A) Downvote the comment because the I think the conclusions are grossly wrong.

B) Upvote the comment because I'm so enamored of the elegance of the argument, it's contribution to the conversation, and how it made me think more carefully about my own argument.

C) Ignore it - and unjustly not provide my reasoned critique on what clearly took a great deal of effort by the author.

If I could only Flag it as "DISAGREE WITH CONCLUSION" and "ADDS TO CONVERSATION" then everything would be good.

The four flags that I think would be useful would be:

  FLUFF vs CONTRIBUTES
  AGREE vs DISAGREE
  SPAM
  INCIVIL

An author may then write an unpopular, but carefully reasoned discourse, and still enjoy the positive feedback of his peers on his contribution, even if he is challenging their perspectives.

Edit: On consideration, I should have followed my own advice and started this with: "A well reasoned, well written and interesting argument. But why aren't you adding the alternative of ANSWERING the argument instead of just moderating?!"

There is also the (strong) possibility that I really have nothing to add to the conversation other than a DH3 "Disagree" - which perhaps suggests that I have no right to be moderating that particular comment in the first place.

I sometimes invite people to write such comments if they disagree with me. Often, when a comment makes a factual assertion for which I know of no evidence, I will reply to ask for the evidence. I will say things such as "Citations, please?" or "What are some examples of this?" or in some other way ask for the participant to back up the statement to which I reply.

But there are cases when a comment is totally wrong, and not a good contribution to the thread besides (which has always been a generally supported basis for downvoting), and it's just too much bother to post a countering reply or to ask follow-up questions. In such cases, I don't feel embarrassed at all in downvoting the wrong comment. The site founder pg, who announced news of changes he is considering in flagging options on HN in this thread, wrote 681 days ago,

I think it's ok to use the up and down arrows to express agreement. Obviously the uparrows aren't only for applauding politeness, so it seems reasonable that the downarrows aren't only for booing rudeness.

http://news.ycombinator.com/item?id=117171

Voting and flagging options today are still much as they were then, and so for the moment I don't think it's a bad idea to downvote for disagreement, particularly if a comment has other undesirable characteristics. I will happily adapt to a new system if the site management implements a new system. And I invite participants who disagree with this reply of mine to write a comment and explain why my argument is totally wrong, if that is not too much trouble to you.

That's the inherent issue here. It's that the arrows used to represent quality in theory are actually used in practice to indicate agreement. And then with the down voting, you effectively have a system of negative reinforcement because your comments are being downvoted because you're a dissenter, but interpreted as a sign of bad quality and discouraged to speak further. It's interpreted as a sign of bad quality because all the signs say so. You get grayed out if it's too negative and your comment gets pushed to the bottom. Perhaps you'll become alienated and leave the community, making the HN community slightly more homogenous.

I agree with you in that it's a good idea to allow people to indicate their agreement and disagreement, but to have it clearly defined as an indicator of popularity.

1) "worth reading" / "not worth reading" 2) agree / disagree

Then add a system of whatever flags you think are necessary, like "spam".

I treat the current voting arrows as if they meant (1). I vote up things that I think are worth other people's time to read; vote down things that are a waste of time to read. I very rarely downvote. In some cases where others might more liberally downvote, I simply don't vote.

I think the practice should at least be modeled liked a number-line. The default is 0, neither an upmod or downmod. Just as it is, an upmod (or +1) should be given if the post was particularly interesting (not agreeable, interesting), and a downmod (or -1) should be given if the post was particularly non-interesting (not disagreeable, non-interesting).

I think too many people only see the world as either +1/-1 with no option to just leave a comment alone.

Downvoting something you disagree with kills a discussion, IMHO.

Perhaps a way to balance this is making flagging (up or down) cost karma points, but that would lead to them (KPs) being visible to give immediate feedback on the user's actions. Likewise, upvotes give back the karma you lost modding some comment. A different cost for up/down seems also a good idea. This, simply, could create an incentive to post even when one has nothing to say and prevent constructive modding by people who post little.

Hmm.... Perhaps an agreeing vote would redeem part of the karma paid to up/down vote.

This balancing act is a very hard problem.

I asked last week

http://news.ycombinator.com/item?id=1006946

whether attempts to make the site better, which is what many downvotes and flags currently are, ought not to be rewarded with MORE karma rather than less. I think people who clean up the trash should be paid for their work.

It seems to me, respectfully, that using the downvote to try and clean the community is like a small-town sheriff trying to keep the perceived riff-raff of city-slickers, with their strange ways and funny hair, out of his community.

I think a bigger danger is the suppression of dissent. Rewarding it with more karma would be a huge mistake.

By the way, it's an EMPIRICAL question whether downvotes or posting disagreeing replies does more to discourage expressions of certain opinions. It seems to me that someone thin-skinned enough to cease posting his opinions because his comments get downvoted (rather than reconsidering how he comments and seeing if he can improve how he comments) might be just as discouraged from posting minority opinions if he sees a lot of replies that disagree with him. If a person has courage of his convictions because his opinions are based on evidence and careful thought, there is little reason not to come back here and express the opinions, even if his comments don't get a huge number of net upvotes.

I respectfully disagree. A downmod supplies almost zero feedback to the recipient. Many posts I've seen with negative mod scores are complex and have a number of topics in them. A downvote at no time explains what precisely was wrong with the post.

It's like a presidential candidate not getting a vote, he doesn't know what part of his platform lost him the vote, just that he didn't get it.

In real life a candidate can pay for a poll and find out why, here it seems that asking why immediately results in more downmods.

(I admit the analogy is imperfect, but hopefully you see where I'm trying to go with it).

And meta-comment, really? Your post was downvoted? Way to discourage ideas and effort....sigh

I'm anxious to see this experiment. If flags are not binary, I'm curious how the number of different types of flags will be used to bubble up articles and comments.

I like the idea of separating "worth reading" from "agree". If only "worth reading" makes things bubble up then people that strongly agree might mark things as "worth reading" because they want them to bubble up. An interesting problem.

Personally I never even notice the person's name unless I'm going back and forth, so I'm not even sure how much effect this has.

Besides, it's human nature to always ask "who". Any system that tries to override human nature has its work cut out for it.

Not just nature, but efficiency. The statement, "In argument subject, not author, holds interest" may be true in theory but in practice this breaks down. People without established expertise may have worthwhile contributions but it's easier for people to receive contributions from people with known expertise. For example, I would just as wholeheartedly receive a comment from cperciva as one from "A well-known cryptographer and talented mathematician with a relevant and competitive PhD."

This is mostly notably seen in the physical sciences wherein most theoretical physics treatises are given little attention from the mainstream physics community unless there is a credible and relevant PhD backing it. Physicists simply don't have the time to pick through all of the papers from amateur hour and point out where they messed up the calculation that they should have learned in Quantum Optics 332.

This expedience is partially reflected in HN comments. While ideally everyone would have a chance to respond to every comment to point out where it is factually incorrect, few have the time. It is also disconcerting to see a factually incorrect comment with more points than the factually correct reply. In each of these cases, one could comment to express agreement with the reply and reinforce the incorrectness of the original comment, but a combination of vote up/vote down fulfills a similar function in a fraction of the time. Perfect? No. Expedient? Yes.

I mean I can already see some problems to being able to see a name, and having that "communicate a great deal" just by the name. If the comment communicates a great deal, then why can't the comment stand on its own merits? It just seems that more people will be inclined to upvote because of the person who said it.

And I understand the concern about being able quickly skim and find notable experts replying. This is a potential problem with hiding author names. But I think some of this problem can be alleviated by the fact that better comments usually float to the top anyways.

Also perhaps this feature can be something that is only disabled after reaching a karma threshold?

For example, I remember reading earlier comments about the negative impact of working all night long on a start up. If I had said this, people would (rightly) ignore me. But if someone who has a startup had said this, it should carry more weight.

(Obviously however there are some comments that do not warrant discussion but this probably would end up fitting into the types of flags.)

It can result in less accurate filtering. Popular comments may be voted up simply because someone sees the amount of points, skims the comment with perhaps a bare understanding, and thinks "hey, this guy's probably right." And then, ding, a possibly unwarranted upvote. The same effect can apply to an unpopular comment, but the effect is probably stronger for unpopular comments, which is why I think it is so dangerous. An unpopular comment would suffer more of this effect because people who disagree may build up more negative emotions, and this emotion may lead to more impulsive, irrational behavior making them more susceptible to the bias.

Seeing a comment he/she disagree with, and seeing that a bunch of people agree with him/her about disagreeing gives that person a sense of approval which erodes his/her ability to truly evaluate a dissenting comment objectively.

If people have no incentive to refrain from bad behavior except the occasional posting of the commenting guidelines to the front page, then they will engage in bad behavior. People respond to incentives.

Flags can play the role of downvotes for controlling spam and totally rubbish comments.

An upvote on the otherhand is "I'm providing positive feedback".

Another interesting experiment would be to display the top downvoters from time to time. We'll at least be able to weed out some of these invisible trolls.

And if you notice, it's like pulling teeth from an angry lion to even get any of these guys to bother posting a reason. The answers so far seem to amount to either

1) My time is too important, I'm too important.

2) The downvote is enough reason, it should be obvious to the recipient that they weren't contributing anything.

I'm sorry, but I don't find either of those reasons even remotely satisfying.

I think you are taking downmods a bit too personally.

I don't want to limit the amount of times a person can disagree with me (at risk of allowing for disagreeable persons).

However, a lot of that might be solved by using words instead of arrows. E.g. a "Agree", "Disagree" and "Detracts From Conversation" button.