I think it's hard to talk about solutions when we don't have a solid definition of the problem. To define the problem well, I think access to the news.YC data would be very helpful.
So a meta-idea would be to open the news.YC database of users, submissions, posts, and ratings. That way, we can look for specific changes in voting/submission patterns. In particular, I'd like to know:
- how post length has changed over time
- how strong the correlation between post length and score is
- if participation has dropped among those who used to contribute the most valuable comments
- the average comment score over time
I don't see any downsides - news.YC isn't looking for profit, or to be acquired, so there doesn't see to be much use in protecting the data...
What is the reason of increased noise? Isn't it the popularity?
In my opinion, yes. Whenever the site got mentioned somewhere, a new stream of unimportant topics appeared.
Why can't or shouldn't the gravity be dependent on submitter reputation? Wouldn't we find any techniques to prevent gaming such a system? If the raw importance, not gaining another karma points, matters, shouldn't we consider moderation by selected hackers?
Sorry for asking so many questions, but I find these being very fundamental to any further considerations about limiting anything. If you want to do this, you have to have some authority.
I strongly support the idea of opening the HN database. This would let the others build interfaces around it, experiment, and hopefully find better approaches to limiting the noise.
As a summary, please note that in case of increasing the quality, one has to do any of the following:
- increase the signal,
- limit the noise,
- develop better techniques for dealing with poor SNRs (but I find this being unapplicable by humans).
To make things clear, my comment is solely about limiting the noise. I'd be very happy to hear ideas on actually increasing the signal.
Submitting relevant links increases the signal. Voting limits the noise. I've never done either despite following Hacker News since it came online. Now I'm going to submit and vote more often because I want to see it continue to be the best. If each of us does this we'll have interesting news to read every day.
It's easy to forget that we make Hacker News. We can fend off the garbage by remembering to participate.
This is getting ridiculous. I don't think I've ever seen so much self-congratulatory navel-gazing in any online community. Am I the only one for whom wading through stuff like this is as bad as wading through tasers & lolcats?
My idea is to vote for the submitter, not the story. Each user sees a different front page based on how he/she has voted in the past. I see more stories from people who have submitted stories I liked in the past. I have a test site running at www.votetheuser.com . You may notice a slight similarity to Hacker News - being the cheap rip off artist that I am.
What are the incentives to post something on hacker news? If the incentive is "get noted by YC so they'll fund me", then posting LOL cats is unlikely.
If gaining karma is the incentive, then LOL-cats have a chance, with a critical mass of supporters (more upvoters than downvoters).
A universal incentive on the net is to drive traffic to one's own web site. That also get's undermined by LOLCats, because as they drive away visitors to the news site, they drive away prospective referals.
Judging by my comprehensive list, karma is the main threat (a ka recognition by peers)?
What exactly is so bad about Reddit? I don't read it, so I don't know?
A guess: too many users might also result in repetitive submissions, as the new users are unaware that certain topics are already common knowledge within a community. Technological approaches might be possible, like better recognition of duplicates (analyze the content, not only the URLs).
>What exactly is so bad about Reddit? I don't read it, so I don't know?
The cliche is that reddit is turning into digg. I recently stopped visiting reddit because I think there is truth to the cliche. There are two dimensions to the the quality problem with reddit: content and comments. For 80% of the cases the content (i.e. the posted links) problem is solved: a few weeks back reddit did some excellent work on the algorithm to aggregate user's subreddits (e.g. politics, programming) and present links on the top page. Users seem to be happy with the change, so for example if someone doesn't want to see lolcatz, they just unsubscribe from the lolcatz subreddit. There are problems around people posting to wrong subreddits, but these are minor.
On the other hand comments are getting really bad. To understand this point, check out the "most dugg" comments on digg for pg's disagreement hierarchy essay. The top comment is just the word "No", and the 2nd and 3rd aren't that much better. So based on that I think whatever karma system you'll have, at the end it will reflect the overall sentiment of the user population on the site. For example, if the majority of the users believe that the comment "u r a fag" is a "good" response and give karma to the comment, then regardless of how you write the karma algorithm, the comments with the highest karma will reflect this primitive nature of the user population. Another example is youtube comments where there is a voting system, but reading comments makes you want to poke your eyes out.
There doesn't seem to be a solution other than to restrict comment voting/karma giving to a subset of the users who represent and vote in line with the behaviors that should be considered "good".
Are things really getting that bad here? I have been lurking and occasionally commenting here for quite a while and I haven't really noticed much change recently - the only thing I did notice was a slight slowdown in posts over the Easter period, which I might have expected anyway (but it may have been because I wasn't at my day job and spent more time here than normal!)
I left Reddit for the same reason I left Digg. Both have dishonest voting engines. Recently, Reddit was caught making comments only show up for the poster, nobody else. http://reddit.com/info/6drwl/comments/ Reddit also will blacklist your name such as the submitter of the above article, and then all submissions (not comments) points are reset to 0 every minute or so.
Ideas: I want downmod buttons because people post (and up) too much non-news. I think there should be points and total vote count. Then there can be a contrivercial page (ranked by points+total votes){or something like it.} Also there should be a "safe" non-contravercial page, a "worst" page, and a "zero gravity zone" where old good posts dominate.
communities must grow or they will eventually die off.
as communities grow their aggregate intelligence goes down.
the problem is to reconcile these two facts.
good luck, because society hasn't solved this problem yet either. The most successful societies have been those that spread the most effectively, not the ones that treat their citizens the best.
I think part of the problem are people who create polls such as the "what part of the world do you live in?", attempt to patronize pg in a thread, while talking about "meaningful HTML" when that person's site (it's the same character BTW) had HTML errors in it . . . to which he responded "it's not a bug - it's a feature!"
I mean, c'mon . . . that's kiddy, Windows user stuff . . . this is a site for HACKERS.
So a meta-idea would be to open the news.YC database of users, submissions, posts, and ratings. That way, we can look for specific changes in voting/submission patterns. In particular, I'd like to know:
- how post length has changed over time
- how strong the correlation between post length and score is
- if participation has dropped among those who used to contribute the most valuable comments
- the average comment score over time
I don't see any downsides - news.YC isn't looking for profit, or to be acquired, so there doesn't see to be much use in protecting the data...
How about it, Paul?