Hacker News new | past | comments | ask | show | jobs | submit login
Dear YC: Please stop downvoting thoughtful comments you disagree with.
174 points by andreyf on July 24, 2009 | hide | past | favorite | 111 comments
When someone has obviously put effort into contributing to the conversation, it's rude to downvote them to negative scores. Most recently, there some major offenders here:

http://news.ycombinator.com/item?id=720779

and here:

http://news.ycombinator.com/item?id=720215

and here:

http://news.ycombinator.com/item?id=720569




As someone who has received a string of downvoted commented recently (admittedly they were warranted given the nature of some of them) I really don't see the problem.

Someone disagrees with you and does the initial downvoting, then groupthink kicks in and your score hits the floor.

It's a simple fact of life in social media sites (it occurs on many sites with differing communities), this is why PG put in the -8 limit, to combat the detrimental effect of this kind of groupthink.

Honestly, deal with it, it's just online Karma, a simple numerical value... what effect does it really have after you've unlocked various features on HN.


I kind of see myself as the balancer. I tend to actually read a downvoted comment to see if it was worthy (and vote it up) or see a highly upvoted comment and possibly vote it down if it's just getting upvoted because it puts forward a popular view (e.g. Microsoft is evil) rather than something of value.


If downvotes are used for disagreement, then I have to read them in order to know if they are relevant. If downvotes are saved only for irrelevant, one-liners, me-too's, etc., then I can safely ignore them.


I worry that when so many comments have negative votes, readers will be more inclined to downvote future comments.

While I agree with you that it's really human nature, I still think it needs to be discouraged, if only to keep HN the kind of place where debate is encouraged.


"While I agree with you that it's really human nature, I still think it needs to be discouraged, if only to keep HN the kind of place where debate is encouraged."

I agree with you, absolutely. I think we're mostly adults here and hearty debate should be encouraged.

I think that if you really want to counteract the douchebag effect (the groupthink downvote pileon on thoughtful comments) then make who has downvoted you public to everyone but the downvotee - transparency would be one way to help combat it.

Think about it for a second, if you have a legitimate gripe with someone's comment, then whether you downvote them or not shouldn't matter. Those comments are easy to single out.

However, if you're piling on just because of groupthink, then why not let other's know. I'm sure someone could come up with an algorithm for working out who's being a dick with the downvoting consistently.

I think there should be a clear distinction between legitimately disagreeing/downvoting someone's comment and being a dick, so why not try and expose the latter.

We've seen in many, many cases that anonymity on the web has given rise to the kinds of behaviours that many of us dislike. Why not bring it back a little towards people being responsible for their actions.


Why do you think that the 8th downvote isn't as carefully considered as the first? Maybe the comment really is that worthy of being downvoted.


Groupthink is a silly thing. Besides, if your comment gets downvoted, it becomes gray and much harder to read.

It's unfortunate if thoughtful comments get "hidden" by downvotes.


I usually check out grayed out comments on the posts I read; I have yet to see one that was "thoughtful". Maybe on posts I don't bother to read, because the topic doesn't particularly interest me, there are lots of thoughtful comments downvoted, but I haven't seen it happening.


I often check them too, just to see if they were actually offensive, or just victims of groupthink etc.

Well, if by thoughtful comments you mean obviously intelligent, well thought-out and reasonable comments, then yes, you don't see (m)any of those downvoted.

But sometimes people downvote comments where the author has clearly put some thought into them, and tried to contribute to the conversation. Those shouldn't be grayed out just for failing to be as spectacular as some of the others. Everyone does what they can.

I do realize I'm trying to defend my own comments here too..


Does this not apply anymore?

http://news.ycombinator.com/item?id=392347


I suspect most of us have seen that. I am mostly in agreement with the way pg runs this place, but in this one case, I think he's wrong. I must not be the only one, since this topic comes up every so often.


Yeah, giving him the benefit of the doubt, I imagine PG meant that in a different context than I'm putting it here - I'm saying that reflexive downvoting of thought-out challenges to your opinions is impolite, and asking people to please think twice before doing it.


Though buried in the pile, I think more people need to notice this link.


The biggest problem I've seen grow recently are healthy comments quickly downvoted from 1 to 0. I immediately vote them back up but I want to reach through the monitor and throttle the voter for spinning such a sensitive-to-initial-conditions process in a negative direction.

If I were the HN Sovereign, I'd consider having moderators casually review all 'first votes' on comments, especially 'first downvotes'. If a certain user is injecting a lot of 'sting' into threads with a hair-trigger down-vote, I'd put their franchise on double-secret-probation.

What is double-secret-probation, you might ask? Well, they'd still see their votes take effect, but no one else would -- at least for initial votes, maybe for others too.

Their negativity would be quarantined.


It would be nice if you linked to the actual comments that were down voted instead of to article pages with dozens of comments, where, presumably, the reader has to guess which comment you were referring to.


I thought there were too many comments to link to:

Create's and bitdiddle's responses here: http://news.ycombinator.com/item?id=720924

This response: http://news.ycombinator.com/item?id=720909

Is identical to this one: http://news.ycombinator.com/item?id=720766

My comments in this entire conversation, also: http://news.ycombinator.com/item?id=720569 (at one point, each of mine was <0, now upvoted out of nagativity, thanks! :))


Wrt create and bitdiddle, downvote seems like the appropriate response. It appears that create is factually wrong, and bitdiddle was saying something irrelevant about stock price.

Wrt your comments on the 1984-Amazon-copyrights thing, I think you have something there. HN has too many people who think copyright laws are bad. Many here even seem to think that the GPL is bad (see comments on Zed's GPL rant). And many of them feel this very strongly. That's just the community; when you come in here and say that following copyright law is a good thing, to many people it doesn't sound like a reasonable thing to say, it seems like a harmful thing to say. They turn into white blood cells and stamp out the infection. :)

I don't even bother opening the comment pages on GPL and copyright articles anymore. I don't see any light being generated there; only heat.


Got to say I love the quote "no light generated here, only heat". Did you just make that up? It's a great quote for discussion groups.


I feel it as a pun on a well-known idiom in other languages, that would translate verbatim to more about smoke than fire i.e. something of no substance or use, most often discussion. It is a certainly saying in my native language.


MJF: Hemminger is claiming Microsoft put the LIC code under the GPL because it was in violation of the GPL. Is this true? Did you have to suggest to (Microsoft Platform Strategy Chief Sam) Ramji & Co. that they were in violation in order to get them to agree to release the code under GPLv2?

GKH: I didn’t have to “suggest” anything, I only had to merely point out the obviousness of the situation :)

MJF: If this isn’t accurate, could you let me know how to interpret (Hemminger’s) comments on his blog.

GKH: No, that sounds accurate.

[...]

Ramji didn’t come right out and deny the GPL violations claim, but I guess that’s as much as we’re going to get.

http://blogs.zdnet.com/microsoft/?p=3433


Where the 3rd one is a simple WFM, which is helpful because it establishes that not everyone is having these problems, the 2nd comment sounds condescending and ignores the problems that the parent is having in favour of bringing up an unrelated bug that should be fixed soon.

See the parent post: http://news.ycombinator.com/item?id=720525


Seeing the first link, I find this: "Although it may be enticing to spend a few years with a slick point and click IDE such as VisualStudio and pretend you are programming..."

Is that a thoughtful comment? I don't think so. I think that this kind of comment is what groupthink is really made of.

People thinks that the audience of the site is mostly anti-microsoft and happily insults "the others", hoping that this kind of remark will be rewarded.

Heck! If it turns out that enough people dislike that behaviour and downvote, you can still make a story asking for a revision. It seems to have worked, BTW.

OK, it might not be the case for all the comments you meant, but there are many and the general tone of them really didn't invite to more detailed read.

IOW: if you have something interesting to say, try to write respectfully and to the point. It's posible to say you don't like certain IDE without saying their users "pretend" to be programming.


Fairly easy to scan the comments the OP is talking about -- they're light-grey with negative scores, and it's clear they've had effort put into them.


"Fairly easy to scan the comments the OP is talking about -- they're light-grey with negative scores, "

not necessarily true at the time of reading (vs at the time of posting) . You forget that comment scores are dynamic and even very old comments can be upvoted past zero (downvotes have a time window) and maybe no longer "light-grey with negative scores" by the time someone reads this submission and looks for the comments.


noted, thanks.


IMO All the comments on those threads seem fairly downvoted; either for knee jerk inaccuracy, flogging a point or just generally ranting. EDIT: the FF3.5 one I might agree with - did that get edited in http://news.ycombinator.com/item?id=721314 I didnt see it the first time?

I think it's wrong to call it rude to downvote them......

EDIT: according to the rules comments like "why was that post downvoted" are discouraged. Just FYI.


Somehow this comment seems related to the recent recursion post.

BTW, I think it's fine to complain in a new post. I guess it's annoying in the body of a long string of comments, but the entire point of the posting system is to allow group moderated discussion. If people don't like your question, they'll downvote it.

This is a much better response than ignoring idiotic behavior and/or spamming a good comment thread with whining.


> Somehow this comment seems related to the recent recursion post.

oops not intentional... Im not sure how that got in there (fail copy/paste)

> BTW, I think it's fine to complain in a new post.

Fair enough ofc. I just thought I would raise it in case people had missed the guidelines :)


Cool. Agreed.

But just like the pirates say, it's not a code. More like guidelines.

I think that's what makes HN so frustrating sometimes to people (including myself). There is quite a bit of flex involved. For instance, I just submitted a blog entry I have on arguing which looks at religious-type arguments. It's probably over the line (perhaps not?) but I'm not completely sure, so I just threw it out there to let the community decide. Posting, to me, has a much looser set of rules/guidelines than commenting.

But that's just my interpretation of the community mores.


I have noticed this effect as well. Many of my comments fluctuate massively; up to +10, down to -8, back to 2, down to -5, ...

I think this basically means that we are a few months away from being Reddit. Pretty soon, any comment containing $HILARIOUS_MEME will be upmodded regardless of its applicability or actual humor value. (We are already Reddit with respect to downmods; say something bad about the iPhone or something good about the GPL, and the downmods start almost instantly. Also, I think there used to be articles about programming here, but I haven't seen any in a while. Maybe today I can read about how programmers are smarter than everyone else, or how to sell a $35,000 watch. Tomorrow, I look forward to some cat pictures.)

Anyway, moderation is a waste of time. Hitting "flag" will get comments and articles that you disagree with censored a lot quicker.


Unfortunately that's something of the lifecycle of online communities. Once they reach a critical mass people that probably shouldn't belong start congregating, sooner or later there seems to be a tipping point where things start racing down to the LCD of humor and insight. People start leaving and form another community.

HackerNews is surprising for the length of time it's managed to exist without turning into that. I have noticed that the quality of submissions has decreased over the past year.


I downvote when people post short comments that are basically "you're wrong" without any helpful criticism. When someone criticizes me in a thoughtful way I respond instead of downvoting (and often vote up if it brings up an interesting point).

A good metric that will keep conversation constructive is this: the purpose of your debate isn't to convince the person you're arguing with, it's for the benefit of the bystanders. you'll find yourself constructing much more professional arguments if you keep this in mind, the temptation to engage in "status plays" diminishes. your goal isn't to prove that you're right. your goal is for there to exist at the end a string of intelligent back and forth that display both sides of an issue in a useful way.


I think the people that downvoted you in the amazon threads were doing so because they thought you were trolling.

You seem to be smart, any rational person knows why jeff bezos was apologizing, so it seemed to me like you were saying something controversial to illicit a responce, and that should be discouraged imo.

(there is obviously a line betweens devils advocate and troll which I dont think you crossed, and so didnt downvote you).


To be frank, a lot of the comments on http://news.ycombinator.com/item?id=720779 that you made were not "thoughtful comments that people disagreed with". They were sophomoric. I read through most of it, and didn't vote one way or another, but my take away from it was: people kept pointing out some very obvious things, namely that even within the legal framework that is setup by our copyright laws, there's still a very wide berth of behavior that ranges from tolerable to asinine. Even taking the legal copyright framework into account, Amazon behaved asininely, and that is what Bezos was apologizing for. You stuck to your argument about legality even when it no longer made logical sense, which is just a hairline away from trolling.


Kudos is not about etiquette it's about what motivates people to upvote or downvote and this differs from person to person.

Personally I value truth, humour and interesting information. You will note that etiquette isn't part of my up/down voting practice. Other people have different rules.

Kudos is a reflection of the community which means that the community is (if we follow your theory) slightly rude. I don't have a problem with this. I do however have a problem with NOT downvoting thoughtful comments I disagree with. I think it's rude to tell people that they should think like you. I would downvote you for this but etiquette isn't part of my kudos practice and I think this discussion is interesting.


The purpose of karma on HN is to promote civil discussion, not to show agreement or disagreement. Hence, downmodding comments that you disagree with, yet are still valid, runs counter to the system's intended use.


Maybe that's not the purpose of karma on HN. Maybe the purpose is to assist its users in making judgments about various issues. If a bunch of smart people read a thoughtful comment and indicate their agreement or disagreement via voting, why isn't that useful to someone when forming their own opinion?

As an analogy, consider the issue of global warming. Very few people can invest the time to decide the issue based on primary research. Instead they rely on secondary or tertiary reports, consensus of the scientists involved and their reputations. This is the comment karma system made large.


I use the moderation system here as an asymmetric control: if I agree/like something, I upvote it. I don't downvote if I disagree/dislike something. I only downvote when I think something is not a part of legitimate discussion.

If I encounter a 0 or -1 post that I think is legit, I will upvote it.


Aye but purpose ain't always what implementation becomes, people game systems. So the utopian ideal is civil discussion but the reality of it is whatever people decide and you can't control the people.


This is true, which is why it's worthwhile discussion what the karma system here is supposed to be. If you deliberately don't use it that way, then I find that sad because I like PG's goal. As he explicitly states, this site is partially an experiment. I hope it's not a negative result.


This is because many people here still vote by agreement, rather than for comments that are interesting, insightful, contradictory, terse, and so on. This happens on a much larger scale with lower-rated comments. I once saw the word "no" with a score of 50~, which made up for the contributor's running streak of -5~. It wasn't that their contribution quality became better, it was that their comment hit that major "upvote when you agree" fault in our system. It reduces our quality. Seeing even a poorly-composed brute force attack on the ideas in the article is what I come to the comments for, because even reckless disagreement is often a gold mine of those little bits of information that you didn't take into consideration, but should have. (I've said all this before, and it should be repeated.)

Vote for what you think others should see, not for what you agree with.


In the same vain, we should also discourage upvoting simply for agreement. There is no reason why one-liners should get upvoted to very high scores. Upvotes should reward content and thoughtful discussion.


Ultimately, it's a culture thing. I've been here for the better part of a year, and I've noticed that comments which I feel are worth lots of merit (I wrote a script that makes your life noticeably better! here!) and ones of less merit (It's the same as a link 30 pixels away!) are voted to almost the same karma. This was not always the case - historically, things seemed to be voted far more consistently.

The reason for this, as I ponder it in my head, might be that the threshold for downvoting on comments is too low relative to the amount of people granted that power. In my sole opinion, as the number of people able to downvote goes up, the worth of the comment voting system goes down.

Perhaps, instead - a small economy based upon earning karma with good comments, and spending it when you make negative votes - might be a good idea.


I've had a few comments downmodded, and it sent a clear message: my comment was simply not valuable to the people who read it.

I use downvotes in the same way. There's many factors that contribute to a comment being valuable: factualness is just one of them. Poor commenters defend positions without evidence; try to cast personal feelings as rational arguments; fail to properly read and understand the article; attack an imaginary slight from the parent poster.

In a real conversation, I'm not going to rebut everything disagreeable that you say: sometimes I'll just frown, turn away, shrug, or interupt you. Downmodding is my way of quickly saying that I didn't get anything out of your comment.


N.B. The Hacker News Guidelines:

Resist complaining about being downmodded. It never does any good, and it makes boring reading.

http://ycombinator.com/newsguidelines.html


It's a common conceit among engineers to believe that they can be objective in all things and cleanly separate their opinions and their knowledge. We may be a little better at it than the average person, but we are all still human, even the most socially detached savant is still subject to the influence of passion and bias. Realizing this, the original question is a fantasy, it is utterly impossible because there can be no agreement on what is even considered "thoughtful".

My recommendation is to take it easy, let people be people, and stop daydreaming.


Quick: can someone create a proxy to hn which does the regexp s/\d+ points? by//g

I wonder what it would "feel" like to read hn without the quantification of quality.

After all the most interesting stuff would be still on top.

Edit:

I've done a quick bookmarklet, it seems to break links, but it gives an impression (so enter in in the urlbar at any hn site and press enter):

javascript:document.documentElement.innerHTML = document.documentElement.innerHTML.replace(/\d+ points?| by |/g,"");


I was toying with suggesting this "improvement" the other day - i.e. remove comment scores.

I think it makes things better; but I suspect there would be a lot of opposition.


Maybe one could make that a user setting so that everybody who likes it that way could switch points off. I think I could be more objective when I wouldn't see the exact points but just the relative order of stories and comments.

Another interesting idea would be that a reply to a comment would count as +1 (or maybe +0.5) for the parent. Because I think that if a comment is interesting or controversial enough to justify a reply it deserves a point or a half for that fact alone.


There are too many replies that correct factual errors or explain/reinforce HN social norms to assume reply == upvote. Though you could downvote and reply.


As mentioned in many other comments - there is no good way for a commenting/voting system to determine a user's motivations for voting (in either direction).

...But there are subtle ways to provide incentives for user voting behavior.

For instance, why not attach greater risk/reward to voting?

Incomplete ideas that come to mind: 1)scarcity - finite votes force greater selectivity in voting 2)vote transparency - anonymity permits nastier activity 3)user transparency - highlight ornery vs constructive users

As for doing away with downvotes: Zero is an artificial floor that would simply bundle mediocre comments with comments that are off-topic/spam/in poor taste etc.


A friend of mine is doing his phd research on online discussion to solve this exact problem. Digg, HN & other up/down voting sites struggle to express the viewpoint of the minority. He is working on algorithms that basically let you vote on one side of an issue.

Take for instance a political debate, where Digg and many other community sites lean to the left. Once you have expressed yourself as a "leftist" on a thread, it removes your downvotes from "rightist" comments. This gives the minority a voice, resulting in pluralism as opposed to "winner takes all".

An interesting approach to a common problem.


I wouldn't be happy if, say, objections to an unpopular scientific theory were simply expressed by scientists down-voting and burying it. Unless the idea's completely insane or facetious, I'd want them to respond to it with a measured discussion for many reasons, one of them being that ground-breaking ideas don't always meet with widespread acceptance as soon as they're introduced. I know from personal experience that I've slowly changed my mind after hearing someone argue against my opinion, no matter how much I wanted them to shut up and accept my views!


As much as I agree that voting should be used to indicate value to the discussion and not agreement with one's own opinion, I think it is a hopeless cause. It requires that participants curtail their instincts for the sake of the greater good, an unlikely prospect in the best of circumstances, much less in a community of semi-anonymous strangers.

I think it better to simply suggest that people not take scores so damn seriously. To stop interpreting them as a measure of thoughtfulness or truthfulness and start thinking of them as just a reflection of people's attention.


Honestly?

I downvote comments I agree with, if they look upvoted out of proportion to their value. I also upvote things I disagree with if they've been downvoted unreasonably low. I don't think a stance on an issue that's plausible and civilly stated needs to go below 1, for instance.

Different people have different approaches; if we could get everyone to vote based on one standard, we wouldn't need a karma system in the first place.


I don't downvote comments I disagree with. I do downvote comments that are off-topic, don't add anything constructive, or predictably rehash the same old arguments that have been used in certain online debates thousands of times over the years.

Which on the surface might look like downvoting comments I disagree with... but there's a difference.

Is it just me, or do the weekends tend to attract more meta threads?


Sometimes a reply to a down-voted comment is more insightful than the original. The only comments I down-vote and don't reply to are the "hey this is great" or the one-liner joke comments.

If I think the commenter is wrong or a douche, I'll down-vote and reply with my own side of the debate.


It would be very interesting to see a list of the top 50 or so downmodding offenders next to a list of their "karma score" (or whatever it's called here). Even anonymous statistics would be interesting. Some people are just mean because they can be.


I'll second this with one minor caveat. They're not necessarily "downmodding offenders" until proven as such by the data you're requesting. They're just the top 50 downvoters.


That data's not meaningful without also considering upvotes - they could just be prolific modders.


I too am curious to see the downmod/upmode ratio (anonymously of course). I think mine is probably 10-20 upmods to 1 downmods, and I mod sparsely (maybe 1 comment in 20? depends on the thread). I downmod "noise", not signal I disagree with.

While we're geeking out, I'd be more interested in some kind of karma/comment metric than absolute karma.


Maybe. But some people tend to see the good in things, and some tend to focus on the bad. I think it could be reasonably hypothesized that persons who downvote a lot focus more on the negatives.

It's hard to read contextualization or to infer what we refer to as vocal inflections through writing; I am as guilty as anyone of the occasional "posting while drinking" (which I've noticed does tend to make my writing seem a little bit more hostile than intended); however, some attempts at humor here just end up falling pretty flat.


Down-modding is not what brings the interesting conversations out of the woodwork; up-modding does. So they are "downmodding offenders", especially if the comment was previously up-modded from its original default score.


It would likewise be interesting to hear what are your thoughts about anonymity and freedom in general.

Should we publicize a list of people who voted for the Canadian Lobster Party, because they obviously did it out of spite and should be punished for their lack of patriotism? Or do the benefits of anonymous voting (or modding, in this instance) outweigh the costs of people who don't take the rules seriously?

There's a little society going on here at HN. If you feel compelled to punish people who stray from the "true path" here, it's probably just a reflection of your greater worldview.

And in a final attempt to be downmodded, let me quote scripture: "He who is true in a little, is true in much; he who is false in small things, is false in great." Luke 16:10

Edit: Typo


It would likewise be interesting to hear what are your thoughts about anonymity and freedom in general.

Should we publicize a list of people who voted for the Canadian Lobster Party

Hey, if the numbers were interesting, why not? I'm INTP; I love looking for patterns in things, especially numbers. Social media is almost like poker in that it has a psychology issue.

But that said, I don't really downmod people very often (inadvertently sometimes, since the up and downmod arrows are so close together). My upmod/downmod ratio is probably something like 79 to 1. I'm curious about others'.

Yeah, sometimes certain people irk me, but usually do I respond to them before downmodding. And I usually attempt to befriend them after an apology, or at least an attempt for a truce on a subject we might disagree about. But some people are just vindictive, or extreme cynics.

As a result of this, I have a lot of one or two-point comments. This doesn't bother me; I'm not here to karma whore. I like the articles and the discussions, and tend to lurk most of the time anyway, posting only in spurts.

P.S. Why did you say that you're "attempting to be downmodded"?


Fair enough. I agree the numbers could be interesting. But it's quite different if you want to see users' names.

I grow weary of HN's Singaporean (overly sanitized) approach to geek discussion. Geeks suffer enough from intellectual narcissism without trying to constitutionalize it. Thankfully there's enough brain candy on the site to compensate.

And people, let's go light on the hacker this and hacker that. We all get it. We're all keen to learn and explore. But... was Michael Jackson a hacker? Geebus. Labels are overrated.

God Bless the USA, but there's something distinctly American about coming up with a subtly differentiated product and then hyping it as the total opposite of what came just before it. Yeah, I mean HN / Reddit / Slashdot.

P.S.: I made the remark about being downmodded because I assumed my gentle admonishment would be unpopular, plus I noticed that almost without fail it happens when people quote the Bible.


Fair enough. I agree the numbers could be interesting. But it's quite different if you want to see users' names.

Well, a name is just a variable. If that variable can be correlated to a group status of another value within a social network, it can reveal patterns about social dynamics and even things about viral distribution of certain prominent ideas / memes / buzz-worthy topics.

It is interesting how, for example, the Google zeitgeist works, or the "trending topics" on Twitter. There are big pipes and there are small pipes, but in the oligopoly of mainstream media, it is only a select few who control the flow of data.

P.S. By "mainstream media," I mean the conglomerates who control the major networks of NBC, ABC, CBS, FOX, and the other free airwaves that don't require a cable subscription payment every month.


I wish I had the time to read enough comments to have an intelligent opinion on this.... I generally vote things up that I strongly agree with and vote things down that I strongly disagree with, without putting much more thought into it than that.


Its just a number, this is just a board where we discuss what others have written (and then people find time to VOTE on the discussions about what someone else has written!). stop taking this number so seriously people!!


I have little hope that Hacker News will not suffer the same fate as reddit and digg. Hopefully there will be a new refuge by then (on the other hand, it could save me quite some time).


It's a number that really means nothing in the whole scope life. Why are you so concerned with it? This is a legitimate question. No one has ever answered it.


I think its actually more valuable if people do vote their agreement (and disagreement), think of it a s very techie peer review process.


Meta is Murder

http://www.codinghorror.com/blog/archives/001282.html


I am impressed by the number of votes (up and down) on this topic. A nerve was hit, obviously.


i don't see what's wrong with downvoting on disagreement. does it hurt people's feelings?


Yes, a downvote can 'sting' -- because even if the person casting it means to say "I disagree", they are sending the exact same signal as is sent when they mean "Your behavior is unacceptable". And, downvotes are a 'scarlet letter' of community disapproval, and cause comments to fade in color and sink in position. It is a status kick-in-the-groin.

And that's OK, for truly unwanted behavior.

But the right kind of disagreement is good, healthy, necessary and should be celebrated.

One option would be to have two axes of single-click vote-reaction. One is classic up-down promote/demote, for whether a comment has quality. The other would be left-right agree/disagree, a mini poll to capture sentiments.

Unless there is an easy outlet -- a click-region -- for sending a 'disagree' signal separate from the 'disapprove' downvote, it is inevitable the two will be conflated, causing angst and periodic threads like this one.

For previous discussion on this 'two-axis' concept, see:

http://news.ycombinator.com/item?id=613112

http://news.ycombinator.com/item?id=214521


Thoughtful comments that are factually wrong should be downvoted.

Seriously. If your code doesn't compile, the compiler doesn't care how much effort it took you to write it...


"Thoughtful comments that are factually wrong should be downvoted."

Maybe. Maybe not. But one thing's for sure. It you think something is "factually wrong", your reply provides much more feedback than a vote.

I'd like to think that judgment based upon data is what separates hackers from posers. Let's see that data please.


Let me pose a hypothetical: I'm reading HN while I'm waiting for a particularly long compile or backup or whatever. I've got about 5 minutes and about 4 minutes in, I find a comment that honestly and eloquently makes a claim based on the perceived fact that the world is flat. This comment is highly rated. Should I:

1) Withhold my down-vote because I don't have time to explain it?

2) Withhold my down-vote unless someone else has taken the time to explain it?

3) Down-vote and hope someone else explains it or at least hope that this helps other comments percolate to the top?

I honestly think 3 is the best option and I can see the point in 2, though I think it's withholding signal from the system.

Related question: is it relevant what the comment's score was? If it was -7, should I make it -8? If it was +200, does it decrease the burden to comment?


This hypothetical "factually wrong" dilemma is a red herring. The real issue is disagreement, not factual correction. Much more often the case is that someone has written a thoughtful, honest comment based on some worldview that is very different from your own. In that scenario you should not downvote at all. Rather, take the time to explain your own worldview or just move on.


The "factually wrong dilemma" is in no way a red herring. It's actually precisely the point I was responding to in the grandparent post. Down-voting based on opinion is something else entirely. Of course, people will confuse opinion and fact sometimes, and I would encourage people to be skeptical of themselves on delineating the two, but that doesn't invalidate the idea that people can say things that are factually wrong.

Yes, if someone says something that is factually correct (or at least not incorrect) based on a different world-view, then I agree, don't vote them down. If they say something that is factually incorrect but accepted in their worldview, I don't know that that changes what I was saying before.

Edit: you know, it's usually my policy not to comment on how comments have been modded, but I do find it hilarious that in this discussion, someone's down-voted this comment (at the time of this edit) without replying.


To the sibling comment: yes, I agree on the main point and am just discussing a finer point, an issue that was raised by an earlier comment on this chain.

Voting based on opinion reinforces groupthink. Voting based on fact helps keep the level of discussion high.


It's a red herring because the problem that the OP is trying to get at is not about factually incorrect comments. The important message is that you shouldn't downvote comments you disagree with, and instead we're nitpicking about how to be the fact police.

Maybe it's because we already agree on the main point, and are just trying to settle the finer points?


What I would do personally:

  if score < 1
    option 1
  else
    option 3
But of course it's a judgment call. If it seems like the person is intentionally trolling, then by all means, go negative. If the person is just misinformed, zero is a sufficient slap on the wrist that hopefully says "make sure you know what you are talking about next time", and a reply makes it obvious why it got downvoted.


For me, it's:

  if score < 2
    option 1
  else
    option 3
Note that I don't want to penalize them for being wrong-- we've all been there (and sometimes when we think they're wrong, really we are). It's better to let them lie and reply with better information or a ("Are you sure that's right? I recall blahblah.")


I'm not totally sure, but I would say either upvote your favorite amongst the rebutting comments that will likely crop up if the incorrect post is highly rated, or stay quiet.

Here's a different hypothetical where upvoting based on agreeing opinions, rather than quality of reasoning, can be a problem. It sometimes crops up in sites with rating systems like the one we are reading right now, and it could be on any blatantly wrong, or merely not-widely-held opinion.

1) A person writes "a comment that honestly and eloquently makes a claim based on the perceived fact that the world is flat" (or any other well-reasoned claim based on an incorrect premise, or simply an opinion which is disagreed with).

2) Either because it was well reasoned or because of a number of otherwise quiet flat earth believers agreeing with it, it gets up-voted highly.

3) A few passionate replies are written to rebut the highly voted claim, but the majority of which do so caustically, without good reasoning; perhaps even name-calling.

4) In order to push forth their opinion, people who believe the earth is round up-vote the rebuttals, regardless of their quality; and the tragedy is that sometimes (often?) the rebuttal most up-voted is one that simply appeared earliest, rather than the one which rebuts most clearly.

So, I don't know what the ideal voting method would be, but voting based on quality of reasoning, not agreement, would have avoided this problem -- at least in the case of arguments.

A side note: as others have noted, the distinction between "fact" and "opinion" is, well, a matter of opinion in many cases.


A comment like that requires a reply. If you don't have the time, that's okay b/c someone probably will (but you could check back later to make sure someone has).

Don't punish (downvote) someone for earnestly taking part in the conversation. Other people probably have similar beliefs so this person did everyone a service by bringing the question out into the open so it can be discussed.

Votes encourage discussion, not correctness.


"...eloquently makes a claim based on the perceived fact that the world is flat."

You should do what feels right to you.

What would I do? Probably click reply and challenge my fellow hackers with something like, "Who you gonna listen to, Galileo or Thomas Friedman?"


I agree that one should do what feels right.

FWIW, I was trying to posit a situation in which you don't have time to respond. I was using flat earth as a stand-in for a wrong statement, and oftentimes, it takes more than name-dropping Galileo to demonstrate that the writer is factually incorrect. I agree that if you have time, demonstrating the truth is the best option.


To continue with your analogy, I would assume most people would be very irritated with a compiler which only reported that you made an error but didn't give you any context into where the error occurred.

Most compilers make a large effort to give you good error messages with file names and line numbers so you can correct your mistake.

Downvoting is like reporting an error with no context: not very useful in determining what is wrong with the comment.


Too many people in these discussions treat the issue as a false dichotomy. They fail to take two things into account:

1) Downvoting and responding aren't mutually exclusive.

2) Downvoting and the presence of comment that already states the exact reason for the downvote aren't mutually exclusive. You downvote one post and upvote the comment that explains the downvote.

You really wouldn't want everyone that downvotes a post to repeat the same reason over and over again.


Disagreed. If downvoting meant "you are wrong", then it would be kinda like a black-hole if it weren't accompained by a reply, because the obvious follow-up question "why is that wrong?" would go un-answered, and thus nobody learns. Since learning is what we are here for, please correct "factually wrong" posts with a reply, not with a downvote.

It seems more correct to downvote those posts or replies that do not follow the vein of intellectual conversation that HN tries to promote, those comments that distract you from the main topic instead of giving you more info to chew on.

Similarly, those that do make you think, whether you agree or not, have contributed to the conversation and should not be downvoted. If you only wanted correct comments, then this wouldn't be a conversation, it would be going over what you already know, saying out loud what you read in a book or just plain brain masturbation.

Let's have a conversation.


It is not an either/or - you can both vote and reply.

Thoughtfulness has little correlation to the quality of a comment. A one-liner from a subject matter expert is worth more than an entire paragraph of mere hearsay.


If you replied with a correction, it's implicit that you disagree, and then downvotes would be redundant.

I think one liners, unless they are reminders of something that's universally obvious, are worse than no comment at all, because they can't contain much information.


'If downvoting meant "you are wrong", then it would be kinda like a black-hole if it weren't accompained by a reply, because the obvious follow-up question "why is that wrong?" would go un-answered, and thus nobody learns.'

On Reddit, since they added in the "suppress entire thread once the parent goes below -4", yes. On HN, which never suppresses the thread, no. I for one occasionally check even the -8 posts, though usually if you make it to -8 it's for a reason. I've "rescued" a few -3s with replies before.


On reddit when things go to -8 it's for one reason only: They are pro religion, or pro israel, or conservative. (OK, that's 3, but it's basically the same thing.)

On reddit stupid stuff usually gets a plus vote, and people express agree/disagree via votes.

This is NOT the way HN is run, and people coming here from reddit have a hard time adjusting. So we need threads like this approximately monthly.

Maybe PG can require all new members to read some kind of etiquette guide (with a quiz).


Having “come here from Reddit” some time ago, I can vouch that for newly arrived redditors this is moot, since I’m still not offered a downvote button.

I’d guess that by the time the downvote button shows up, most people should have had time to realize HN generally offers more interesting discussion because it works the way Reddiquette FAQ says Reddit should work: do not downvote points of view.

// Now I’ll go back to lurking.


I disagree. Just adding a comment to explain exactly what is factually wrong is a much better way to continue the conversation. Downvotes should be reserved for vulgar, rude, or off-topic comments.


Does something which is factually wrong improve the quality of the conversation? I think that downvotes should be used to push down comments that reduce the quality of the conversation even if they are not vulgar, rude, or off-topic.


Yes. People who contribute things they believe to be correct, even if incorrect, gives other people an opportunity to demonstrate the correct behavior and others still a chance to be educated.

Additionally, sometimes the original poster was in fact correct, and the responder correcting him is proven incorrect.


In that case, the comment does add value to the conversation, whether it is correct or not, so I agree with you that in that case downvoting is wrong.


Comments that are factually wrong should go to 0 and no lower!

Below 0 is only for spam and things like "Me too".


I disagree -- someone else could come along and upvote it again.

Since HN reports the final number, not +/-, information is lost.


Yes. I'd really like to have: "$upvotes - $downvotes = $score". I wonder if the past upvotes and downvotes count for each comment is stored in the database or just the final score figures?

I posted this on the suggestions page long ago: http://news.ycombinator.com/item?id=107060 It has only 20 points, I don't know if it means almost nobody else wants that...


Few comments are plainly "factually wrong." For one, the facts are often a matter of debate. For two, many comments represent judgments, your own being a perfect example: is the statement "thoughtful comments that are factually wrong should be downvoted" itself evaluable as either "factually wrong" or "factually right?" Only if you believe that the "should" in this case exists independently in the world could you answer yes to that. Then it would be a fact to be discovered. Otherwise, (and most people take this position), we can agree on the facts and disagree on the "should."

So it goes with most prescriptive uses of the word "should," actually.


People are not compilers. The difference is that people are able write and explain their point of view. If you think that something is factually wrong, I'd be interested to hear what is right. How do you explain that by downvoting?

Voting is good for filtering large amounts of content. It's useless, even offensive, to make a point that way.


Sure, here's one that's popular, but factually incorrect (+5), with my request for correction at -1 at one point: http://news.ycombinator.com/item?id=720644


I read that example. It's factually incorrect if you take the comment literally, but if you understand the underlying point that the author was trying to make, it is a legitimate addition to the conversation.

This goes to show pretty clearly that someone's facts are someone else's opinion. How much of a nitpicker do you want everyone to be here?


Good point. I'd also note, however, that as hackers, we should attempt to be as precise as possible in our discussions.


I upvote comments that are factually wrong, but where the phenomenon of someone on HN saying it interests me.


> If your code doesn't compile, the compiler doesn't care how much effort it took you to write it

Trying be a compiler for somebody's though ;)




Join us for AI Startup School this June 16-17 in San Francisco!

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

Search: