This, and many of the suggestions in the original thread, and pg's default response, are all falling into the same trap: they are attempting to fix a people problem using technical solutions.
Technology can make community management easier. Better algorithms and visual design and so on can improve the way that a community polices itself. But, I think eventually you'll run into vanishing returns from that kind of stuff. (I don't have any good science for this, but I've been spending too much time online since the dial-up BBS days, and community development is one of my interests.)
Eventually every community needs one or more managers, or gardeners (http://lesswrong.com/lw/c1/wellkept_gardens_die_by_pacifism/) -- people who have been around a long time, who embody the "spirit" of the site, and who have, not absolute, but special influence on the community. (RiderOfGiraffes immediately comes to mind; I really wish he was still around.)
This is absolutely analogous to Google attempting to avoid customer service completely by piling on more and more automated systems. While those systems help, they can't yet replace an actual customer service department -- and I think the frequent complaints from people who've been impacted by Google's lack of customer service bears this out.
I like some of the suggestions in this post, but I don't think that they can resolve the main issue with HN at the moment: it has grown to a size which needs one or more community managers.
There are some architectural issues too. I think the current trend in online forums, where new things are always better than good things, is ultimately a step backwards in terms of encouraging thoughtful discussion (http://www.robsheldon.com/conversations-online/). If a really interesting item comes up on HN that requires, say, a solid hour to compose a thoughtful response -- something very technical that would benefit from some number crunching -- there is very little motivation to commit the time to write that response. By the time it's written and posted, the site has moved on to the next new thing. However, the first person to respond to an item with a quick, witty, or snarky response gets rewarded with lots of attention.
And that is a problem that can not be resolved by any of the suggestions so far.
> By the time it's written and posted, the site has moved on to the next new thing.
I noticed this within the context of submissions and /new when I posted two blog posts of mine. Both posts were x000 words so minimum reading was over 3-5 minutes (average time on site was 1.5 minutes so not everyone read to the end).
Both times neither posts got any upvotes until half way down the first page. They were surrounded by topical submissions, news or etc. Stuff that either had a simple message in the title or was only a few paragraphs long.
Except a strange phenomenon occured when the posts fell to the second page, readership doubled. Whereas on the first page Google analytics real time reported 10-14 active readers this jumped to the low 20s. By this time the posts had received 5 upvotes. On page 1 they were maybe the 4-5th most upvoted submission but once they reached the 2nd page they stole 1st or 2nd place among their non-frontpaged peers.
The lesson I took away from this is that short or topical submissions have a strong advantage. It is easier and quicker for someone to decide to upvote. In same cases they may upvote without reading at all if the title is descriptive enough. Meanwhile people do like longer posts but reading a full essay is barrier to upvoting.
One way around the burying effect of the new page would be to not let it be ordered by age but delivered in a random order (but removed altogether after a certain amount of time).
Perhaps votes should be weighted by the length of the submitted post. So, e.g., votes on short submissions are worth less than votes on full essays. Length of the post can be gamed, but I guess when that happens some form of information theoretic measure can be used as part of the weighting.
I don't suppose this is already being done, is it?
Actually, I strongly doubt it can be easily automated in a way that doesn't reduce readability, which I expect to lead to a sharp drop off in number of up-votes. If the number of up-votes drops off faster than the advantage gained from the longer post, the system can't be gamed.
Certainly some of the suggestions in the article have merit. Encouraging people to browse and vote on stuff in new would be good.
But I agree with you. You don't fix human problems by adding technical controls. You fix human problems by adding technical measures designed to enable human controls.
My proposal is relatively simple: figure out a high karma number (high enough few people have it) and allow these people to issue an incivility warning. If you get more than a certain number of incivility warnings in a month you get banned, first temporarily and on the second time perhaps permanently. These reprimands would be public.
People need feedback. People need to know when they are stepping out. One of the real concerns I have with hellbanning (though I recognize in some corner cases it may be useful) is that it denies people this feedback and therefore does not encourage people to stay within the lines of acceptable conduct. So make hellbanning only for cases of people who are long-term repetitive problems or spammers. Implement a system that gives feedback to users for the rest. And allow the most senior people to enforce civility constraints.
Technology needs to empower people to manage people. Software should not manage people.
> "Technology needs to empower people to manage people. Software should not manage people."
I agree, in a way.
The only communities I have seen that associates numbers with human merit or trustworthiness are the ones populated largely by programmers. I feel this is telling.
I always have low scores on these sites because I am just not motivated to contribute to them. While Jeff Atwood was on to something when he said that users can do a good job when they feel invested in the quality of content on a site, we just cannot expect numbers representing reputation or consensus to always be set rationally. That, and I just hate thinking I will be assigned a score for trying to contribute to a conversation when I have no idea how the site's culture will interpret what I do. We may as well carry those 1-10 judge cards around our necks and stick gold stars on outsiders' foreheads. It's absurd... but apparently it's necessary. It shouldn't be, but I know I can't handle 1 million+ users as a moderator.
One of my old jobs was to facilitate a chapter of Socrates Cafe, a national discourse organization. We allowed anyone at all to come in and share his or her experiences. My job was to facilitate the discussion and put out flames fanned by conversations between opposing parties. So, when atheist/theists, right-wing/left-wing wars erupt, I was supposed to take varying measures to restore order without evoking enough negativity to scare away members. That is, I had to try and settle everyone down while making them feel BETTER about everyone else.
I cannot begin to tell you how ridiculously hard this can be, and technology does not help unless you use it for shameless enforcement.
Intuition and context-awareness helps tremendously, and there have been times where I simply could not think of any solution to one particular heated debate outside of, no joke, an air horn. I hated having to blast all of the words right out of the room, but the group we had (to me) represented a community gone wrong that needed flat out intervention. We never once used tools to try and change the rules of conduct. Instead, we used technology as a brutal and unapologetic last-ditch enforcement of existing rules.
Now... That's a group where everyone saw each other in person. Clearly not the same thing as an online group. That's worse. In this setting, technology is the only thing you HAVE to mediate all interactions, even the "human" ones.
I don't have a real solution to the sorry state of large, rude online communities. In fact, if you made me a moderator, I would probably go on a banning spree until everyone was too terrified to stay on their own accord.
Even so, but I doubt anyone else really has a solution either... Including the OP. Improvement can only be measured until online cultures change in a way to push newly set boundaries.
The problem, I think, is that engineers have been taught over and over to see the humans as outside the system. You have the system components, you have a user interface, and the user. This makes sense in a way because you can mathematically understand the components of your software, but you cannot mathematically understand the user of your software.
Recently I have started to realize just how wrong this is, first in reading what others have written on IEEE spectrum on the so-called "Automation Paradox" and later internalizing that business systems usually treat people as integral components. Even if humans are relegated to a supervisory role it is important to given them enough to do to maintain the system and failure to do so in vehicles can result in air travel catastrophes or ships running aground. But the same holds true for every other field. There is such a thing as too much automation. Especially where we are dealing with social environments, like HN, that human touch is what is most important.
I think in this the first question is what should humans be doing, and the second is how can the computer system enable them to do it. Unfortunately engineers tend to forget the first question and so never reach the second.
There are technological solutions to some of the problems.
For example, there is an increasing tendency to create dummy accounts for the sole purpose of dissing another user without getting your own account dirty. I was recently at the receiving end of this, and I have since noticed this happening more often all over HN. The solution for this would indeed be simple: don't allow new users to post comments for, say, 8 hours. In general these kinds of measures aim to make destructive behavior less fun, and I think they do work to an extend.
Then there are problems of style that can be solved by having guidelines.
Getting rid of "lolwut", "ORLY", or "Pro Tip:" comments is something we should all agree on; yet those are regularly upvoted, as are remarks that personally attack other users. This can be solved by declaring a ban on these kinds of things, because I believe the majority of users would honor those guidelines of they were made clearer.
Of course, the larger problem is that this even needs to be stated officially in the first place. It should be obvious that lolspeak, at hominem attacks, ridiculous trolling, and employing the basest rhetorical fallacies is not something that we reward here on HN, but nevertheless voters do reward them. That's indeed a problem of audience, not something that can be completely addressed by programming or guidelines.
Then there are people who see destructive criticism as the main problem, which I don't agree with. Personally, I try to keep my criticism constructive if possible and I do appreciate others doing the same. But when I want to receive feedback, I'm glad about any kind of response really as long as the tone of the discussion remains civilized.
The worst thing that can happen with feedback is not getting destructive criticism, it's the absence of any reaction at all.
On the "don't allow new users to post comments for, say, 8 hours":
Don't forget lot of people create a first account just to do a first reply. I would have feel not welcome the first time I try to comment if this protocol was implemented.
More than that, I'm not sure that your proposal will resolve the fake-account problem: if I love to be a jerk I will play with 2 accounts, your 8 hours delay will not block me.
I don't think it's fair to call any of the post's suggestions purely "technical solutions". They all involve changing the user interface and that can make a dramatic difference in how an online community functions.
(In contrast, pg's idea to assign different weights to comment votes would be invisible, leaving the site UI as is. That is a purely technical solution.)
I'm not clear on how changing the user interface isn't a technical solution, especially in contrast to such a low-tech thing as a community manager.
I didn't mean to imply that technical solutions wouldn't make a difference, even dramatic ones, just that they could not completely take the place of some manual effort.
If your definition of "technical solution" is as broad as to mean "any code change at all", then by definition, all changes will be technical. But at least the first two options here are social, not technical (though of course, as with any change on the site, they will require code to change them).
Just because a solution is implemented with technology doesn't mean it can't be a people solution for a people problem. There are a number of technological changes that will target the psychology of the people, guiding them toward different behaviour. I agree that the proposals in this post don't seem to address the actual issues very well but I think it would be a mistake to completely disregard technology changes as an avenue for affecting people changes.
> Eventually every community needs one or more managers, or gardeners (http://lesswrong.com/lw/c1/wellkept_gardens_die_by_pacifism/) -- people who have been around a long time, who embody the "spirit" of the site, and who have, not absolute, but special influence on the community.
1) Don't you think Wikipedia administrators and Stackoverflow mods fulfil that "gardener" role? (I think so).
2) Consider that above "gardeners" are often criticized harshly on HN (even though, in my opinion, they are a positive force on Wikipedia and Stackoverflow). Don't you think strong moderators would have a bad time on HN?
>they are attempting to fix a people problem using technical solutions
That's catchy but isn't that what a large handful of pg's startups are doing?
Don't get me wrong. I'm glad we're not throwing "technology" at the problem. But I'm glad for a selfish reason. I think adding features and modernizing the site would draw attention and users, probably the "wrong" kind.
> That's catchy but isn't that what a large handful of pg's startups are doing?
Yes, and I cringe every time I see the next startup come in trying to do just that. The reality is the people putting money behind this whole movement don't see it as you and I may see it. So as long as people are making money and they continue to prey on people's cognitive biases then we will continue to see this happen.
A comparable social site, reddit, didn't draw attention from the wrong kind of user by changing their site. They drew attention from the wrong kind of user because their biggest competitor changed their site (Digg).
Sure, some of the suggestions are technological in nature (purple for distinguished veterans). The author seemed very aware that tweaking the algorithms wasn't the solution.
Well, he did say, "here is what I think would benefit the community the most." But, I wasn't really responding to him directly anyway.
I just picked this particular thread because it was new. The other one was too busy and already halfway down the front page; the same comment there would have worse chances of actually being read.
This, and many of the suggestions in the original thread, and pg's default response, are all falling into the same trap: they are attempting to fix a people problem using technical solutions.
If all you have is a hammer, everything looks like a nail.
Yeah RiderOfGiraffes is cool. He is still around but was sort of hostilely driven into retirement. He goes by ColinWright these days. You're most likely to find him in math related topics.
The other thing is that his proposal was a proposal to "improve" Hacker News, not a promise to "fix" hacker news. Trying to do a grand massive "fix every problem" is more likely to introduce new problems than just making relatively safe improvements.
This is essentially the old argument over whether it's better to patch old bad code, or whether it's better to rewrite old bad code. I think all of the same responses would apply, i.e., whether or not new problems might be preferable to the existing ones. Since much smarter people than I have written carefully-considered essays on that argument (and without coming to a clear conclusion), I'll just let them speak for me. :-)
I think of looking at new submissions as a kind of "service to the community"-- more specifically, to the community I want HN to be. If the front page is filled with links I don't value, then it's partly my fault if I've failed to upvote submissions I would rather see on the front page (this, of course, means browsing "new" is a reward unto itself, but it's a somewhat abstract reward nonetheless).
Sometimes I wish HN wouldn't track average karma. That discourages me from commenting on new submissions that I believe won't get many eyeballs. Intellectually, I know my average karma shouldn't mean anything to me, but the fact that it's there for all the world to see exerts a gravity of its own. "You get what you measure."
I'm someone else who refrains from commenting in order to protect my average karma.
The thing is, that it isn't as irrational as you are making it out to be. Posts from users with higher average karma are sorted higher. Presumably you are posting because you want your post to be read. Either you want it to influence people or you want your question answered, or whatever. If you didn't want it to be read, you wouldn't have posted it.
I'm pretty sure this has a 'rich get richer' effect. On a page full of a lot of comments, most people will only read the first few. Being at the top leads to more up-votes, leading to more posts at the top, and so on recursively.
I am totally guilty of gaming this system. I normally only post in threads where I think that I'll get up-votes, and I only post things that I think are worth posting. The latter effect is good. There are too many people on HN and it's a good thing to encourage us to maintain a high signal to noise ratio. But the former, the fear of posting in unpopular threads, or on comments near the bottom of the page, is awful.
I suspect the average karma preference has a negative effect on quality, as you've pointed out. There is incentive for the best commenters to only comment occasionally, when they'll get a lot of karma for it.
Digg kept piling on stuff like that, making some users more valuable than others according to the algorithms, and I think it hurt rather than helped. I'm not sure it hurts on HN, but it doesn't seem to solve the core problem, either.
I think it does come down to a social problem being tackled by technical solutions. It may be that the only way to make this problem go away is to empower moderators (of which HN has more than a few) to simply remove the most negative stuff before it has a chance to alter the tenor of a thread. Hostility breeds more hostility and it spirals ever upward, until the conversants are more frustrated than talkative.
Wow, I had no idea average karma sorted people higher/lower.
But you actively try to protect your karma? After my karma hit 1000 I basically stopped caring anymore. What incentive do you have to protect your karma (serious question, I'm not sure I understand why it matters)?
I didn't know this, either. This idea will only work if karma measures what we want it to measure. If karma isn't associated with quality but with site-gaming, or snarky commenting, or other bad stuff, this system creates a positive feedback loop for sending the site in the wrong direction. Yikes!
I thought the sorting of comments was based on the votes for that comment in that thread, not based on the average karma of the poster. Is that not the case?
If 5 of us post in a thread at the same time, and no one upvotes any of the comments for an hour, then the ranking of the posts is going to have some kind of dependency on our karma metrics, rather than a "first come first serve" ordering.
Fuck average karma. Who even looks at it (besides the owner of the karma)?
I post a relatively high volume of comments on articles that get little visibility. Sometimes I answer an Ask HN type question with what ( I thik) is a good/useful answer and don't even get an up vote from the asker (but I have gotten 'thanks!' replies).
Oh well, I come here to discover new things and try and return knowledge I've contributed over the years. Karma has zero inherent value to me. The day I can spend it at Starbucks, the gas station, a sushi bar, or any place else I'll start go give a fuck.
Fuck karma in general. I can't think of a practical use for publishing the karma number. Real karma ripples throughout the world on it's own, you don't need a number to record it.
There's far too much ego worship in communities like HN, where nerds are often too concerned with it to begin with. The whole 'honoring the most popular community members' is the same. It's horrifically vain and just gives otherwise normal people's commentary much greater weight than it needs. Good comments will stand on their own.
Though you are correct in theory, I think the karma points are beneficial in driving the community. I didn't know what it was before I signed up, but now I check it out every now and then (though you can tell by my score, I don't obsess over it). Though maybe having a user know their overall score is a detriment.
It doesn't matter how much Karma is accumulated overtime, but consistently benefiting the community is valuable. Maybe we shouldn't be shown an overall karma score, but just how your karma has changed over the last month. While keeping th e overall score hidden from users.
I think levels would be better, taking the idea of a veteran to the next level. But I think the levels should be based on time and "karma". So you start off green, if you stopped on the site 5 years and never added anything you would be green but the same would be said if you were on the site for 5 minutes and got 1000 karma in a day, you would then move up a level or something but didn't get the "veteran" accolade. You then wouldn't be able to receive such higher levels without being a member for a certain period as well.
Maybe the overall levels relate to how much you can help moderate the site?
I know karma doesn't matter but I do find myself looking when I come onto the site and equally I don't post as often as I should because I find idiots down voting something for the sake of "happy clicking".
Life is too short to worry about votes and karma. If it really matters to you, just timing posting the right stories will net you hundreds or thousands of karma at a time. Then you can burn a couple karma expressing what you wanted to say even if it's unpopular.
But I don't understand the need for identifying who is a veteran or not. People who use the site will know who is a veteran and who isn't. In addition, someone without much of any karma will be treated probably poorer than many other users, or their ideas dismissed. Just because someone is new to the community doesn't make their comment less valuable.
The site will take care of the sorting/displaying of comments. There's no need for me to see the numbers that make the site run. It just ends up giving some people a false sense of importance. Keep the karma, but hide it, and maybe people will stop changing what they say just because of who they're talking to.
I'm actually amazed at the amount of time that happens actually. I've had the same experience. And I tend to upvote people who respond to my comments as long as they don't trash what I've said.
I guess some people hope, that high Karma will improve their chances when they apply to Y Combinator.
Actually, I vaguely remember a post or essay from pg, which suggested that one of the reasons to create HN was to assess potentially good founders based on the content they post on HN.
I'm not saying YC is actually skimming applications based on user Karma. But even if they don't, people might believe it.
Actually, I vaguely remember a post or essay from pg, which suggested that one of the reasons to create HN was to assess potentially good founders based on the content they post on HN.
Hahaha. PG should be doing the inverse. If someone is spending enough time on HN to have a really high karma then they aren't spending that time building something.
Yeah, I like to see my absolute Karma score tick upwards (it's how I know that I'm adding value to the community) but I don't stress out about saying something that might not be a huge hit to my average karma.
I always thought it would be interesting if you're 'average karma' number was rounded (floor(avgK)) to the nearest int and then applied as an upvote (or down vote) if that would change things. It is always dangerous though to embed in the system a feedback mechanism like this.
I downloaded the source once again to see how hard it would be to implement my 'directional multi-root' sort of karma system in HN [1]. Its one of my 'spare time' projects.
"Sometimes I wish HN wouldn't track average karma. That discourages me from commenting on new submissions that I believe won't get many eyeballs."
Maybe some sort of comment karma normalization based on the upvotes for the topic as a whole or the total number of votes on the topic, to try to adjust for popularity of the topic vs quality of the comment...
> I think of looking at new submissions as a kind of "service to the community"
Why not have the 10 or so newest submissions on the top or bottom of the main page, so that everyone gets to see them, without having to explicitly click on a different page?
e.g. if the new submissions are at the bottom of the main page, the "service to the community" of reading the new submissions is done almost automatically once you have gone through the list of the top submissions.
How about a mathematical reward system for new threads? Something like:
Vote Value = (Log[base:Median(AvgUserKarma)](AvgUserKarma)) * (Log[base:(25)](ThreadTop50Rank + 1))
That way if you contribute something great in a new thread and someone upvotes you, you'll get more karma from it.
A warning though, this reward system would encourage equal distribution of comments, versus the typical 80/20 rule. Not sure if that is what is best for the community.
Why is karma tracked at all? I know that it's used to unlock features of the site, but that doesn't need to be visible. The point of voting is for ranking comments and stories (which is great!), but it seems to have negative effects when public.
I disagree. I put a lot of effort into finding good stories and posting more useful comments when the /leaders page was a big deal, solely because I wanted to reach the top half of the page.
Not everyone is motivated by point scoring but some people are, and that's not always a bad or selfish thing.
I think this lays the real issue out but buries it:
> The only way to guarantee any visibility is to time very carefully using HNPickup, be a celebrity like John Gruber or Dustin Curtis, organize an up vote cabal, or write sensationalist content.
HN has become mainstream and it's subject to all the spam, pandering, submission strategies, power users and all the other bullshit that comes with that territory.
I'd start by getting rid of every user who's submission to comments ratio is ridiculous and every website that would be better suited to digg or reddit.
> spam, pandering, submission strategies, power users and all the other bullshit that comes with that territory.
that's actually what bothers me more about current HN state than the 'harsh comments' everybody complains about.
and one another thing: I remember when the frontpage was full of business and tech advice from experienced people who shared valuable knowledge hard to obtain on my own. now I wonder how many more git and fabric tutorials do we need, or why should I care what every blogger on the planet thinks of twitter.
Dude, I'm happy when I see a git / fabric tutorials. I'm always looking for new tips.
Looking at the current front page, the lonely VIM tutorial is literally the only thing a programmer is going to find useful right now. Oh, there's also a PCA overview using mostly Excel (two lines of actual code - from matplotlib.mlab import PCA; pca = PCA(x)), something about a new editor called "Lighttable" and a few sciency things, but mostly the submissions wouldn't even be worthy of Slashdot.
when it comes to technical knowledge on HN, I'm happy when I see a talk f.ex. how huge companies solve tech problems, or conference talks around specific technologies, because it's knowledge very hard or sometimes even impossible to acquire on my own - no matter how hard I would try, I won't have a clue how facebook handles massive deployments or how twitter scales unless somebody from the company publishes it and the community shares it. there is a vast ammount of technical advice/news to share, but toolset tutorials are one google search away, you don't need hn support for that, rtfm.
I don't think they're different issues. I think there's a bunch of people who want to be spoon fed arstechnica/torrentfreak/extremetech/etc just like they were on digg and reddit who cannot even pretend to be impressed by weekend projects or early stage startups.
That's why I like the idea of going after the users and sites that are better suited to other social news sites - the crossover in stories and topics between HN and http://www.reddit.com/r/technology is significant when it should be incidental.
> HN has become mainstream and it's subject to all the spam, pandering, submission strategies, power users and all the other bullshit that comes with that territory.
I think this article falls exactly under the heading of a pandering submission strategy. It's link bait. It's a semi-controversial topic trending on HN, better jump in and vomit up a blog entry to get some traffic! Nevermind that it says nothing at all.
Here's a thought for improving your image of HN: Read less of it.
Read a few submissions with interesting titles, post a comment or two and go back to doing something worthwhile. I know I waste too much time on HN as it is but many HN'ers spend way more time on here than me. With some topics pushing 200+ comments who has the free time to read everything? Don't. You'll have less noise to signal. You also won't get so emotionally caught up in the drama that inevitably accompanies human interaction.
Stop crying. HN is fine.
EDIT: 'Stop crying.' isn't directed at Benologist. Just realized it may come off that way.
Yes high submissions with no or low involvement in the community. Either they really, really, really, really love submission karma or they're here to exploit the site.
It should be pretty hard to automate spam comments (without the comments getting downvoted to hell). Whereas automated submission of stories is super trivial.
Average votes/submission seems a better gauge than raw number of submissions. Maybe there are people who are just really good an finding and posting a lot of stuff that's interesting.
Possibly off topic, but when you title a post "A Modest Proposal" it signals satiric content to a lot of viewers (as the original Modest Proposal was to defeat famine with cannibalism")
You are the second person to comment about this. I merely intended to humbly propose things, not allude to this historical work. I've changed it to avoid controversy (as that is not my aim).
The original - Swift, 1729: A Modest Proposal For preventing the children of poor people in Ireland,
from being a burden on their parents or country,
and for making them beneficial to the publick.
http://www.gutenberg.org/files/1080/1080-h/1080-h.htm
While not all of us can spend the time to become an HN celebrity, there are many of us who have worked to get a high karma over time. HN should highlight the names of such users in purple, similar to the way it highlights newbies with green.
Considering that the post is 3.5 years old, it makes me wonder if this problem has always been around. I wonder if this problem is cyclic and peaks at different times during the year or during certain events, such as summer or demo days.
Maybe instead of adding a comment, you could also have a button to add another link covering the same topic. Good for when someone links a BGR story that's just a linkbait headline summary of a BI article that's just a rewrite of a NYT interview. So let's say all articles for a common topic become linked. Allow upvotes on articles and the highest one becomes titular for the meta-thread.
This would (1) condense topics so the front page isn't deluged with say 5 twitter api threads, (2) unify discussions and avoid repetition, (3) help enforce good net practices i.e. linking to the best source.
I do think the "New" page is a significant part of the problem, but I'm not sure I would focus primarily on how to reward users for browsing /newest. I think I would lean towards trying to figure out how to improve its mix of content first. The often-low-quality content is why many people (including myself) have dropped off from reading it as much, so the two are closely related. The content needs to get to the level where someone reading /newest can rationally believe that giving each link a fair shake, by clicking and actually reading it (not spending 10 seconds skimming it), is worth their time. If people did that, more quality content would be unearthed, as opposed to just stuff that's appealing from the title and a 10-second skim.
One part is just spam-filtering, which is a never-ending arms race. But it seems low-quality even past that. Something to lower the total volume of the firehose might help; some people who don't even really participate in the community submit 5+ articles daily, or more. And there is a lot of reblogged content as well.
Could you incentivize people that upvote new items that ultimately cross some threshhold (front page/votes) such that their upvotes translate into some kind of karma or reward points of some sort? The earlier you are as an upvoter the bigger the reward? Of course that may just incentivize people upvoting everything in sight :) Perhaps historical upvoting patterns on the New Page could be incorporated to penalize that kind of behavior. Just throwing it out there. I do agree, I occasionally browse the new page but there is just so much cruft there that looking for the gems is time consuming so I generally just (lazily) wait for them to hit the front page.
Edit: Once again bitten by reading HN comments before the article as he proposes something along these lines in the post, doh! :)
No. I disagree that we need some sort of way to protect / alternatively value the old folks with valuable contributions. This is a bad idea.
At first glance it seems good, since the old & upvoted are known entities - on average, they are valuable to the community. Thus, give them something that others can aspire to, even if it's just a purple link.
I would caution that this will introduce strange, unforeseeable side-effects, not least would be that you would necessarily have to classify new users as dodgy, until such time that they have added enough data to be re-classified.
Doing that to new users might even be beneficial in the short term. But in the long run I fear that adding any frictions and subtle "you-aren't-as-valued-yet" around new users might introduce subtle behaviours and keep certain types of people away - something not easily quantifiable until it becomes too late.
I have a pet, untestable, and probably worthless theory that the tone of HN has slowly drifted toward the bitterness we see now because (a) you can't downvote links, so you have to express disagreement through the comments, thereby adding negativity to the comment thread that might have otherwise just been 'n downvote, and (b) the fact that comment scores are invisible doesn't give new adopters an idea of how strongly people agree or disagree with tone or topic in people's posts. So they don't learn by osmosis.
Both these ideas gave short-term benefit but I believe (totally unscientifically) that they slowly, invisibly, led to the tone of the site as a whole.
Be careful when changing the system, as the human part of it is difficult to manage via proxies like scoring and similar. And if we introduce a way that subtly penalises new people (or values old contributors, alternative view of the same idea) we might just change the way we attract new people - and a community needs all sorts of new people to stay alive.
It is an attempt to treat the symptom and not the cause.
I run large communities, and I've implemented "hide" and "ignore" functions for pretty much everything in my time.
People rub each other up the wrong way, so they want to ignore anything the other person says.
People get rubbed up by all threads on a certain topic, so they want to ignore the topic.
When a topic hits the news in a big splash and half the front page is the same thing, they get annoyed and want to ignore the whole of that topic for a short while.
Failing to hide/ignore all of that stuff results in them lashing out to the detriment of all.
But, if you indulge it by providing the hide/ignore all you've done is hide the symptom. Now people think it's cured and it is not.
The problem here isn't the content, it's the reaction to the content.
The problem here is with the reader and contributor, not pg and the code base.
The problem is that there needs to be a shift in attitude and etiquette from a lot of users.
It's a people problem, and people need to be shown what isn't an acceptable way to behave.
I shudder at the ramifications of some of these suggestions. Users do what you pay them to do. For example, paying people to up vote/down vote from New would result in just that -- not in people clicking the links and thoughtfully evaluating the content. And denoting high karma users more visibly would result in more karma boost activities. The end result of that is National Enquirer style headlines.
There is a change that would probably help: allowing a difference in signaling between agree/disagree and signal/noise. The definition of up/down voting is closer to ham/spam than agree/disagree, but people tend to use the arrows for agreement instead. It's hard to behave otherwise, by up voting a well argued comment you think is wrong, for example. It would be easier to highlight quality if multiple axes were available.
I agree that multi-directional voting is worth exploring and that promoting karma points leads to more sensationalism. Possibly even some form of democratic tagging?
While I agree that tptacek can be blunt to the point of cruelty, his arguments are well reasoned. I've also never seen him belittle the efforts of anybody (though I have seen him torch week reasoning). Even if I've disagreed and silently wished he would walk away from a thread, his contributions to HN are worse the bluntness.
I agree 100% (which is why I went with a silly "wat" instead of challenging the statement directly).
Every forum has a small group of people who tend to lack social graces (or what passes for them online) but aren't just random assholes and are correct most of the time. As long as these people don't stray too personal too often, their contributions remain a net positive, IMO.
tptacek certainly fits into that, he just seemed like a funny choice to lead off with in that statement.
I'm not sure whether this is a good or bad thing, but most sites (including HN) with partially-automated moderation systems make the same assumption:
*You only need one "karma" variable, for both posting and moderation*.
That is, users who post popular content (i.e. comments that get a lot of upvotes) and those who moderate well (i.e. votes strongly-correlated with those of the admins) are rewarded in exactly the same way, that is by increasing the value that unified "karma" variable.
IMHO this is not the best approach. For every prolific poster (who probably has decent karma already) you probably have 10 "lurkers" who have read most stories posted in the last month, know what they do (and don't) want to read, and are just as well-qualified to vote on stories and comments as those who regularly post comments (and earn karma from them.)
So what would happen if you made "posting karma" and "moderation karma" independent of each other?
(sort of a trick question, since it has been done.)
The title of your piece suggests a satirical tone and I was ready for a delicious morsel. Alas, I could not detect the slightest hint of sarcasm or of a darker, ulterior aim. It could still be there, mind you, but on subtler wavelength than my resonant cavity can detect. It might as well have been titled "HN Considered Harmful", I suppose.
"there are many of us who have worked to get a high karma over time."
I'm of the opinion that arguments must be judged by itself, and not by the "Aristotle said" conformity method that so many problems brought to society in the past and in the present.
I'm probably one of the older people here and had like 7 different accounts or so(I keep forgetting passwords and accounts when my machines change-update).
I had a big karma on some of those accounts, but honestly I don't care the least. I have more important things in my life to do-care, and so most of the interesting people in HN. A great PhD with great things to say is going to have work to do better than collecting "karma" on Internet all day long.
I'm relatively new to the Hacker News community, so forgive me if this is totally off-base, but in my opinion the issue is that the community feels relatively anonymous. Perhaps this is by design, the point being that the comment should be more important than the commenter, but it has its downsides.
It's easy to act viciously when you're only identified by a username, and this lack of identification, I think, breeds contempt. Though it's possible to dig deeper and find out more about each commenter, while skimming comments on an article, I don't "see" a collection of accomplished human beings having a conversation. At a glance, I don't know whether I'm talking to a bunch of snarky twelve year-old know-it-alls or founders with a wealth of real-world experience. To me, this matters.
I wonder whether there are simple ways to shift the tone of the HN community - for example, by assigning a short byline to each username that gives an indication as to who the user actually is, and why I should care about what they have to say (the "about" field already exists, but maybe it needs to be brought front-and-centre somehow).
Again, I've jumped into the HN community late, so perhaps I'm being a little naive. However, maybe a newcomer's perspective could be helpful.
I think there is a lot of identity associated with users. I'm relatively young, and often I will see one user identifying another user by their first name, which I don't know at all. People know each other here.
This might be bad. If so, I'd say all comments should be anonymous for 2 days. Give out a pseudo-random pronounceable handle to each user for posting on each thread. People making dick comments will still have to deal with their commenting history as posts are "unmasked" in 48 hours. (This well could cause more problems that it solves.)
That's interesting, and it's probably something I'll come to recognise as I spend more time here. However, from where I'm standing, it does seem as though the issue is that those ties are breaking as the community grows. I'm not totally sure adding a further layer of anonymity would strengthen discourse (though it could be an interesting experiment).
The 'more' link regularly fails to render on my IPhone 4 leading to a further long term aversion to going there. Not sure if anyone else has had the same experience? Just saying if it's affecting many devices this Is just compressing the issue further.
I have a different problem in that the 'more' link uses some crazy key instead of 'page=2'. If you press it a relatively short time after being loaded, you'll get 'unknown or expired link', meaning you need to backtrack, reload the page, scroll again, click... it's trained me out of bothering to do it anymore.
In my experience the _more_ link works for about 5 to 10 minutes. If you slowly scroll down the page reading things on it, by the time you get to the bottom of the page it no longer works. You have to reload the page to get the link to work.
I'm an algorithm designer, so I don't know about all of these changes but I do have a mathematical solution to PG's vote value problem:
before edit: Value of vote = Log[base:TotalUserKarma](UserAvgKarmaPerComment)
after edit: Vote Value = (Log[base:TotalUserKarma](AvgUserKarma)) * (Log[base:(25)](ThreadTop50Rank))
another edit: Vote Value = (Log[base:Median(AvgUserKarma)](AvgUserKarma)) * (Log[base:(25)](ThreadTop50Rank + 1))
EDIT: added a +1 to prevent a 0 value for the #1 thread. (unless you actually want votes not to count for the number 1 thread, in that case you can just pull out that "+ 1")
This would reward contributors with higher relative average karma per comment. If you have 0 karma your vote isn't worth anything and if you do have karma but your average is only 1, once again your vote isn't worth anything. You have to consistently contribute useful material to have a say whether or not someone else's material is userful to the community.
This seems to be a decent reflection of social circles in real life.
Try it out on users you think contribute very little and see if it is effective. Naturally, I don't have that data so I can only speculate.
The problem with average karma is stat it strongly discourages people from posting replies to less popular or older messages. I've made a conscious choice to ignore my average karma. I dont think it is conducive to good conversation, as it tends to be heavily biased towards the first few posts.
I wonder if we should go the other way -- I always liked slashdot's system of a 5 point cap on any post. Is a 500 point post really enter than a 100 point post, or is it more likely it is seen more often?
This is true, but it is a problem that currently exists.
For orthogonality's sake, I would suggest that if most of your posts come from 1:1 conversations, post by post, then perhaps there is a better medium to communicate with that person than through a board like HN.
The people that I have those conversations with are people I only know through HN, and the discussions come out of what happened here. Those discussions would not have naturally happened in another medium.
As an example I give you this discussion. In this thread I have 2 posts, both with 1 karma right now. I do not expect them to go up much if at all. Where would you suggest that I should have gone for this conversation with you?
EDIT: I changed the evaluation to not make the avg karma relative to total karma but instead to the median value of average karma of all users. The downside to this is that Median(AvgKarma) would need to be calculated quite often throughout the day because I doubt pg wants to pay for the resources to run that everytime someone votes.
The problem here is that someone who consistently provides high-quality (or at least frequently-upvoted) comments / submissions will likely have less of a contribution than someone with significantly lower karma. For instance: you would have (slightly) more influence than someone with 3000 karma and an average of 5.
Then this will reduce the amount of impact they have on the community as a whole. They will be doing us a favor by starting out ruining their karma.
Also, to state the obvious, every member here was a new member at one time and I'm not ready to say that is "typical" behavior. It's just very memorable behavior.
The whole reason why I submitted this was because PG said it wasn't the comments that cause issues but that bad comments were getting upvoted too much.
Actually, karma is very good market indicator for any board and filter bubbles tend to be specific to filtering out certain types of content. This won't be filtering anything by content.
"What doesn't matter"
...
"Complaints of 'too much startups, not enough tech' or vice-versa"
Is there a place to go for real hacking articles and none of the startup news bullshit? I don't care if it's called Hacker News or not. I don't even care if it's /new/. I just want a place to discuss interesting articles about making software and hardware do interesting things.
The site was originally called startup news. I think that in retrospect the rename to HN was a bad idea, because it changes peoples expectations of what the site is for.
Regardless, what you are describing would be much appreciated.
I find that paradoxically, the biggest problem with Hacker News right now is the hot algorithm we use. It's great for bringing good stories to the attention of the community and having a burst of short lived discussion over those stories, and then dying out. What it's not great for is long term discussions.
For example, with Quora, when someone answers a question, or posts a comment, you never care how long ago it was posted, because there are notifications, and the front page is based on what you follow, not how popular the questions are.
This is a hard problem to fix, but a couple suggestions I have are to improve the threads feature, such that it resembles a real notification bar. Secondly, I think a new page should help, a mixture between the front page, and the saved stories, which shows active discussions, and an emphasis on one's that you have upvoted.
I think those might help reduce the ephemerality of the discussions here.
How about this: charge a subscription fee to perform "destructive" actions on HN. This means posting comments and voting, including upvotes. Perhaps submitting stories can remain free — the paying users will determine the quality of the submissions.
Some of us (used to) get a lot of value out of HN, so $10 or $50 a year would be a bargain.
I would not blink twice at this. Few better ways to weed out the trolls. It would be interesting to devise a way to utilize money and karma in a way that basically says, "If you are provably improving the community significantly, this costs you nothing; but if you are average or, worse, a drag, you will pay."
Agree. Creditcard payments, however small, seems to be the by far most efficient way of filtering spam. CC payments make banning have a real effect.
Metafilter is the NIH syndrome on HN.
The problem with the New page is the extreme volume of entries due to everybody being able to post. Voting only goes so far, it doesn't work when you have too much stuff and users feel that their time spent voting doesn't make any significant difference.
The signal-to-noise ratio on the New page must increase before most users will go to the trouble of voting.
And now all your opinions on the internet are linked to your credit card... while I'm not so concerned about HN, I don't think this is a great solution in general
Between the to bar and the content, for a week or so, place a centered link reading "A personal message about the future of HN from Paul Graham." Make it a thread maybe. Talk about the expectations of visitors. The need to build up, rather than drag down. To hold back the snarky comments, no matter how funny we find them. (And not to vote them up, no matter how funny we find them.)
Talk about the virtues of reddit's better maintained communities if you're not looking for serious conversation.
Make the post exactly what you want from the community.
Treat them like the serious business and tech minded sluts you want them to be woke under your roof. And point that out too. You're choosing to engage, like you often do. You don't want to, and shouldn't have to, seek technological answers to people problems.
Maybe you'll convince a few people to think before posting/voting.
Surely one badly-needed feature is needed is more different views. As the community gets bigger, the single 'front page' doesn't cut it anymore.
Reddit's subreddits have become enormously successful in this vein (probably to the detriment of its main page, which nobody in their right mind would read anymore). I'm not saying this system is perfect for HN, but other systems in the same direction might be very nice eg. tags for things like 'please provide feedback on my new page/product' (I'd love to see them on one page), or different types of comment view on a per-user-basis (how do I want to see comments, not one system for all HN). For my mind, an encouragement of diversity along these lines would help the site thrive whilst allowing users to keep (and define) their own community standards.
Wrong. Instead user karma system should be deleted completely. (Yes, I know that I am a "newbie" here. Think what you want.). I have seen too many portals died or degraded by the weight of "karma". The only working example is probably Stack Overflow and I think they won't last long because of it.
Only reason why SO can survive is continuous strict moderation. I like "C++ vs Java", "Most influential book" and similar topics very much but they really don't belong to SO by their rules.
The same should be applied to HN - ban those millions Twitter and FB topics. Ban topics with complaints (however reasonable). Ban SOPA/ACTA etc. topics.
In short - filter everything that is not educational for IT specialists. Yes, it would be more boring that way (or not) but it would ensure HN quality and survival.
There is a tendency to upvote comments you agree with, and downvote comments you disagree with.
This has two ill effects:
1) Witty and popular viewpoints rise to the top, creating an atmosphere of superficial groupthink
2) People with unpopular (but insightful!) viewpoints may be discouraged from posting them.
Given this, would the level of comments improve with the following change?
Replace the up/down vote buttons with a single button next to each comment that reads: 'insightful', and use that to vote up comments. Forget about downvoting stuff - the low content viewpoints and rudeness will both sink to the bottom. Or if you wish, add a 'flag' box for comments made in poor taste.
Using such specific labels would make people think twice before voting, both about the purpose of their vote and if the comment needs a vote at all.
1. Don't let people create multiple accounts so easily. Switch to OpenID and have the ability to associate old accounts with the new ID or have the old account deactivated within 90 days. I don't even know how many accounts I've created, but it is a lot. This is basically anonymity. This does not work.
2. Don't filter comments and not hide topics that anyone disagrees with unless they are extremely offensive, and those could be marked as offensive and then PG and those with enough karma will check those and if they are offensive, they will hide them, but anyone could still see them, vote on them, and comment on them, if you clicked a flag to see them. You have to login to upvote, downvote, mark as offensive, post, or comment.
3. Show who upvoted or downvoted to everyone. This will make people think twice before downvoting.
4. Allow private communication between users so that if someone downvotes you without saying why, you can ask them why.
5. The new page is grouped by: last 24 hours, last 3 days, last week, last month, last 3 months, last year, and then by year. Everytime something is upvoted it goes to the top of the new list for the group it is in.
I'm sorry that some people have complained that HN works fine technically as it is. It certainly does, for those it serves. However, people complain that there are negative views here. I think the changes I proposed will help with that. But a bigger problem is that people's posts and comments get whacked. If you handle things in the way I described, and stop silencing me and others, I think the community will be much happier.
It's unlikely that a purely algorithmic ranking solution will be adequate. HN should better leverage the judgment of the site's veteran members across the board, rather than only for article ranking. Awhile back I posted a self-organizing way to filter comments that works in this way: http://news.ycombinator.com/item?id=3473753.
>HN could benefit from implementing infinite scroll, so the next page is appended asynchronously to the current one. [...] The problem is only a small minority of readers bother to click the tiny "More" button on bottom of the page.
Yes. More so, if I spend considerable time, reading the second or even third page I can't get to the third or fourth page because the "More" link becomes broken after a while.
I suggested this on a previous 'improve HN thread' (instigated by pg himself):
Shutdown HN for a week or so and when reopening only let a certain quota in per day/week/month. Give priority to old and/or high karma members. I think that it will have the effect of churning the low quality membership.
Perhaps allow these elected to elect others to enter "promised land".
I think showing veteran users in another color would really help a lot to set the tone. If the community rewards snarky posts with many upvotes, this alone will not be enough though.
Perhaps marking up veteran users by karma and/or seniority and weighting upvotes by the same metric?
I agree with showing certain users in certain colors. I don't think karma is the right metric for it.
I'd like to see green for new users, maybe some specific color for throwaway accounts of someone with a real account, let people with special other colors reveal that color if they wish, and maybe good and bad colors for people who meet other requirements -- there are some people who probably deserve a "red flag" for 30+ days, but not hellbanning, and others who deserve the honorary purple.
The only karma related way I'd award purple is for submission points, or for being early upvoters on submissions in /new which are later upvoted by a lot of other people. Sheer accumulated comment karma is probably not a great metric, since someone who is on HN a lot, chooses to write "populist" comments, etc. is likely to accumulate a lot of karma without adding much value.
> I think showing veteran users in another color would really help a lot to set the tone.
Why should we set apart veteran users? Only because they are longer here doesn't mean they have a more important opinion on a certain topic.
Setting them more apart is like having village elders. Even if they talk about something they have no idea of they are still right because they are so old?
I think a lot of new users are unaware of the HN guidelines [1]. Linking to it from the nav bar could help showing what is and isn't an acceptable way to behave.
Personally I think I'd raise the bar to moderator rights to those with several thousand karma, and encorage strongly to flag stories they think are inappropriate. That should get the site closer to what it was a few years ago.
What criteria would make someone purple ? Popularity ? Number of karma points ? What would be the incentive to respect them more then others ? And why would they deserve more respect than others ?
I'm a long time HN reader and I don't have much karma. Its because I try to post and comment wisely. Why wouldn't I deserve respect ?
My impression is that radical comments are more prone to votes since they pull on the emotional strings.
It would be a bad idea to favor people voting in the new section because I bet people would vote blindly just for the benefit of voting.
Why not show the "new" tab as the front page and put "top" as a menu option? I know it's one more click for a lot of us, but it could be a nice change! Or, at least make that the default.
There are a handful of domains that host pretty shoddy content - poor research, poor writing, link-baity titles, bias, etc. Banning those might be too extreme. But it'd be nice if there was an A list and a B list and then a A and B list.
the A list would be carefully curated set of URLs proven to be of interest to the community (but a domain can easily be removed if they start getting too many bad stories); the B list would be everything else' the A and B list would be both.
While reading this article through HN, (http://slifty.com/2012/08/a-tor-of-the-dark-web/) I thought it will be a good idea if HN can be moved inside an onion network like tor. People need to go through tor in order to access the site. It should not be visible via normal http/ip addresses. This can reduce the number of users coming on HN, and hopefully HN maintains the authenticity.
I think "reward for browsing New" could be a good idea yet very difficult to implement it well. What I'd like to see is a front page with various lists of submissions according to different criteria, some ideas: 1st| most votes this month - this week, 2nd| "Hot today" (this'd be similar to the current front page I believe), 3rd | "Promising" (Posts from New section that begin their momentum), 4rd | Most commented month / week. Basically a bit of filtering.
Sridatta, thanks for taking time to articulate your ideas on how to improve Hacker News. I enjoyed your post and really like the idea of distinguishing veteran users.
But is there a "Hacker News Manifesto"? No there isn't. HN is what it is because of the community. Sure, there should be moderators to control the profanity level.
But overall, if content is interesting and it's up-voted, it means it's meaningful to someone and therefore should maintain its place on HN.
The content should be driven by the community - if it's too startup related or too technology related is really a subjective matter.
My immediate reaction is to say that the downvotes serve a useful purpose. They can indicate social disapproval without adding noise or fanning flamewars.
The downside, of course, is that you see so many "WTF? Downvotes? Why" Comments.
Honestly, I think this is kind of much adieu about nothing. On the original thread, one commenter looked at the OP's original submission and found a lot of praise etc. Similarly, wheher I see ShowHN posts with projects that seem trivial or even frivolous to me, there are usually many well wishers offering a lot of praise.
> only a small minority of readers bother to click the tiny "More" button
The size migh be an issue, but the greatee problem is that by the time I reach the bottom of the page, the link has usually expired. If you want your users to see the next page, that really is inexcusable.
I would also remove the duplicated content/submissions which create or could create fragmentation of good comments, besides a personal frustration for having just wasted time opening again the same link only because the title seemed different.
Add to the list: Keep HACKER NEWS FOR AND BY HACKERS..stop posting other useless content. I want this place to learn and read about new coding practices, api's, only technically relavent stuff.
This is a good piece on why aggressive moderation is necessary to keep online communities from devolving into cesspools, even though it may attract cries of censorship:
I never thought about it until now, but the best IRC networks I hung out on in the 90s were ones with really good net admins who enforced the rules. Even when I was younger, I didn't belong on certain IRC networks, and I was quickly banned from them. That same rule applies to all the other communities (mostly gaming) that I was involved with.
One problem with banning users is that they're not difficult to prevent from returning. At least on IRC, you had to pay for a BNC to change your IP address because whatever your dynamic IP resolved to, the hostname was fairly easy to reliably ban.
Technology can make community management easier. Better algorithms and visual design and so on can improve the way that a community polices itself. But, I think eventually you'll run into vanishing returns from that kind of stuff. (I don't have any good science for this, but I've been spending too much time online since the dial-up BBS days, and community development is one of my interests.)
Eventually every community needs one or more managers, or gardeners (http://lesswrong.com/lw/c1/wellkept_gardens_die_by_pacifism/) -- people who have been around a long time, who embody the "spirit" of the site, and who have, not absolute, but special influence on the community. (RiderOfGiraffes immediately comes to mind; I really wish he was still around.)
This is absolutely analogous to Google attempting to avoid customer service completely by piling on more and more automated systems. While those systems help, they can't yet replace an actual customer service department -- and I think the frequent complaints from people who've been impacted by Google's lack of customer service bears this out.
I like some of the suggestions in this post, but I don't think that they can resolve the main issue with HN at the moment: it has grown to a size which needs one or more community managers.
There are some architectural issues too. I think the current trend in online forums, where new things are always better than good things, is ultimately a step backwards in terms of encouraging thoughtful discussion (http://www.robsheldon.com/conversations-online/). If a really interesting item comes up on HN that requires, say, a solid hour to compose a thoughtful response -- something very technical that would benefit from some number crunching -- there is very little motivation to commit the time to write that response. By the time it's written and posted, the site has moved on to the next new thing. However, the first person to respond to an item with a quick, witty, or snarky response gets rewarded with lots of attention.
And that is a problem that can not be resolved by any of the suggestions so far.