When it comes to so-called "public knowledge," obscurity (and more largely, apathy) is the only kind of security there is. All websites eventually die (or, at least, all that I have observed); you can't "keep a community alive" so much as you can just let it die peacefully, leaving its children (perhaps under the same ownership!) to continue its legacy.
It is irrelevant whether or not obscurity is the only kind of protection available; obscurity is temporary and only needs to be defeated once. Furthermore, obscurity is not the only kind of security there is: Benevolent moderation is another form of protecting a community. And even if obscurity was the only kind of security available, we could simply invent a new kind of security.
I am unaware of any websites that have "died". I assume you refer to the digg.com and reddit.com exoduses that are so often brought up. Comparing those two sites to Hacker News is not comparing apples to apples. The commercial purpose of those sites is to sell advertisement space to advertisers, so trading a small group of thinkers in favor of a large group of more advertisement friendly users makes commercial sense. Hacker News, on the other hand, probably has the commercial purpose of driving business to YCombinator. It is in Hacker News' best interest to keep the small group of thinkers in favour of the larger population. So there is good reason to believe that Hacker News will be able to overcome these growing pains and be a community for the thinkers for awhile yet.
The only site that I used that and reminds me of HN is Perlmonks(.org). It never had any commercial ambitions and the quality has been maintained for a decade. There's tons of lessons to be learned from it. They use blind voting, have user levels, and the higher level users moderate. The community there is unbelievably awesome and it has a very similar feel to this site.
Obscurity in the sense of "how many people know about this site" is not binary - every inbound link makes it a little less obscure.
Popular forums (at least in my experience) stop being worthwhile to me past a certain point, and sometimes do just die outright from the problems caused by their popularity.
The sample size is small so far. I can't think of a lot of sites that have tried really hard to avoid becoming lame, even at the expense of growth. Most put growth first.
Less than it would with other people. I get a lot of info about which forums exist from referring urls. That was how I learned about both Delicious and StumbleUpon. So I can at least say with fair certainty that there's no good smaller forum where people link to my stuff. Except here of course.
Metafilter might count as an exception. They often have interestesting links. But reading the comments is like eavesdropping on a bunch of freshmen at Oberlin.
TechMeme is not bad, but I don't think the sites they link to are decided by user votes.
I'm sure there are thousands, for everything from hydroponic gardening to fursuits. I wasn't saying there aren't good small communities, just that there aren't any except News.YC in the subset of the world that reads the kind of thing I write.
> All websites eventually die (or, at least, all that I have observed); you can't "keep a community alive"
All this makes me think that maybe community forum software should be taking reddit's recommended page to the next level and only showing me posts/discussion from people who have similar voting habits to me or even based on some social networking style "proximity".
Having all submissions in one pot and giving everyone an equal vote averages everything out so the only way to maintain whatever niche the site caters for (in this case thoughtful and interesting posts/discussion) is to maintain the niche of users.
I bet Gnome-theme-designer forums have it easier because not everyone wants to be a part of that niche, but forums based around interesting and intelligent discussion are going to continue to slide gradually towards digg no matter how well they start because (a) everyone thinks they're more interesting and intelligent than they really are, and (b) lots of people want to join an online community filled with people "just like them".
That sounds pretty computationally intensive, if each page load is going to result in recalculating all scores for a relatively large universe of objects, but maybe "people who have similar voting habits to pg (or some small, select set of trusted quasi-moderators)" would be feasible.
I'm not so sure a technical, mathematical, and purely self-organizing solution exists -- I think at some point, standards have to come from the top. In the past, I think pg has exerted his influence by commenting on things he likes to see and doesn't like to see. He continues to do this, but either his comments will continue to occupy a diminishing share of the site as its user base grows, or he'll spend too much of his life commenting -- even lecturing and nagging.
I like the idea of giving each user a weight, calculated from how similar their votes are to some trusted set of voters. Each user's vote would then count for some fraction between 0 and 1. Scores could be updated periodically en masse to reflect new weights instead of with every page load.
I like that idea, pg and friends would be the founders of the community and would moderate it simply by voting on submissions.
Maybe such a system could also give people the ability to "found" new communities within some sub-domain and have all the posts from the main site submitted to their site but ranked differently based on different voting weights.
I agree that forums and sites like this have a lifespan. They don't die outright, but they become less and less useful/interesting as they grow and age. Techcrunch almost certainly reduced the useful life of news.yc, so I would have prefered it not happened. But how much did it reduce the lifespan by? Probably not that much, I suppose.
