Hacker News Reader

Comment by jacques_chester | Hacker News Reader

Ah yes, the cycle of website life.

* Hot new community forms at Site X.

* Site X residents refer to themselves as the New Wave of whatever. Much better than older Site W because of features/members/dynamic/demographics 1, 2 and 3!

* Site X's reputation spreads to former hot new sites T, U, V and W. Site X begins to attract more and more new users.

* Site X denizens begin linking articles at T, U, V, W and vice versa.

* Site X begins to exhaust natural topics of conversation. Denizens of more than 3 months standing become sick of 100th "What does Site X think about AlphaGamma?" post and begin to slap down newbies.

* Someone reminisces out loud about the Golden Days of Site X.

* Discussions on Site X become more and more about Site X. Extremely intelligent individuals begin to earnestly argue that their proposed feature will save Site X from itself.

* Someone proposes or launches Site Y. A how new community begins to form there ...

I've been watching this same story play itself out since Slashdot circa 1998.

View on HN

> Site X begins to exhaust natural topics of conversation. Denizens of more than 3 months standing become sick of 100th "What does Site X think about AlphaGamma?" post and begin to slap down newbies.

This implies it's hostility to "old hat discussions" that have already been around the block that people are worried about.

I don't think that's primarily it. The civility levels have slipped a lot... this place used to be as civil and friendly as meeting a bunch of people at a dinner party where everyone admires the host.

Still super-civil compared to the rest of the internet, but the levels have come down a lot. Raw outrage, profanity, and outright insults used to get downvoted to oblivion even if people agreed with the general argument the commentor was making. Rudeness really wasn't tolerated, so new members learned quickly that you had to think for a minute longer before replying.

It's still a great site, but it's not elitism and reminiscing that the old guard is scared of. It's comment quality slipping to rest-of-internet level.

>I've been watching this same story play itself out since Slashdot circa 1998.

And every social subculture in history. See: the constant evolution of the social scenes around music genres. It's just a fact of social psychology we have to live with.

"Dubstep was sooo good back in 02 before the students found out about it! I mainly go to future-garage raves now man..."

And Linux was so cool before the Ubuntu kids showed up.

Good point.

On reflection it resembles flocking behaviour. I imagine that website members can be modelled as simple agents with two rules:

1. A cool site has few, impressive members.

2. Move towards the cool sites.

Given an arbitrary distribution of "impressive" members, this should cause a constantly moving flock of members to move from site to site regardless of its purpose.

Add to that:

3. A member's threshold of minimum community coolness below which they will leave, is correlated to their impressiveness.

I figured I'd covered that in rule 1, but upon inspection, it's really two different rules. Your rule 3 is a better statement of one of them.

It strikes me that a logical step to prevent #2 is "allow only new contributing members that have been vouched for by n old members".

Yes, I'm that elitist. And yes, I'm well aware I probably wouldn't get voted in :)

But here's the other part - keep the site open for reading. That means it (hopefully) self-selects for quality, and other people can observe a high-quality discussion.

Do you have any examples of a site where treehousing has improved things?

Apple comes to mind, app-wise.

I was thinking in terms of communities.

Metafilter is often cited. They have a $5 fee for accounts.

Then again, Something Awful is $10...

That's clubhousing, not treehousing. Treehousing would be putting social barriers on entry. Clubhousing would be financial.

These terms are 100% arbitrary and invented an hour ago, but these seem like good definitions.

Dribbble, then.

It'll be interesting to see how that model holds up to another 2-3 years. The few invite-based communities I've been a part of started to degrade as the social graph grew out and started to pick up the occasional undesirable.

It was harder to get rid of them because they knew someone who knew someone, and trying to push them out had a network effect. It might be obvious to someone removed by 3+ levels that a bad element is bad, but the person who invited them might not see it.

Advogato investigated the problem of transitive trust networks pretty thoroughly as I recall. Anyone here able to comment on how they turned out?

I am seriously considering pruning branches, not just leaves of the social graph to keep communities healthy.

I.e. if you invited a bunch of bad apples, you will be pruned too. I have no idea if it would work, but it's pretty obvious that open social graphs beyond a certain threshold are not conducive to high-level discussions.

Zed Shaw's Utu had that idea. Doesn't seem like he's still working on it.

I'm not surprised that I like how Zed Shaw thinks :)

This is how many private bitorrent sites work, or at least claim to work. You are responsible for your invite tree.

A better way might be to nuke the undesirable, put their branch on notice if they start to cause trouble, and remove them if necessary.

It avoids unnecessary disruptions. Not all branches will shake if a leaf falls.

There is a story making rounds about hiring policies of Saint-Petersburg Mint (back in Royal Russia times). Every new hire not only needed three recommendations from current employees, but that should the new person ever get fired, those who recommended him would be let go as well.

Brutal, but brilliant... even if it is not necessary true :)

I assume, of course, that this process wasn't fully recursive, lest the entire mint be fired for one unfortunate hire.

Even then - it's only going UP the tree, not traversing down again to the siblings :)

This is similar to how many private BitTorrent trackers work. If someone you invited gets banned, so do you.

But here's the other, other part: keep the site open for writing as well, but make the non-elites' contributions only visible when you are logged in as a non-elite. Then the elites can continue their excellent discussions unhindered, and any passing stranger can browse the elite discussions: yet everyone can have their say in the agora.

Let's not forget it happens in the actual cities we live in as well! Hipster-ism gives way to gentrification and then the cool kids move because there are way too many yuppies. Same things happens with suburbs - they get too crowded and people start moving to the country to make new suburbs.

This is just human nature, it manifests itself in the clothes we wear, the cars/bikes we drive, the places we live, it certainly also applies to social websites. I'd be surprised if it doesn't - we'll see how facebook fares, or if there will be a coolkid exodus as the rest of the world gloms on.

It's basically a (sub)cultural arms race that everyone fights in.

Rent control has been a pretty effective way of preserving the identities of communities in cities. It's not perfect, but it helps. Concepts similar to rent control can and should be applied to sites trying to preserve certain characteristics.

In my experience (SF, LA) the rent control is there to prevent the gouging of yuppies, not to keep the poor people in. The idea is the city wants to keep rent just high enough to retain the rich people for the tax and spending base but low enough that the landlords to squeeze them out, which they most likely would do if given the chance.

It might be different in places other than California though, like New York.

You can't squeeze out rich people by raising prices without ending up with empty units, which landlords don't want.

Lots of landlords are stupid and will raise rent beyond the market clearing rate just to make a quick buck, resulting in empty units and shitty communities.

No, actually it's rent control that gives way to that. Landlords don't rent at a loss, and prefer keeping their apartments empty, resulting in shitty communities, then squatters, etc.

Both mechanisms seem somewhat plausible. Either of you got any evidence about how often each actually happens?

From the Wikipedia page ( http://en.wikipedia.org/wiki/Rent_control ):

Most economists believe that a ceiling on rents reduces the quality and quantity of housing available.[41][42] This view is based on analysis of empirical evidence as well as the understanding generated by theoretical models.[42] Economists from differing sides of the political spectrum, such as Paul Krugman[43] and Thomas Sowell,[44] have criticized rent regulation as poor economics which, despite its good intentions, leads to the creation of less housing, raises prices, and increases urban blight.

It probably works both ways, depending on exactly how low or high the rents are held. I've lived in some really, really nice areas that are rent controlled (SF Pacific Heights, Santa Monica, Downtown San Diego), apparently others have lived in relatively crappy neighborhoods where rent is also controlled.

Now that I think about it further, it seems obvious that rent control would be a tool that can be used by a municipality to control rents in either direction, depending on their goals.

Here in SF the geniuses of the Lembi Group bought a bunch of apartment buildings on credit (something like 1 billion dollar's worth), used questionable/borderline illegal tactics to squeeze people out of rent controlled units, and then jacked the prices way up. Then the real estate bubble popped and they went bankrupt.

Off-topic, but rent control has terrible problems. It lead to some grave NY neighbourhoods problems from the 60s to the 80s; back in the 20s in Paris...

Everything in moderation.

A big part of political-economic maturity (for me at least) realizing that just because the supply/demand lines from your Econ 101 class don't line up, doesn't mean that something is worthless. It could be that you're not measuring everything. Culture matters, preserving social mobility matters.

Well, to be perfectly exact, in the real-world, for a limited duration, controlled rent can work fine. It just has been applied without discernment too many times.

Sure, you can screw up anything, and I certainly wasn't saying "everything should be rent-controlled".

Just kinda irks me when someone comes in with the Econ 101 argument regarding a policy that's being implemented by people with advanced degrees in city planning. These people took Econ 101, and know more about the subject than you or I. They could still be wrong, of course, but it's not for a silly simplistic reason like that.

Big mistake mentioning future-garage, now I looked it up on Google and know what it is. Pretty soon I'll be showing up uninvited to your raves.

While I basically agree with you--I too have seen this cycle many times before--what I don't understand is why people feel the need to slap down newbies in an attempt to recapture some since-passed golden age.

If I get bored of a site or simply get nothing more out of it, I just move on.

If I see topics that are boring or simply have no interest to me, I just ignore them.

If I see rude, insulting comments or personal attacks, I just flag them and move on.

I really don't understand why people need to express so much hate.

Loss aversion, maybe?

People who are deeply invested in a site want to defend their view of how it should be. Change is threatening, and it's easy to react with knee-jerk hostility.

How about, instead of a single community, the site runs as a set of "cohort" communities that each only allow direct sign-ups for their first year in existence, but then run a bit like leagues thereafter? If people like you on HN/2011, you might be invited into HN/2010, then HN/2009, etc. It could also be automatic/karma-based: if you earn enough karma on HN/2011, you start again on HN/2010 with 1 karma point (though you can also stay on HN/2011, marked as an alumnus/boddhisattva.

Pretty interesting idea. Has anything like this ever been tried with other communities?

I think this would just breed resentment, or at worst Lord Of The Flies/Robbers Cave style anarchy.

[1] https://secure.wikimedia.org/wikipedia/en/wiki/Realistic_con...

When you have two groups that have no reason to not be the same group, the fact that they're separated breeds resentment. When you tell them that there's a criteria they were ranked upon to put them into these groups (in this case, age + karma), they understand. For example, there's no big rivalry between grad students and undergrads; one graduated to become the other.

>For example, there's no big rivalry between grad students and undergrads; one graduated to become the other.

All undergraduates know that they will soon be at least graduates, if not later postgrads. The transition is inevitable.

Any kind of stratified internet community would have to have the eventual prospect of transfer from one strata to the other, as an inevitable consequence of good/bad participation.

As soon as it becomes set in stone, or the transfer becomes a deliberate choice rather than an inevitable consequence of behaviour, I think people will forget about the criteria that originally divided them and slowly regress to hatred of the other.

A recursive treehouse sounds like an idea worth trying.

This comes across as snide. Slashdot circa 2011 is largely unreadable. So is Digg, so is main Reddit. The cycle may be real, but so is the phenomenon driving it.

Sure. All sites experience regression to the mean[1].

Usually this starts happening about 12 months after I decide to become a regular. In HN's case I've come in halfway through the cycle, about 3 months ago, so apparently the magnitude of my destructive powers is increasing.

[1] and also the snide.

I agree. One thing tho that never seems to come up: we people from older waves ignore the fact that we spent the last $N years growing, getting better, getting deeper into stuff, while the site has not grown with us. So even if the site hasn't changed, our perception of the topics and such has -- intro to X used to be deep, since we were little, but now that we are big, we just see a puddle :)

On reading this, the Evaporative Cooling effect comes to mind.

http://blog.bumblebeelabs.com/social-software-sundays-2-the-...

Interestingly, this isn't a problem related to any particular category of site. Hell, look at 4chan: bitching about newfags and how /b/ isn't good anymore, how /b/ was never good, different factions with different ideas on how to fix the problem (or what the problem is, or if there even is one), spinoff sites where the early adopters migrate to...

This isn't an HN problem, it's an Internet problem. pg could probably do some things to ward off dilution, but it won't last forever. Interestingly enough, Something Awful seemed to be pretty damned successful in this regard by doing the unthinkable and charging people to post there. I'd hate to suggest it as a solution, but it has been effective in the past.

pg could probably do some things to ward off dilution, but it won't last forever.

I think a paraphrase of this sentence has appeared in HN comments 100's, perhaps 1000's of times.

yep. and in all this time, pg has managed to keep the site hovering just above the shark, as he once put it.

SA has internal problems too. There's a constant slow cycle where the individual topic forums become boring and straightlaced because the users get proud of how mature they are, so they start new subforums where people are allowed to make "bad" but interesting posts as long as they pretend to be doing it ironically. Then everyone moves into those and it collapses, because they really did become bad posts.

Sometimes they move offsite instead, usually because the site admins didn't actually like the topic enough to make more forums for them, or because the admins banned all the users because they had formed secret IRC cabals conspiring against them.

The last stage is that the admins change the forum CSS so it's unreadable. I haven't figured out that part yet.

The gist of that (forums "cancer") is that those users "infected" with it have their posts become harder to read outside of those subforums by lightening the font. A quick mouseover returns the text color to black.

In my opinion, the method doesn't really work because they aren't reading their own posts. It should probably change everyone else's text color instead of their own. Greasemonkey scripts and the like would still fix it, but it would be a lot more annoying to the shitposters that way.

Still, for a site that's been up since the Dark Ages ('99), and relies almost entirely on its users to keep it relevant, it's been remarkably successful.

Just because this has been the typical website lifecycle in the past, doesn't mean that it has to be the typical website lifecycle in the future. We increasingly have better and better tools to manage how people interact with sites and communities.

Upvoting/downvoting, filtering, disemvoweling, moderation, tagging, reputation systems, etc. are just a smattering of the tools we have available and with time it's likely that we will see more and better tools for community management available in the future.

The most important thing is for "us" to define exactly what current and past features of the community we want to preserve and also decide what kind and amount of "chaos" we want to permit to allow the site to evolve and improve. It makes the most sense for PG, YC founders and very high-karma users to be responsible for defining these community characteristics, ideals and values.

The problem is that many of those tools are affected negatively by the regression to the mean as well. An average comment, that would never be upvoted during the early elite, will be upvoted by average people during the mean. Reputation follows.

The question to be answered then is "what mechanisms of reputation and combinatorial game mechanics can prevent regression towards the mean?"

I don't have an answer to this off the top of my head, but succinctly describing the problem as you just did oftentimes can be half the battle.

Another formulation:

"What system can allow a site to absorb N new users without causing regression to the mean, with the constraint the membership is open?"

I'm pretty sure the answer is: no such system exists.

Eigenvalue/SVD upvoting. Comments and articles are ranked for you according to their ranking with similarly voting/submitting/commenting users.

The frontpage (depending on how you set it up) could look like Reddit's, based on the majority of users, but to logged-in users it'd be more refined to their tastes. Perhaps you could look at the site from others' perspectives.

Defining the model, computing likenesses and finding the appropriate material could be quite resource intensive if you didn't think long and hard about the implementation.

(You'd might also want to avoid up and down arrows, because "like" and "dislike" quickly supplant "good comment" and "bad comment".)

The very first time I saw a site like this, many years ago, that is what I assumed all of the voting was about. I know a lot more now than I did then why resource wise it is the way it is, but I occasionally think about how one might go about doing a more customized version.

Superficially at least, it seems to be a good way to put everyone into a community whose size relates to the commonness of their preferences. I think this would occur as long as comments/articles beneath a certain threshold just weren't visible to a particular user.

Unfortunately it seems like a lot of calculating to do per hit, but in thinking about it, it doesn't seem to hurt Netflix too badly.

If it's too much complexity to track each user's preferences in terms of every other user's, you could also take a Bayesian filtering approach, similar to junk mail controls, and compute a custom front page based on each individual user's prior upvotes/downvotes.

In either case, the function of the voting mechanism will ultimately shift from generating the public front page to generating a custom view for each user. At that point, you could re-purpose the public front page as a selection mechanism to draw in the highest-quality users - perhaps by choosing a cohort of existing users to have public-upvote rights, or even by just applying old-fashioned editorial controls.

The core issue here (IMHO) can be summed up as:

* We all (and I'm aware I'm a pretty young account myself) vote on stories and comments to rank their quality

* Newer users vote less in line with the historic ethos because they're less aware of it

* All our votes have equal power (the downvote filter here is a crude effort to suppress that but isn't powerful enough to be fully effective)

* Network effect means a successful site's growth will be closer to exponential than linear

* With exponential growth the new who are unaware of the community standards are quickly able to overpower the existing who are aware

Hence almost inevitable degradation in quality which people here are complaining about and which I can partially see myself.

Two ways to partially address this spring to mind:

* Express vote power as a function of reputation - an upgrade to the downvote system. This has the problem that it provides a direct incentive to align with groupthink and post the easy but commonly held stories and threads so as to attract repuation and so increase one's voting power and.... It also, I think, requires floating point reputation and scoring, and so increases computational load.

* Display voting based on a similarity function - 'users who liked this also liked....'. This permits cohorts of users to work en masse without prior coordination to lower the ranking of content they find less interesting / valuable and so get what they want more easily. This means though that it increases fragmentation and reinforces groupthink again by design, which is clearly not ideal. It's also again significantly more computationally expensive.

TL;DR: Computation can be thrown at partial improvements but they're expensive to operate, open to gaming and likely to reinforce groupthink. I'm not sure it's socially or technically possible to keep the noisy brats out without severely damaging what attracted the Grand Old Users in the first place. Welcome to social groups...

What about disabling voting until a user reaches a certain amount of karma? Of course, you'd have to give the upvote ability to the original members (hand selected?) so that the system could actually get off the ground.

It would be similar to the "invite only" community idea where older users have to actually approve new members, but less restrictive. Perhaps even have upvotes on users instead of just comments. It would be similar to an "invite", but just for allowing them to vote. I'd say also keep the hellbanning that we have here. There will sadly be some false positives, but it seems to work very well for keeping out the worst.

Perhaps upvoting users could affect the reputation of the folks doing the upvoting. If you upvote a bad user into the community, your reputation suffers as well.

Now, this does nothing to solve the problem of groupthink, and in fact somewhat enforces it. However, I think some amount of groupthink is inevitable if you want to create a strong sense of community. Because you're trying to cluster users around a core set of values, it's probable that they'll think similarly in other ways as well.

Edit: I think multidimensional voting helps as well. Ignoring the comments voted "funny" on /. helps to filter out what people liked just because it was amusing. I think "agree/disagree" voting would be useful as a separate thing, because too often that winds up conflicting with upvoting/downvoting for quality.

If you are right, a website could automate the cycle as follows:

1) Every user belongs to at most two (or k) classes.

2) Each class can have a finite number of users n.

3) When a class fills up, a new one is opened.

4) There are always two (or k) open classes at a time, and new users join both of those two (or k) classes.

5) The classes are staggered so they don't close at the same time; e.g. if k=2, n=10000, the first 5000 people to join belong to class 1 and class 2, the next 5000 belong to class 2 and class 3, the next 5000 to class 3 and class 4, and so on.

6) Users can only post to classes they are in, and can only comment on posts in classes they are in (or maybe newer ones too).

7) By default, users only see posts to classes they are in (or maybe their classes + older ones). They can opt in to see more if they want.

The idea might work better if it is combined with some more onerous mechanism to join classes late - a fee, a requirement for an endorsement from a member or someone. Charging a fee for access to the more '1337' earlier classes might be a good way to monetise the site; the fees could either be flat, or could escalate (exponentially?) with the age of the class and / or the number of people who pay to join it late (possibly with a discount applied as active members leave).

Classes might get too small due to attrition - coalescing older classes when they get too small might be a solution to this - it might also make sense to increase class sizes (exponentially?) as the site gets more popular.

I wonder how different the website lifecycle is from the cycle of nightlife hotspots.

Reply on news.ycombinator.com