcoming to such a variety of people really astonishes me.
I don't know if the $5 is as significant as the quality of their moderation team. Perhaps it helps regulate the flow of new users. What really wins it for them is that the moderators are willing to respond to each member individually and talk them through the site if need be. I've never seen a Hacker News or a Reddit moderator do anything similar. (The Reddit mods are cool people but moderation isn't their focus at all.)
I think that the $5 fee is important to avoid burning out the moderators with a huge load of work. You can only manage/moderate a limited amount of work, not a tidal wave like noise on the "New" page on Hacker News.
So I think you need both: a high quality team, and a (rough) automated filter to restrain the amount of work put on the team.
To be fair, I can probably count the number of communities that have full-time paid moderators (I mean moderators that actually help guide discussion and interact with the community, not just administrators) on my hands and toes. Being the best isn't hard when you have such little competition.
As Groucho Marx put it, "I wouldn't join a club that would have me as a member"
One thing that HN & Reddit do right is put a time limit on discussions... They don't quite close, but they become less prominent. So if the system "fails" and a bad discussion starts, at least it goes away.
There was a discussion on LinkedIn about the poor quality of discussion on LinkedIn... The quality of LinkedIn users might be a factor, but we see the same problem on AMZN. General discussions on AMZN are dominated by idiots and cranks, whereas the product reviews on AMZN are intelligent and useful. The populations might be self-selected, but mostly you've got the same people being intelligent in one format and stupid in another.
Both LinkedIn and AMZN have the format that the most "active" discussion stays on top. So if you've got some hardheads who've got nothing better to do than argue about abortion or creationism, they crowd out good discussion.
So you don't just have communities burning out, but discussions also burn out, so it's good to retire them before they reach heat death.
Even if I accepted the left side of your first graph, the right side tells me that intelligence continues to fall monotonically during the final burnout of the community. It's just as likely that average intelligence would start to rise again (especially given the basic quality of the archives of the community) and so you'd see a damped oscillation finding some equilibrium.
And in fact, I'd argue that equilibrium has been reached, and is high, here at HNN, as well as at other fora here and there (Making Light being a great example, or PerlMonks.org). There are, in fact, smart communities that don't get overwhelmed.
I disagree; by the poster's definition, HN failed. It started out discussing technical items and startups, lasted for a while, and the topic has been completely diluted to "anything at all which is either new or interesting." In the words of the post, it has "pivoted to an off-topic focus." PerlMonks is a good example, though.
I agree that the presumed link in the post between topicality and "intelligence" is a strained one.
Ah, but the smart people are the ones who leave first. Go check out reddit.com/r/programming these days and see if you can find any of the smart people who were around back when it was still programming.reddit.com.
It'll happen here, if it hasn't already. People are starting spinoffs every few weeks, and it's only a matter of time before one of them gets some traction. One of them might have already, for all we know, since they certainly won't tell us about it here.
One thing to always keep in mind when starting a site on a particular topic, don't go too narrow on the domain name.
Choosing a name like ____forums where ____ is some brand (or nickname to avoid trade mark issues) becomes stale once that brand/model is out of style. Having a narrow name yet broad enough to last through the years is always a better approach (unless of course you don't intend the site to last pass the shelf-life of the brand/model).
Sentence one is good enough for me: I'm of the opinion that online communities that focus around a specific topic or group of topics... That's not really the kind of community that you expect to last very long, is it? At least without a change of topics.
He also seems to assume a priori that effective moderation is impossible, and thus that off-topic and banal discussion is inevitable -- which seems to me like assuming the conclusion.
I'll agree that most online communities seem to follow this pattern, but I'm not sure it's inevitable. There are at least a few community sites that (IMHO) have managed to maintain a high level of intelligent, polite discourse. MetaFilter comes to mind as an example as well as, so far at least, HN. Anyone else have other examples?
> Either way, they will fail eventually (the former due to dilution, the latter due to alternatives) and this is perhaps the reason why no online community can last for relatively significant amounts of time.
No online community can last because they will fail eventually... Does anybody else see anything horribly wrong with this kind of thinking?
Churn is inevitable and unavoidable. If the influx of new members doesn't match the exit rate of existing members then a community would shrink. Constant shrinkage is not sustainable.
The point of the article is to argue that all sites like this (HN, Reddit) will devolve into 4chan. The fact that it has a large user base does not make up for the fact that it's full of banal memes, porn, racism, and idiocy.
HN is an excellent example of progress compared to those that came before. It almost certainly won't last forever but whether it succeeds or fails it has already raised the bar for what is possible considerably.