>>solips+(OP)
Given the current trends in migration from other communities, I wouldn't be surprised if Paul is pressured into adding a "tl;dr" link to the header. ;)

>>solips+(OP)
HN was launched in Feb 07. The complaints that it had jumped the shark began about a year later, when we got a sudden influx of new users after Arrington wrote about us. To measure any possible decline I created an alternative version of the frontpage that only counted the votes of users who signed up in the first year:

http://news.ycombinator.com/classic

As you can see, it's still not much different from the regular frontpage, which is encouraging.

The median comment is probably not as good as it was several years ago, nor as civil, but things aren't dramatically worse, partly because the sorting of comments (which I've tweaked a fair amount) means the lame ones are less visible.

HN traffic roughly doubles each year. That growth rate shows no sign of decreasing. If it continued we'd have 32x the traffic in 5 years. We currently get a bit over 80k unique visitors on a typical weekday. 32x that is 2.6 million. It seems overoptimistic to expect the site could grow to that size and still be bearable to use. So probably either the growth rate is going to have to decrease, or HN will be destroyed. I'll do my best to make sure it's not the latter.

>>solips+(OP)
If the atmosphere and functionality are to continue coping effectively with the ongoing growth, I think new ground will have to be broken in how to manage large online communities. The human brain only deals effectively with a community size of about 150 persons. More than that causes issues (strife). Most online communities only have somewhere in the neighborhood of 10% to 30% of members as "active users" to some degree. I don't remember the statistics as well as I used to, but I think (typically) about 20% are "regular" posters and another 10% are one-time or very rare posters. I have in the past found that this meant that a forum with more than about 700 members would be effectively bumping the ceiling on that 150 person "community" mark. What I have seen is that this tended to cause spin-offs in some fashion or another.

My thought is that older cultures tend to be more formal as a means to effectively cope with the larger-than-150 community size of developed cities. I don't know how to apply that to the culture of an online community but I think that is where there is hope for effectively coping with the size of HN. This is the largest forum I have ever participated in. I know there are some other large forums out there, but I don't know how they compare, size/traffic-wise, to HN. In terms of culture, my general understanding is they tend to be less polite.

>>solips+(OP)
I fully expect that in 5 years a good portion of members who are active on Hacker News today will still not have launched any startups, there will be plenty of debate over which frameworks/languages are the best, lots of discussions about how to handle scaling problems, and plenty of threads about whether it's better to stick with a six-figure job or go the startup route.

>>solips+(OP)
I don't see much change in terms of appearance over the next five years. It's unlikely there'll be a need.

Where I think HN will need to change significantly is in the way submissions are handled. Currently you need a score of about 4 - 7 in the first hour to make the front page, as the number of users increase, the score threshold increases with it. HN is at the higher end in terms of front page quality, but there's already a lot of good stuff that gets missed. Ultimately that will be compounded as things go on unless something is done to cope with it.

>>pg+T
Would you ever consider implementing subreddits? The reddit team has said that they were implemented in order to help the site scale without losing the small community feel, and that certainly holds true for some of the smaller, more focused subreddits.

Alternatively, would you take steps to limit new membership? HN these days is often mentioned in the same breath as Twitter, Reddit, and Digg by many bloggers. I don't see that trend stopping any time soon, so if you didn't plan on implementing subreddits, then something like this might be necessary.

The Eternal September phenomenon has always been fascinating to me, not least because it is one of the Internet's oldest problems, yet it remains unsolved. I'd be interested in reading your thoughts on how a community can protect itself from dilution in this manner. Beyond the ideas expressed in the following link, did you have any other thoughts? Has anything changed your mind in the past year?

http://www.paulgraham.com/hackernews.html

>>solips+(OP)
HN is a great reference library as well as a news source and a community. The 'front page' is great for keeping your finger on the pulse, but what about more tools for finding and keeping up with subsets of the astronomical amount of information.

For example, what if you could look at all postings that have at least 10 but fewer than 30 ups? What if you could look at the posts with the most ups of all time? A data-mining point of view might accentuate the depth of the knowledge represented here.

>>pg+T
If HN has jumped the shark, what's the next hip spot?