Hacker News new | past | comments | ask | show | jobs | submit login
Snobbism-less way to save hacker-news from becoming another digg/reddit
25 points by tzury on March 24, 2008 | hide | past | favorite | 53 comments
The only two aggregators I consume in my google reader are this site and http://friendfeed.com/.

I recently noticed that hacker news is gaining popularity which makes posts less interesting.

I was wondering what if the karma will removed, would it cause karma-motivated posters looking for other places to be top ranked

I just wish we will remain a small community with common interests but I am not sure it is even possible nowadays.




The site has been diluted a bit. The TechCrunch post increased our traffic by 20% like a step function, and since then I've noticed slightly dumber stories getting slightly more votes. I don't think getting rid of karma would solve the problem, though. The problem is voting, not submission.

I may finally try turning on some kind of vote weighting. But there is a more organic solution:

Long-time users: please vote more.

If people voted up good stuff more often, that would fix the problem. And I know there is a lot of room to do that, because we now get around 12k unique visitors a day, and yet top stories rarely have more than 100 votes.


The best voters are going to be the ones who take the time to read and evaluate what they're voting on. They're naturally going to get outpaced by people who vote reflexively based on titles (Ron Paul syndrome) or who go for short attention span content. So even if you somehow manage to convert habitual readers into active participants over the long term you're fighting an uphill battle in terms of numbers.

I, for one, would love for you to turn on vote weighting because I'm really curious how well it will work.


They're naturally going to get outpaced by people who vote reflexively based on titles (Ron Paul syndrome) or who go for short attention span content.

Perhaps a little javascript could be used to disable the article vote arrow until the page was actually read by the user, or maybe the arrows should only be on the comment page.


What about generating more 2 front pages versions:

- One page listing the top posts, but only counting votes from the leaders, say, 10% of all users. So if Hacker News has 100,000 users, only the votes from the first 10,000 users would be counted.

- Another page listing the top posts counting only votes from the oldest users. So from 100,000 users, only the votes from the 10,000 oldest users would be counted.

I'd like to see pages like that, and notice the differences between the actual front page and those counting only the fist 10% users slice by these criteria. And also, because of their antiquity or karma, the 10% slice would receive a sort of reward.


One thing that I've been thinking is that... there just aren't that many really good HN worthy articles out there. Are more people likely to each post at a more or less constant rate? If so, that's going to dilute the site, if the number of "good articles" is constant in a given period. No idea how to fix that.


Write an HN-worthy article and submit it.

Easier said than done, I admit. ;)


I've done a few in the last couple months. It's much easier to vote for the good ones than write your own.


"... It's much easier to vote for the good ones than write your own. ..."

More to the point you are rewarded more to post articles on a regular basis. Writing your own articles requires a far greater energy input. This is exposing a flaw in the system in a way. Maybe HN user submitted stories should get recognition in the submission process? Maybe a start at 2 points instead of one?


Maybe by implementing that gravity* slider alluded to in news.arc, or at least decreasing gravity so higher-rated articles last longer at the top of the front page.


A recycled idea:

Make submitting a story cost 1 point. If you have taken the time to reflect on the quality of your submission you should be confident that you will get at least 1 up vote to break even. Users with a negative point total would not be able to submit again until they have gained back enough karma through commenting.

I think one main problem, or a symptom, is that we now have a new page with the vast majority of the links having no up votes. No one wants to sort through uninteresting links to find the good stuff - even to "save" a favourite site.

I suspect that davidw is right: "the number of good articles is constant in a given period". If that is right then attracting more submissions will result in fewer "good articles" finding their way through.


I think treating karma like currency will only exacerbate the situation. I feel like the underlying problem is that the audience HN targets is a generally outspoken one - we have lots of opinions and little problem expressing them. That being said, what we say is not always accurate or worthwhile. So, the more people that are attracted to HN, the more comments there will be and thus, the more inaccurate/worthless comments there will be.

Most of the members of HN seem to have their own blogs, each with at least some readership. If it were just really insightful comments all the time, they would probably mention HN on their blogs, and word would get out. This isn't that hypothetical as that seems to be exactly how HN got where it is today. Anyhow, the growth will always be substantial, so the only way to keep it how it was would be to apply some sort of invite only membership or externally limit participation.


> what we say is not always accurate or worthwhile

Comments that are inaccurate or worthless can be voted down. An increasing number of low quality links has no real cost to the submitters but it has real cost to the quality of the site.

Right now there is nothing to stop the number of low quality submissions from continuing to grow. I don't think invites or membership is the answer. The challenge is to be open and high-quality.

How would treating karma like currency exacerbate the situation?


Before I vote more I need a "this belongs here" vote that is different from my "this is good enough that I might want to find it in my 'saved' list" vote. My saved list is already unwieldy as is, so I don't want to start voting more using the current system.

An adequate solution would be a way to prune items from my saved list without undoing my vote.


I second this. I upvote comments much more than I upvote stories, since there's no drag on my usage later. I usually only upvote stories that I really, really want to be able to find again later.


We don't need a second type of voting, just let us unsave stories later.


Here's some ideas:

Votes don't count until you've been a member for X months.

Weigh votes based on a function of join date and karma.

If you have less than X amount of karma, your votes don't count. That would force new users to submit worthwhile content and comments before they have been "accepted" into the community.

After thinking about this for a while, I really like that third option I listed. If you don't contribute to the community, why should you be allowed to vote on the content of others?


"... The site has been diluted a bit. The TechCrunch post increased our traffic by 20% ... I don't think getting rid of karma would solve the problem, though. The problem is voting, not submission. ..."

Take a look at http://www.perlmonks.org You can vote but are limited by a daily ration proportional to hierarchy. But permonks has a different DNA compared to HN, less hierarchy more structured anarchy. So maybe that's not an ideal solution.

"... Long-time users: please vote more ..."

Willco, but I don't scale well. Of course there must be plenty of better ways to manage voting. Sounds like a new problem to solve.


Give a down vote to selected users.


That strikes me as a really good idea, even if it only went to people already designated as editors. My guess is that there are borderline stories that editors are close to killing but ultimately don't because of (perceived) community support and wariness of out and out censorship. Giving them a less extreme negative option could help.


How about having a leader board like page, listing users sorted by votes cast + submissions?


I always voted up good stuff from the new page. I have since stopped, of course.


the voting system is wrongly designed. you should be able to vote while you are reading the post/site you are voting for.

it should be easy to implement as a bookmarklet.


The issue to me is between these two things: Alienate potential users versus allowing the mainstream crowd to alter the community. Digg and Reddit are prime examples of what eventually happens once sites become mainstream. To determine the best action we need to know what the long term goals of the site are, without knowing that its tough to generate any 'fix'.

How about going the invitation only route? Or creating a series of basic coding/php type questions in order to sign up (think captcha)? I think either of those if implemented would cut down on the drive by comments that bring down discussion.


Personally, I'd go the mefi route and charge $5. That seems to work there, and (since this site is primarily about YC's aims and its potential founders) I'm sure serious entrepreneurs wouldn't think even once about paying.

Look at it this way: people will pay for the privilege of an intelligent conversation, but trolls feel so entitled that 99% of them will never pay.


It felt funny how different my reaction was to the idea of paying at Metafilter vs news.yc. I scoffed there, but I'd pay $5 here in a heartbeat. If that's all we have to do it seems a small price to pay.

Maybe if it added up to enough there'd even be a YC company funded by it :-)


Oh, man -- your post made my math-addicted brain start estimating the financial cost of the time I spend here on news.YC. Now I need a drink, and it's only noon.

Believe me, $5 would be nothing.


I'm not convinced that this "less interesting" problem is real.

For one thing, talking about the "quality" of news.yc posts is like talking about the weather. It's partially random, and it fluctuates. Sometimes it's grey and drizzly! Sometimes it's sunny! Sometimes you have Hurricane Katrina!

There certainly are lots of people complaining about the quality of posts, but that in itself isn't evidence of anything. People constantly talk about the weather, too. It's a safe topic, it's a shared experience that connects you with your audience ("Did you get wet yesterday? Yeah, I got wet, too!"), and it changes a lot, which guarantees that there will always be something to say.

Any community larger than five will never agree on the definition of an "on topic" submission. One person's sensationalist linkbait is another's clever marketing hook. One person's banal retread is another's timeless classic.

Finally, I don't think we need another system. We have a system: PG edits the site using whatever tools that come to hand. (Yes, the karma system is there, too, but it's kind of lightweight -- karma does not convey ultimate power, it's subtle in its influence, and it mostly serves as a way to send small-scale social signals.) Ultimately, I don't trust systems to build readable sites: I trust editors. I'm confident that, if the site starts to suck, PG will start arbitrarily pruning and transplanting and reworking the rules until it doesn't suck anymore. I certainly trust him more than I'll ever trust any mindless, gameable, automatic system, or any silver-bullet gimmick.


Believe me, it's real. I've been on reddit for a long time and I've seen it go from a great site with quality links to kitten pictures and "Vote Up If You H@te Hillary" posts. It's not as bad as most people say, but it's certainly a different site than it was a year ago. I'm not sure how to prevent it here, other than manually editing the system and deleting certain users.


Perhaps I should clarify that I'm not convinced that it's real on news.yc. Not yet, anyway.

Yeah, reddit and Digg are disaster areas. I don't know why, and your guess is as good as mine. My impression is that, in Digg's case, they tended to play up the gaming aspect early on -- e.g. they put up huge banners trumpeting the number of diggs for each article -- and built their audience out of people who think that competing to push articles up the leaderboard is a fun sport. Now they've reaped the whirlwind. Reddit, I'm not so sure I could even guess. I didn't stick around to analyze the phenomenon; I just left.

I do agree, however, that the only proven solution involves manually editing the system to some degree.


One thing that might be useful is for users to be able to block other users. If I don't want to read any more posts/submissions by WakeUpSheeple or AlienFromPrisonPlanet, I can just block them.

One could combine this with flagging accounts for possible deletion if they are blocked by too many people (as WakeUpSheeple probably would be). Deletion would occur only after editorial review.

This wouldn't work for that other site, since trolls have critical mass. But it might work for us.


Good question, as I've noticed a few posts completely irrelevant to everything that goes on normally here at Hacker News. It's a thin line to tread about removing karma and I'd love to hear some users thoughts about the issue.


I would seriously consider changing the appearance of the colored header bar. There are 9 items on this bar, but they are not all the same kind of item.

The first 4 (Hacker News, new, threads, and comments) are different views of the same data, so they should be tabs.

Leaders, jobs, and <login id> should be links (blue underlined) because they take you to another page, not another view.

Submit and logout should be buttons.

I would change "Hacker News" to "Popular" and I would put the "New" tab first, because the Popular view is the default. I still get confused by the "threads" and "comments" tabs (now what are these, again?)

Why do I make these suggestions?

Because we are all so used to the site, it's almost impossible for us to visualize how noobies see it. We really need to think about how to better direct people to the new page more often and more easily, without making it the default page. I think the biggest problem is that too many good stories never gain enough traction to graduate to the popular page. If a good story falls to 31st on the new page without an upvote, we'll probably never see it again. This happens often because new posts come in faster than people get to vote for them.

I often read a story that I think is fascinating and wonder how people here will react to it, so I submit it, only to find it disappearing without any upvotes or comments. On the other hand, sometimes my submission takes off to the top of the front page. I realize that I probably think a little differently that a lot of people, but not THAT MUCH differently. Sometimes, I think, it's just timing. Thus, the "It fell off the new page too fast" hypothesis.

Just a thought. Of course, if the data suggests otherwise, then never mind.


Sorting, searching, filtering, tagging and personal preference should solve this problem

Moral is, this site should add features that solve the problem instead of asking the users to ... just behave!


> Sorting, searching, filtering, tagging and personal preference should solve this problem

and ruin the sense of community (baby + bath water)


i think going social and "asymmetrical" is the way to go.

think twitter. their subscription system does not obligate you to follow people following you. this is the key to good content filtering.

of course we can keep a "public timeline" (for new users which don't know where to begin)


I agree that the posts are getting more mundane as some try to get 'the top score' possible. I'm all for this place being incredibly pompous if keeps up the quality of the content. I question the purpose of a number based karma system. Couldn't we just have a thumbs up and thumbs down approach on individual posters and not their posts. This way if you personally recognise that someone has a higher quality of posts, then you will go to the effort of recommending that person. Then you may choose to look at those who are connected to them. This way people don't have the same motivations. I'm all for a subscription based scheme as well if it keeps the quality up. Maybe even charge for it.


use a pagerank-like karma assignment (high karma people give more karma). it's better than having a slashdotesque editor. (and the only thing our grahamist overlords would have to do is to kill someone's karma if they go off topic)


all these ideas are a little experimental. how can we try them out without nuking this community?


I've also noticed a lack of interesting articles on the front page ever since pg's "troll" essay. The current general hubbub over zoo vs savanna comparisons takes up too many slots.

I think there are a couple ways to go. One thing to try is to show points, but no upvoting on the front page, and allow voting on the 'new' section, but no point totals. That way, articles that that get upvoted are interesting articles from people browsing new articles. Without the influence of points tallys, people will be voting more or less independently.

As it stands now, social news isn't a 'wisdom of the crowds' sort of thing, but more like a positive feedback system that relies on a catchy title and a chance on the front page. This probably makes it more of a hit-or-miss for good articles.

Secondly, as any community gets bigger, the common ground between them gets smaller. There was some quote about television being stupid and inane, not because people that watch it are stupid and inane, but that the common interests we have are stupid, but our refined tastes are varied.

I find myself upvoting very little on the front page now, and the I have to spend time scouring the 'new' section to find the interesting stuff. To be honest, I see no reason why we shouldn't fragment the top news. This isn't like a newspaper where we only have one page to print on. However, I don't mean, breaking it up into "subHNs" like reddit did.

I mean, I should be able to alter the weighing of the front page. The simplest scheme would be to allow a user to follow certain people's submits. I regularly like nickb (other than his xkcd submits) and chaostheory's submits, but not as much edw519. A mere difference in taste. However, my own news snobbery shouldn't affect what goes on the front page. Everyone then has their own 'view' of the top news. That way, if Hacker News sucks ass for you, it's because you voted for "haha news" and followed submitters' articles that you don't actually find interesting.

One of the arguments against this is that the articles would homogenize, because people would follow the same types of things. Though potentially computationally expensive, one might be able to characterize a user based on the articles submitted/voted in some parameter space, and then you can show a distance measure between this user vs you and all the people you follow.


I just wish we will remain a small community with common interests

It's interesting to see the comments in this thread while others are talking about optimal company size (http://news.ycombinator.com/item?id=144969) and why large companies can be dysfunctional (http://news.ycombinator.com/item?id=144904).


How about imposing a small karma penalty to post a new story?

Then, set a minimum karma requirement for posters. The karma requirement could be positive or negative.

It would eliminate spammers and reduce bad stories in one fell swoop.

I noticed a couple users tend to post a large number of stories which generate no discussion and earn no points, but there's no loss to the user and no prevention mechanism.


It is possible if you add proper trust metric, karma as it is right now is just a little bit too primitive for that.


What do you mean by a proper trust metric? I see systems of karma on sites like these eventually evolving into reputational economies so I'm curious about any way of making karma more complex.


Trust and reputation are always in context, current karma system ignores it. But once you add context visualizing and interpreting karma becomes hard. Take a look of how we're trying to deal with this problem (link in my profile, sick of being accused of spamming, sorry).


I have an account there but I'm unclear on what you're doing differently with reputation..


Utilizing reputations through clustering of voting histories in order to produce their recommendations is a pretty good description of what's going on there now. Next they're planning on giving you even more control by allowing you to control who you trust directly. This combined system should be the special sauce that really solves this problem, I think.


It is using machine learning approach to estimate what the impact of your feedback should be. That way it is depends on context (only topics / tags right now) and is gaming proof (i.e. you can give as much feedback as you want, that would only improve your reputation / predictions for you by the system).


How about this approach: the earlier a member signed up, the more their votes count in the rankings.


> I just wish we will remain a small community

This might be the problem? Simply the size?


so remove the karma and there's no more motivation to gain some sort of social ranking cred? interesting idea.


I don't know that could go both ways: Those who post in order to spread information for information's sake may continue to do so, while those who spam may simply go away because there is no "reward/recognition" anymore. However those who spam may also feel that they now have nothing to lose and may post more worthless stories even more often.


remembering the name of everyone interesting works less and less well as the site gets bigger.


"... remembering the name of everyone interesting works less and less well as the site gets bigger. ..."

I've noticed this as well. I keep a running tally of the frontpage with hacker, points, karma where I can sort the front page to read the top pages by person/karma. A quick scan of the front page shows me almost all are new names, most have karma of <700. Another check is the /news page compared to the leader board.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: