It's kind of amusing that the entire premise of this article is undone by the aside:<p>> <i>update: Apparently it is possible to downvote comments, which I never realized.</i><p>But while downvotes seem useful for <i>comments</i> where there's a lot more scope for false information, trolling, etc., I'm not sure what the additional value would be for <i>posts</i> -- they're either popular or not, and you can still flag posts that are actively harmful.<p>And then this:<p>> <i>Is it realistic for users to expect to post in an environment where there are no penalties at all, no way for their peers to express disapproval or disagreement with their post?</i><p>Huh? Peers express disapproval and disagreement by <i>commenting within the post</i>. As I am doing precisely right now. ;)<p>Jeff Atwood is a super-smart guy who's done amazing things but this particular piece leaves me quite baffled.
I've been through several forum deaths. The cause is usually large growth from people with very different mindsets that downvote for essentially political reasons instead of encouraging good discourse. This makes the best people leave in short order.<p>In a sense there's no way to structure a forum to avoid this. Being outnumbered by people who don't value open communication is unpleasant with or without downvoting. Forums that are highly specific to a small audience can last a long time. Popularity is the death-knell.
Getting it wrong for 12 years now:<p><i>The Value of Downvoting, Or, How Hacker News Gets It Wrong (2009)</i> - <a href="https://news.ycombinator.com/item?id=25633668" rel="nofollow">https://news.ycombinator.com/item?id=25633668</a> - Jan 2021 (1 comment)<p><i>10 Years In, Was He Right? “Value of Downvoting; How HN Gets It Wrong”</i> - <a href="https://news.ycombinator.com/item?id=23409231" rel="nofollow">https://news.ycombinator.com/item?id=23409231</a> - June 2020 (1 comment)<p><i>The Value of Downvoting, or, How Hacker News Gets It Wrong (2009)</i> - <a href="https://news.ycombinator.com/item?id=13325726" rel="nofollow">https://news.ycombinator.com/item?id=13325726</a> - Jan 2017 (12 comments)<p><i>The Value of Downvoting or How Hacker News Gets It Wrong (2009)</i> - <a href="https://news.ycombinator.com/item?id=10875619" rel="nofollow">https://news.ycombinator.com/item?id=10875619</a> - Jan 2016 (36 comments)<p><i>Reddit's Discussion about HN</i> - <a href="https://news.ycombinator.com/item?id=508801" rel="nofollow">https://news.ycombinator.com/item?id=508801</a> - March 2009 (24 comments)<p><i>The Value of Downvoting, or, How Hacker News Gets It Wrong</i> - <a href="https://news.ycombinator.com/item?id=507948" rel="nofollow">https://news.ycombinator.com/item?id=507948</a> - March 2009 (114 comments)
Interesting, making downvotes costly seems like a good system. Though downvotes seem to work pretty well as-is for HN, I rarely see dead comments that were killed by downvotes rather than anti-spam features.<p>If we take HN as reddit 2.0, what would be HN 2.0 I wonder? If I were the one to make it I'd create a system to promote the comments of people with professional expertise on the subject being discussed. HN has mostly solved the politeness and flamewar issues but IMO one its major flaws is that people speak authoritatively about things they don't know about.
I think a critical missing context of reposting an old 2009 article is that a <i>newer post in 2011</i> makes downvotes <i>free</i> instead of costing reputation:<p>>Jeff Atwood announced change of policy -- <i>"Downvotes on questions no longer cost the casting user 1 reputation, so they are effectively “free”. [...] So, it’s imperative the question list have a high signal-to-noise ratio, and removing the penalty for those users who do take the time to read a question and later find it to be useless so they can down-vote is conducive to that."</i> -- excerpt from : <a href="https://stackoverflow.blog/2011/06/13/optimizing-for-pearls-not-sand/" rel="nofollow">https://stackoverflow.blog/2011/06/13/optimizing-for-pearls-...</a><p>Why? He saw that users were <i>too hesitant in downvoting</i> which allowed bad content to grow and further degraded the site. Changing the downvoting mechanics fixed that.<p>(I don't know if there's been an update since 2011 to the downvoting system and incentives.)
The glaring issue in this article (which I couldn't bring myself to finish) is that he equates downvoting to "evil or incorrect post(s)". Which would be true in an ideal world, but the reality becomes downvotes are equivalent to "I disagree with you".<p>If plentiful downvoting is allowed, discourse becomes a popularity contest. Fortunately most of the experienced users here seem to understand this and reserve their downvotes for appropriate situations.
No system will be free of problems or gaming.<p>But, doesn't hacker news actually have implicit down voting by way of time? HN isn't designed to be a searchable Q&A or forum. It's a "what's a currently popular" site. If something isn't good, it just disappears faster than the opposite.<p>Have I wished I could down vote things sometimes? Sure. But then I just wait 5 minutes, refresh, and it's gone.
I really like downvoting where you lose one of your past upvotes in order to downvote. It's a nice little filter that people who care about how many upvotes they have won't use, and puts an end to some amount of frivolous downvoting.<p>In a meta way, you kind of have to use a little bit of your own point capital to downvote something, so if you're highly upvoted by the community then calling BS on something is easier for you and the risk taking for doing so makes more sense, versus someone who just signed up.<p>I also wonder, for a site like hacker news what's the point of even showing the user the # of upvotes they've obtained?
I don't know if these are popular opinions or not:<p>1: I really wish I could downvote Hacker News articles.<p>2: I really wish there was some friction to downvoting. Perhaps something like choosing among: disagree, factually wrong, mean, incoherent<p>Regarding the downvoting of articles: For awhile I used to see a regular pattern of weird articles on the front page. The discussion would then predictably involve a mod defending the article. IMO, I think downvoting an article would help in this situation.
I initially loved the idea of voting posts up and down. Thanks to seeing how it actually works on SO and here, well, I am not much of a fan anymore.<p>At the very least, if I ever implemented it myself, I'd require that people explain why they are voting up, or voting down, before they could vote. And I'd not limit users from being able to vote until they magically get to the right amount of "karma".<p>Without requiring an explanation, votes are more akin to boos and cheers than anything useful. Not a very good way to encourage a diverse culture.
What I find interesting is that users voting on other's comments seems to have become the assumed best way of organizing user-generated content. In my mind, I think the question isn't "do we allow or disallow down votes," but "should we even let users vote?"<p>My impression is that voting systems ultimately end up as popularity counters, and that in and of itself can skew the motivations of participants. Additionally, there's motivation to maliciously hide opposing viewpoints, late-comers rarely get to meaningfully participate due to their "score" starting lower than longer lived and thus higher scored content, and well-intentioned discussion about contentious or nuanced topics can easily get buried.<p>What the "best" way of sorting user content is will surely depend on the goals of the platform, but if deep, informative discussion is the goal, maybe using a popularity contest isn't the right strategy? Perhaps giving the users choices of sorting by new/old, most/least discussion, most/fewest replies, etc... might be a better option? And/or some other non-vote driven method? Of course you'll still need good moderation and a way to flag inappropriate content where a user's ability to flag is kept in check by some combination of the length of time that user has been active, his/her level of participation with the platform, and the number flags said user has received, but overall I suspect this would keep the general user's focus on having actually productive discussion instead of trying to get that dopamine hit from making a well-liked comment.<p>I should also say that I've been mostly pretty pleased with the quality of discussions on HN and I appreciate the good moderation and the general populace's desire to share quality information/opinions/thoughts. But I do wonder what will happen if it becomes "too" popular and if there are things that could be done to keep that focus on quality discussions even in the face of a massive influx of users.
(2009), by Jeff Atwood. I would have been <i>very</i> surprised to hear today's Stack Overflow staff talking about the value of downvoting.<p>Atwood's experience with downvoting on Hacker News is definitely a bit dated as well:<p>> (update: Apparently it is possible to downvote comments, which I never realized. It is buried in the faq)<p>> (I apologize for my misunderstanding, but there’s no visible UI for downvoting, and I can’t recall ever seeing a single negative voted comment in all the times I’ve visited Hacker News! Also, I put these comments in parens to make them extra-LISPy so Paul Graham would see my corrections.)
I've been downvoted to [dead] before, though I've never gotten mad about it.<p>One trend I have noticed is that on HN you won't necessarily be downvoted for your idea but you definitely will be downvoted for how you present it. If you come off as vitriolic or angry, you can expect a proportionate response. Frustration seems to be tolerated intermittently.
I like the idea of paying to downvote, but also an interesting effect of downvotes is that they skew views of a position.<p>Having one upvote and 100 downvotes is the same as having 401 upvotes and 500 downvotes. Isn't that odd? One is a comment everyone hates and the other is a comment almost half of all people like.
> We’ve lost half the potential information. If a post has zero upvotes, does that mean it’s bad? incorrect? uninteresting? mediocre? There’s no way to tell, because zero has multiple meanings.<p>Having downvotes doesn't fix this problem. You still loose information. If a post has zero upvotes, does that mean nobody cares, or a lot of people love it, and the same number hate it? The only site that gets this right is youtube: It shows both likes and dislikes.
One thing I noticed recently about HN voting is that it seems like you can upvote but cannot downvote immediate replies to your own comments. Other users can still downvote the comments. When I noticed that it struck me as very clever and something reddit should immediately poach. It's like a built-in cooling off period.
Personally I'd prefer if comments displayed upvotes and downvotes as separate tallies instead of a total. It means something very different if a post as a lot of up and down votes that sum to about zero or if it has no votes at all.
I think one obvious problem with downvoting where downvotes automatically hide content is that it creates echo chambers. The downvote button is definitely a 'disagree' button here and I think it's not really conducive to great discussions.<p>Then again, this isn't really a forum - it's transitory comments about the news of the day. Perhaps the argument is that there shouldn't be substantive discussion here in the first place. You're clearly meant to respond to the article more than the individual comment - the tree style commenting system and automatically hidden nature of comments down the chain both speak to that.
I never comment on HN anymore, because, ironically, a site ostensibly promoting innovative thinking actually discourages it with mechanized "group think". And ironically I am breaking my rule to never comment, on the completely off chance the observation might sink in. FYI, the down button is right there, don't hesitate, you wouldn't dare want to ever take input that you don't completely agree with.<p>Last post ever, no really this time :)
I find the "I disagree with you so I'll downvote you" approach to comments is quite useless. Should require a comment in the very least if you have any concept of adding value. Especially in political discussions I've seen. Its blatantly obvious someone is downvoting simply because they disagree. That should be trackable and that type of action flaggable in itself. But I understand the overhead it would create.<p>Flagging without reason is likely equally obnoxious.<p>If the downvoted person nas to wonder about the reason "why?" then the system isn’t helping.<p>I've seen plenty of faded out, non-snarky, non-trolling comments that were actually on-topic and relevant. But enough people didn't agree. Making them hard to read just means I now apply a extension to find out why it is being blocked. Often the faded out comment is informative <i>specifically because it annoyed or offended someone who disagreed</i>.<p>Looking back at some of my own comments that were downvoted, I have no idea why it happened. I asked others for their opinion. They have no idea either. So what was their intent? I literally don't know why. No self reflection occurs because there's no clear reason.
Interesting article.<p>From my perspective HN voting, for the most part, works very well. I've been on online discussion groups since the days of USENET. HN is great.<p>What really bothers me is zero information down-votes and the "downvote army" as someone called it in this thread. Nobody learns anything from this at all. What's the value when a downvote carries no consequences whatsoever for the voter while delivering a penalty to the author?<p>One idea I had a while ago is that a downvote should cost the person money in the form of karma. Want to down vote? It'll cost you ten points. No problem.<p>The point of this idea is that a down-vote out of spite or just to pile on someone or an idea is a bad thing. People who behave that way with some consistency would have to give up more and more for each downvote.<p>Paying for a downvote with karma could be an interesting evolutionary solution for downvote abuse, particularly if the cost increases based on a range of criteria. You downvote a comment on a thread and it cost you 10 karma, you continue to downvote the same author on the same thread and it cost you more each time. If people are piling one, the cost goes up. Etc.
>I can’t recall ever seeing a single negative voted comment in all the times I’ve visited Hacker News<p>That's interesting. I wonder how that could be the case.
Jeff’s 2009 argument against HN is a theoretical one, and IMO wrong. He points at a system that works well (HN), and says it “gets it wrong,” for made-up and unproven reasons.<p>Paul’s point about not being Reddit, I think, is you need good moderation, and you need to be willing to block and ban. Despite this, Jeff assumes HN limits downvoting in order to not hurt people’s feelings or something, which is a total straw man or non sequitur.<p>From my armchair, I gather that good communities require good moderation, so if you are trying to scale moderation, that’s your work right there, and a downvoting mechanism doesn’t take the place of that. Ideally mods are people who have organically earned trust and power or are known and trusted by other mods or leaders. Karma scores alone don’t cut it. The mods can use tools to be more efficient and not require an army. If you are running a site that relies on mods, you need a good relationship with your mods and to give them good tools and guidelines.
Hacker News gets it most wrong with the flagging system, which is ripe for abuse (and is constantly abused). I sent this to dang some time ago, and was met with a muted response:<p>I realized something interesting the other day - users gain access to downvote comments once they reach 501 karma. I've always inferred that this is to make it so that users cannot downvote until they have demonstrated some understanding of the community and an ability to fit in with its norms.<p>However... users get access to flagging at 31 karma. A flag is basically
a super downvote in several respects:<p>- It works on comments and posts<p>- It works on direct replies to your own comments<p>- Just a few flags is enough to remove content entirely<p>- A flagged post cannot be vouched for until after it's been removed, whereas a post can be upvoted before it's downvoted.<p>It's a bit weird to me that the OP version of downvotes is available to users 16 times sooner than the neutered version. I feel like HN has a problem with flagging being abused for censorship - this might provide an explanation.
Speaking of shortcomings, how do people remove their account including all of their comments?<p>(Not saying that I want to remove my account of course, but perhaps one day I might)<p>EDIT: some discussion on this topic appeared here: <a href="https://news.ycombinator.com/item?id=16898422" rel="nofollow">https://news.ycombinator.com/item?id=16898422</a>
Personally, I think the "karma" and up/down voting of comments is a cool feature but they need to be taken with a grain of salt.<p>It's fun to see a submission get upvoted and get karma points and disappointing to see one get buried, and frustrating to see a post get buried and than another covering the same topic go big, but taken in perspective to what's truly important the points are trivial.<p>The main reason I visit HN most days, and generally several times a day, is to find something interesting and most days I do, and some of what I find and learn here is important.<p>I never come here looking for stuff to down vote. That's kind of hard to even imagine because you'd almost have to pass over the interesting and important to do that.<p>Be interesting to be able to see how many down votes a user has made. That could make one a bit less trigger happy.
Stack Overflow is certainly one to talk about this.<p>Last time I tried (a few years ago), voting in either direction required a minimum number of points there also.<p>These days, I accept the votes on S.O. for what they are, and hit the "End" key to read the least-upvoted answer first.
Like most comments here - I am a bit divided about the relevance and even the rationale of this article. That said it has given me a different twist on a sentiment I have been dealing with for a long time:
I'm French but have emigrated to the English-speaking world 15+years ago (NZ, AU, USA). I have found the whole experience very challenging in terms of career management. Assessing my own progresses is something I really struggle with. For years, I was sub-conscientiously blaming my language skills (I have the worst French accent ever - when I meet new people they often ask me "when did you land" as I sound like a Tourist in a city I've lived for years lol).<p>I have since acquired more confidence language-wise, but it did not solve my 'confusion' (a la Travolta 404). I realized that what I was lacking was the negative feedback loop.
I haven't had a long career in France before emigrating, but I had been through a large variety of student/casual jobs and a few short term engineering and project/program management roles. Enough to get accustomed to a French "professional context" in which I recall having professional reviews where negative feedback was given formally, and open discussions where employees would criticize the proposal made by a manager, or when a supplier was late in delivering - loud words would be exchanged on the phone.
I haven't seen any of these in the past 15 years. Political correctness rules. Last time my temper took over and I inadvertently let a bit of my irritation be heard on the phone with a late supplier - they literally clammed down and it just added to the late schedule another 10 days of silence from them until they delivered.
To quote Jeff Atwood - this is exactly the feeling I have at work sometime:
"If a post has zero upvotes, does that mean it’s bad? incorrect? uninteresting? mediocre? There’s no way to tell, because zero has multiple meanings.
[...]
Downvotes give you the critically important ability to distinguish between the good, the bad, and the ugly. Without downvotes, how can you possibly tell the difference between a post that is harmless but uninteresting, and one that is actually wrong or harmful?"<p>Am I hallucinating ? I'd be curious to here about other emigrants/multicultural peeps' experience here.
Kind of ironic because HN still remains one of the most interesting and engaging communities whereas StackOverflow has become pretty insignificant, outdated, stale and very hostile and unfriendly to new posters.
> The advantage of this system is that nobody gets downvoted, but at a steep cost: we’ve lost half the potential information.<p>A Mr. Claude Shannon would like a word. I joke, but when you have controls on voting rings and co-ordination, downvotes are necessarily noisier. An upvote is effectively neutral engagement with the content, where a downvote actively reduces engagement. From the perspective of wanting to mine the constructive engagement out of the collective minds of the audience, downvotes are about as meaningful and information dense as a fire alarm.
I thought about making an aggregator when you can only downvote, so that only "surviving" posts would appear.<p>In general most users don't upvote because they're not sure they agree with a post or article or really want it to have more visibility.<p>But I'm betting that most users would prefer to express their disagreement or negative sentiment if it was the only option, and say nothing about things they agree with.<p><a href="https://en.wikipedia.org/wiki/Eliminationism" rel="nofollow">https://en.wikipedia.org/wiki/Eliminationism</a>
I assume that Jeff has changed his mind on this because Discourse does not have down-votes (with the caveat that Discourse has a different audience from HN). I wonder what made the difference?
I would like the ability to see my ratio of upvotes to downvotes. I think it would help show me if I was applying the same zeal to promote ideas as I use to demote them.
Hiding posts is incredibly useful. It's a good way to keep the blood pressure in an acceptable range.<p>I keep thinking that a script or an extension to auto-hide stories from blacklisted domains on the front page would greatly improve my experience.<p>The beauty of hiding, though, is that it only impacts my experience. If people want to discuss and vote up things I find uninteresting, they can have at it, and I don't have to know anything about it.
It would be interesting to have a downvote system where the downvote is added as a comment, giving the downvoter an opportunity to say why the post was downvoted. And also give anybody else an opportunity to downvote the downvote itself (with comment). In other words, downvoting a post might impact you if your downvote is downvoted.
I think the content at HN is pretty good. Some other novel approaches I have seen includes upquest.com which has voting based on where you think the score is headed. I think (as others have mentioned) downvoting introduces issues like for example downvoting articles around the one article you want upvoted.
I just wish you could remove the number next to your name natively in HN. It can't seriously take more than 5 minutes to implement this. There's no excuse IMHO. I want to interact with the site without seeing the number change and wondering why it did.
IMO that seems like a comment that hasn't aged too well. HN remains effective more than 10 years after that, and although he may have studied HN while building Stack Overflow, I don't think it's safe to say that SO has done as well.
I sometimes wonder if the simplicity of the upvote/downvote exposes it to the limbic decisioning, where some portion of the signal is pure lizard brain reaction vs a thoughtful act to help moderate the community.
The other big issue is the ambiguity of the vote: dis/approval, dis/agreement, or low-/high-signal?<p>I think agreement and interesting (signal) need separate votes. It could also be nice to have a hilarity vote that can be de-prioritized, allowing optional entertainment value without shunning it completely.<p>Also, flagging should be reserved for absolutely deplorable comments, not microaggressions or score-settling.
"It’s pretty clear now that the broken windows theory applies to community sites as well. The theory is that minor forms of bad behavior encourage worse ones: that a neighborhood with lots of graffiti and broken windows becomes one where robberies occur. I was living in New York when Giuliani introduced the reforms that made the broken windows theory famous, and the transformation was miraculous. And I was a Reddit user when the opposite happened there, and the transformation was equally dramatic."<p>Paul Graham is such a tool.<p><a href="https://news.ycombinator.com/item?id=24099192" rel="nofollow">https://news.ycombinator.com/item?id=24099192</a><p>Somebody should delete this dumpster fire of a website, perhaps a real "hacker" might take the initiative some day!
To those of you addicted to Hacker News; I can offer this tip.<p>Get a Hellban. It's not hard after all. I'll probally get hellbanned over this post?<p>You will find you will still pruse HN, but very rarely will you feel the need to contribute. Your time here will decrease.<p>Hacker News does not need much improvement. It does seem like it's on the back end of it's business cycle though, like all current social sites? I used business cycle because I don't know the word for a social site's decline. Decline is too harsh a word. Maybe predictable conversations, and a bit of a echo chamber, is more appripro? Hacker News is still the best site for scientific conversation on the the internet though.<p>It seems like we have hashed out the important scientific stuff, and most issues computer related.<p>I'm so greatfull to the hard nosed scientists whom hang here, and take umbridge to unsupported claims. I really do appreciate you guys. You know who you are. You are usually the first one to comment on a study's flaw, or whether the Placebo Effect is rearing it's magical head.<p>Graham should add a few more moderators though. Never give too much power to one person, and an owner. This Plus--if Graham pays more Moderators, he could write off the added expence? It could be his gift to humanity? Oh yea, block a person, but do away with hellbanning. It's wimpy, passive aggressive, weak move. I understand moderation, but just block a person. Don't let them rattle on to open air.<p>I would like to see a few more political discussions though. Not a lot, but some. Maybe one every other week? Expose the political fakers, and hypocrites. Expose the political hacks on both sides. Track every dime both parties bring in by way of Lobbiests?
Belief in Downvoting seems to be an artifact of idealistic naivete (that I used to have as well), in the same way as enthusiastic, unquestioning belief in the writings of Ayn Rand, Karl Marx etc (not calling anyone out, everyone was a teenager once). Downvotes SEEM good on paper. In the real world, they're garbage because they're never used as intended.<p>Nearly every implementation of downvotes assumes they will be used as a way for ordinary users to moderate bad content, <i>and that this power will not be abused</i>. The reality is that we've known mob rule was a terrible idea since at least the Enlightenment era, and probably much earlier. But apparently we need to keep re-discovering it.<p>There is a certain kind of person that mostly likes to attack and destroy, in any given situation. Not for any good reason, but simply because they like conflict. Give them a downvote, and they'll latch on to it and destroy communities. Give them only an upvote and they'll lose interest and go cause problems somewhere else.<p>Moderation is for moderators. Users cannot, and should not, be trusted with it.
>It’s pretty clear now that the broken windows theory applies to community sites as well. The theory is that minor forms of bad behavior encourage worse ones: that a neighborhood with lots of graffiti and broken windows becomes one where robberies occur. I was living in New York when Giuliani introduced the reforms that made the broken windows theory famous, and the transformation was miraculous.<p>The success of "broken windows" policing is not supported by much evidence, so the rest of article based on this concept is suspect: <a href="https://en.wikipedia.org/wiki/Broken_windows_theory#Criticism" rel="nofollow">https://en.wikipedia.org/wiki/Broken_windows_theory#Criticis...</a>
I used to read Coding Horror but stopped paying attention to Jeff Atwood after his anti-coding rant. He repeatedly insisted that "learning to code" is a really dumb thing to do unless you intend to make a career out of it. A really shocking opinion that I was surprised to hear from him, and made me lose all respect for his coding-related opinions.