To make the workings of Hacker News dependent on Reddit or any other site looks conceptually wrong --- can't it stand on its own?

The same holds for explicitly favouring topics like programming, math, science or whatever. From ycombinator.com/hackernews.html (emphasis is mine):

"If you had to reduce it to a sentence, the answer might be: anything that gratifies one's intellectual curiosity."

I think we should leave things exactly as they are. Some of the best "how to do your app" discussions result directly from these "fluffy" posts. Whenever one of these shows up, the ensuing thread is usually very long with many hackers presenting many different opinions. It turns into a "community-based analysis and design session" with exactly the kind of people you'd like to have.

"This is what the up arrow should mean."

"This is what karma should be."

"This is how to get it to behave like..."

And we're all talking about OUR app, the one we all know and love. These are my favorite threads and one of the best ways to learn and get to know each other a little better.

What we already have is better than anything any one of us could have designed.

Keep the fluff coming (just not too much).

(EDIT: This was the 2nd comment posted. One hour later, there are 22 comments about OUR app. This is EXACTLY what I was talking about. Great thread!)

I would not count on that. I have a hypothesis that the character of a site's comment threads is a time-delayed function of the top stories. If we want to keep the 14 year olds away, we have to ensure that the front page of News.YC looks boring to them when they come across it.

I wish I had access to such a great source of information when I was 14. I was just starting to become a competent programmer, and would have greatly benefited from the recurring messages in this site (like release early and often, I rarely released anything to anyone but a small circle of friends). I finally began to learn these principles through a lot of trial and error on my own in my mid 20's.

While I realize that the majority of users will be 20-40 year old web app creators buy don't alienate anyone. Going forward can we look to keep non-hackers away? I think everyone enjoys this community, let's not strangle the spirit of it trying to keep it from changing.

Sorry, I know there are thoughtful 14 year olds. I don't want to exclude them. I'm using "14 year olds" as shorthand for people whose mental age is 14. Basically, anyone, whatever their age, whose comments you can imagine being said by Beavis & Butthead.

I'm not sure users can be neatly split into "mental 14 year olds" versus "mental adults". I act like a mental 14 year old sometimes (usually on reddit), and sometimes I don't (usually here). I think most users are like that, and if we give an appearance of incentive for people to act maturely, then they will. I think this is at heart a psychological problem, not a sorting problem.

Of course, some people will act like "mental 14yo's" no matter what, and others never visit reddit anymore - but it seems intuitive that most people are somewhere in between.

One of the things I actually worry about are the "attractive nuisance" threads about things like politics and economics that are actually pretty interesting, or can be, but often end up being fairly divisive and distracting. This, for instance:

http://news.ycombinator.com/item?id=121137

So far it's civil... But they're the kinds of things that end up drawing people into often lengthly, fruitless debates - they end up being philosophical differences that might be ok to discuss over a bottle of wine, but often end up badly on the internet...

I challenge this.

The decision to broaden news.yc's scope to include other "interesting" information instantly adds fuzziness to the selection criteria. I have seen submissions that fly over my head, but other submissions in areas where I know more than the author.

Next, "interesting" content usually involves a mental challenge just slightly beyond the comfort zone. And so, given greater breadth of topics, which compromises depth, there will be enough submissions that are "interesting" to most people, including the 14 year-olds. To phrase differently, my knowledge in Latin American history might be comparable to a 14 year old, so we both find similar articles interesting. A BBC documentary link can easily be top quality, but appeals to all experts and novices alike.

The way you write it, is like 14 yos have an aversion to knowledge. Even if many Diggers seem immature, I'd bet they are more educated and tech-inclined than most in the same age.

Oops. Just reread the thread and your followup in the other branch.

Paul's point: immature 14 year olds tend to be attracted to shiny things. If they're mature, they'll prolly look past the outer appearance of the site...

I like what edw519 says. The community filter's itself via up arrows. If fluff reaches the top - isn't that what the community wants? Though, I do think we should put the following message at the top of the site: "If you vote for fluff then you really, really suck." :-)

I disagree. As the site gets more popular, there is going to be more incentive to game the system. Digg and Reddit are both victims of their own success. Once you start getting traffic, you start getting page rank. And, once you get page rank, your outgoing links are that much more valuable, and there is a greater incentive to game the system.

I think that safeguards need to be at least thought about to try and keep the quality of the site up. Case in point: http://news.ycombinator.com/item?id=121534

I suspect the site's already being gamed.

I don't understand the need to control things to such a degree. If certain posts become popular its because people want to read them. If its true that a majority of people here only want certain kinds of posts, then these "fluff" posts should move down the list pretty quickly naturally. If you really want this to be such a super exclusive club where only the "right" articles make it, then maybe we should just ditch the whole community thing entirely, make sign ups only through invitation -- AND if you pass a science/CS/math test! Then only people exactly the same as you would ever contribute.

"If certain posts become popular its because people want to read them." Why is reddit's content consistently horrible to the yc community? YC is not serving all people's needs in content - its serving a particular sector of users that expect knowledge/motivation from their content-browsing rather than humor/time-wasting.

Ycnews is figuring how much they should control because the fact is that if you don't have controls, the content and user base becomes like reddit. Also, the extreme degree of control that you suggest is so ridiculous that you're doing nothing but insulting a lot of hard work that people have done to find a good balance between being super-exclusive and open to everyone.

I think user selected filtering is a good half way point. Since this is a hacker site, it'd be cool to expose the ranking algorithm somehow for per user modification. Then, if someone comes up with a good one they can share it. This way effective filtering can happen, but it isn't forced on everyone.

http://arclanguage.org/item?id=1884

However, I know this is a hefty hack, so is probably not feasible.

That is a pretty intriguing idea. Except you'd want to use points as the test, or it would break if you were a day off.

I have another idea--why not just give people on the leaderboard the ability to "downmod", except it wouldn't affect the points of the submission, just increase the gravity so it drops off the front page quicker.

maybe people over, say, 500 karma? seems like enough. reaching the leaderboard is a lot harder than a fixed number.

expand the leaderboard to 100 people

ok: http://news.ycombinator.com/leaders

ty

The idea might work.

I do think that it's a sad state of affairs for Reddit, however, when we're considering Reddit's front page/karma system to be a fluff filter for Hacker News.

heh, I recently built a naive bayes analyzer that decides whether a text is from reddit or from ycnews. The assumption was that Reddit-like comments/stories were fluff and not worth reading.

How about also changing the title and url of fluff links that have been taken over by intelligent conversation?

For instance, that "programming is like sex" http://news.ycombinator.com/item?id=120816 article had good discussion in the comments and the title should have been changed to "what to do about fluff".

I agree.

Also:

While we may want to white list programming.reddit, there should be a way to keep off the fluff that prevails there. The fluff in programming.reddit are the "too popular" posts. We can flag the ones with too many points / too many comments. They scream "bicycle shed".

It's also in violation of reddit's terms of service.

I can think of several reasons it won't work. Off the top of my head:

1) There are just so many shitty links and LOLcats out there that the chances a shitty submission to YC is also on the frontpage of reddit is very small

2) The problem with reddit is that the signal/noise ratio is so high, not that there are no good submissions that make it to the frontpage. For instance this http://drnicwilliams.com/2008/02/22/zero-sign-on-with-client... is currently on the frontpage of both reddit and YC.

There are much better ways of doing it than this proposal. For instance use a bayesian filter to measure the submitted site up against previous submissions with high ratings.

I think the bayesian idea would work for the general case, but it would cause problems for the corner cases, which tend to be the most interesting.

Another alternative is that when a URL is submitted, that URL is checked against other popular sites (Digg, Reddit, Mixx, Sphinn, etc). Then the submission is granted (or debited) points based on where else/how frequently it appears on other sites. Submissions from some sites might get points auto-credited on creation (based on how commonly submissions from that site have accumulated points in the past).

Submissions that link only to another Reddit post, and not the actual story would be debited even more severely (those of us that want to read reddit already go there).

At some phase, a "FoundRead" might become a "Digg", ie: it's fairly well understood that the majority of the News.YC user base knows of the site and goes there directly to consume content. But in the short term it would be credited points for most submissions.

This should have the effect of demoting submissions that are all over the net, and especially those that are just links to digg/reddit. It should also have the effect of causing people to dig harder and deeper for truly good content to post.

how about being able to downmod submisions, or at least flag them as fluf? if enough users flag a post, (weighted by their karma) then that submission should be removed from the main page.

That is interesting. The problem we're facing is fluff, so it seems fair to devise a strategy targeting fluff specifically. I have a hunch that fluff climbs up to the top because of a small group of happy-fluff-voting people (young accounts for some, as PG noted earlier in another thread). If this assumption is right, we could rely on the whole HN community for flagging fluff out effectively, as those happy-fluff-voting people are a minority.

I'd discard ideas like allowing downvotes on articles. I do think the decision (by design) not to have downvotes reduces immature voting and also makes it clearer to detect suspect behavior, just like the fluff problem we're facing at the moment (e.g. 20 downvotes added to 40 upvotes brings a total of 20 votes, and thus reduces the alarm signal).

Maybe we should focus on the underlying users that vote up fluff, and kindly inform them of the goals of ycnews community in content rather than focus on dealing with the fluff itself?

Your "solution" to the problem of a community-driven website is to stop making it a community-driven website and call in the administrators? Isn't that pretty much giving up?

I think an important distinction to make is that there is no such thing as 100% community driven. Certainly reddit is a lot closer to 100% than Hacker News which is a lot closer than CNN, but no matter what there are admin involved. For example reddit decided that posting ascii pictures of Fry was funny... once. Then they changed the rules so you couldn't.

At Hacker News the admin have been pretty up front about what they will do; edit titles, penalize stupid people as defined by the admin, etc. I think the threat of intervention does a lot even if it's turned off at the moment. I'm sure it's in the cards to keep the site as good as possible and that objective is correlated with being community driven as much as possible, but there's a trade-off there and I'd bet sometimes it goes the other way.

Ahem: http://img.4chan.com/b/

(Even mentioning this, though, I theorize that, in the future, there will be a sort of Godwin's law that prevents sociologists from bringing up 4chan.)

An old suggestion I made: Just put a downmod arrow on once a submission hits the top 20.

The solution is way more simple. Since the stories are driven by the users, i would only allow a certain user profile to join. i.e hackers only.

The problem with digg and reddit is they allow anyone to join. In almost all UGC sites the users define the site. However if you define your users you define your site.

I understand where you are coming from.

However, to me and to many others I know, "Hacker" is an ideal that we strive towards... an asymptote... (Stallman is there...Linus is there...I am not going to be there at least for quite some time. But I am trying hard.)

> Hacker" is an ideal that we strive towards... an asymptote...

I disagree with that. There can't be an asymptote because there's no upper bound. You strive toward infinity and "hacker" is just a mile marker along the way.

True. From my point of reference, it at least seems like an asymptote.

The more I learn, the more questions I am coming up with - which seem to increase my mass as I accelerate. May be one day I might have a weird look on my face and be enlightened. ( Airport lady voice gently says "You have arrived" :-p )

No, that day won't come. You'll have lots of little epiphanies along the way every time you grok some new big beautiful idea, but there isn't going to be any sudden moment of True Enlightenment.

I suspect you've already passed the "hacker" milestone without realizing it. The metaphors you're using are a good sign. How do you rate on this rubric?

http://www.catb.org/~esr/faqs/hacker-howto.html#hacker_alrea...

Thank you! :-)

I am not sure of the answer for: "Has a well-established member of the hacker community ever called you a hacker?". But working on cool stuff will earn this soon or later I guess. So, thanks a lot for the encouragement.

Well, bear this in mind: you can consider condition 3 the least important. While conditions 1 and 2 are essential components of being a hacker, condition 3 is just a symptom. (1 && 2 && 3) is sufficient; (1 && 2) is necessary but not sufficient; (3) is neither necessary nor sufficient.

It's not like (3) is some sort of formal induction ceremony. It'll just be an offhand comment that the speaker probably won't remember making a week later. Most often it'll take the form of "you should talk to Fred; he's an XYZ hacker too" during a technical discussion about XYZ. It also tends to occur more often in person than over the internet. If you're worried because it hasn't happened yet, but most of your socialization is on the internet rather than meatspace, then don't sweat it.

I'm with parent. I do enjoy and learn a TON from this site but I doubt I would qualify for any sort of "hacker" title. I'm very much a H.I.T. (hacker in training).

Wow thats a cool term! I am going to call myself a "hitman" from now :-)

This solution is already functioning (via self-selection), and without need a top-down "you are only allowed if you meet X criteria" administrative layer. It should keep working this way (hopefully) so long as we keep the front page entirely boring to anyone but he or she who has the mind of a hacker.

It all relies on perceptions -- our own (those who participate in this community), as our perceptions create the atmosphere of this site, and those of the outsiders (both the outsiders we'd like to join, and those we'd like to keep out), as their perceptions determine whether they decide to join and participate here or not.

As has been stated more simply by PG, we want to maintain the perception to unwelcome outsiders that this site will be a waste of their time, and one very good way to do that is for our active participants to keep the front page on-message and full of considered, reasonable words, which most 14-year-olds avoid like leprosy.

One thing I really like about this site is its lack of flashy headlines. As soon as people start screwing with post titles to make them juicy and full of the "10 secrets to X, Y, and Z", then we'll get trouble, because the 14-year-old minds won't even realize that they have no interest in what we think and write about. They'll just jump in and turn this calm abode of thoughtfulness into a garbage pail of dandruff.

Honest question, do you really want to limit it to just hackers in (what I am assuming) you mean in the sense of "advanced software developers"?

I think (MHO of course) that by "hackers" you should include what I would call the "startup hackers", ie: people that also know how to "hack" business and not just PHP/ROR/etc.

What test do you propose, other than self-identification, that would cut across every possible definition of "hacker"?

You could make the site invitation-only. In order to join, you have to be invited by a member. Each member gets one invite a year.

This is how Livejournal used to work, and it resulted in everyone knowing at least one other person, which was important for a community site.

So not only do you have a recipe for cripplingly slow growth, but you also fail to ensure improvement in the quality of submissions. It's not like people are invulnerable to "Dude, c'mon...invite me!" from their friends, hacker or otherwise.

We have a pretty good crowd here now. I think that most of our friends would also be a good crowd, and most of the annoying 14-year-olds wouldn't be our friends.

And this didn't cripple the growth of Livejournal at all. You only get one invite a year, but you get it immediately on joining. I invite you, you turn around immediately and invite someone else, who invites someone else, etc. The 1/year rule is meant to make me careful about who I spend my one invite on.

Mmm...yeah, I see your point about the 1/year. Hadn't thought about it that way. And

most of the annoying 14-year-olds wouldn't be our friends

That's definitely the case here. :)

One invite a year? That's pretty extreme.

Another suggestion if it hasn't already been suggested before, is that the URLs provided should be checked with previous news.ycombinator.com submissions and if there is a match, the website should bug you about it.

i.e.: person A: submit: www.xyz.com/article1 -- OK

person B: submit: www.xyz.com/article1 -- <msword>"Are you /sure/ you want to submit this article? personA did this <date> ago."</msword>

person C: submit: www.xyz.com/article 2 -- Bug them even for this, just so people can get the idea that its in good form to stay fresh and not have 2 or 3 URLs in the main news feed that will be all from the same website. This isn't reddit or digg. This shouldn't be "oh I found this website slightly interesting and so i'll put up 3 or 4 submissions to specific articles."

What about a more economically-inspired system?

Add a karma cost to actions affecting the headlines (submitting and voting on content). Make users think a little more about how they vote, and when submitting make them put some amount of karma on the line which, if the submission is useful, will easily be earned back.

The karma for all of this must come from somewhere, as it does now. For people such as myself with a pretty low level of participation, maybe this would also be an incentive to speak up more often and submit some interesting material.

I don't know, it's not a complete thought, maybe some folks can improve it.

pg has noted that the fluff stories that this site's target audience might vote up (the "dangerous" stories that would attract those mentally-14-years-old people who are ruining / have ruined reddit) seem to tend to be quick to read. Is there a way HN could keep track of how long a particular story was kept open?

It'd be a pretty contrary approach. Clicking on a link would probably have to open the story in an iframe within HN. Lots of people hate that, I know. Maybe it could be made optional? Would people sacrifice that for improved link quality?

The iframe would be set below whatever amount of HN chrome at the top pg wants plus an "I'm done" button. HN could keep track of time between clicking on the story and clicking "I'm done".

A big question I had was "what about fraudulent reports." It's interesting that this is somewhat self limiting because increasing ranking this way would, by definition, take more time. Also, using the existing anti-spam heuristics (E.g., mostly new accounts upvoting a story), HN could further filter out time reports from unreliable sources.

It occurs to me also that this could be done completely transparently if HN didn't have to deal with the so called "same origin" security policy. A plug in could probably be written for HN to get around it, but bypassing the single most important security policy of the web might just be an intrinsically bad idea.

math.reddit and science.reddit should be considered exceptions as well I think.

Sure! I meant to include that too. Fixed now. Thanks.

Help us rally together against our common enemy. Send us all to Reddit "by mistake" 10% of the time that we try to access Hacker News.

By the time we make it back here, we'll be so distraught by our unexpected exposure to the histrionic twaddle of our enemy that we'll be in the mood to impulsively downvote anything that even remotely resembles fluff on our "home turf."

Defining ourselves in terms of Reddit (like refusing articles because they were posted on Reddit) is (funny and) problematic because:

- HN is independant from Reddit, and has its own identity, let's deal with our challenges.

- it would probably not fix the fluff problem (as it is not a simple solution, and there exists articles on Reddit which do not fall in the fluff category)

Ha, was thinking about that last night. Anything already on reddit should automatically be deleted.

It worked for Slashdot.

Once you rely on centralized human intelligence, expect a decrease in speed. Of course there are other ways to do it. I prefer a market system. If I had critical mass at my disposal, no second thoughts about using such a system.

What about turning on the existing voter power algorithm? (The one that has never been turned on as it was not making much of a difference in tests).

Would we see much of a difference?

What would be the problem by having a downmod arrow for submission based on a karma threshold like for comments? It would allow users to filter by themself.

No, thank you. I don't go to reddit anw.