Warn HN: Downvoting sucks :(
Maybe that is why FB does not have an Unlike button...
So on both sites you can articulate something very well, but if it is unpopular then it will be hidden (Reddit) or dead (HK). That seems, to me, like that isn't why hiding/removing was implemented (more for spam/troll/unconstructive).
What both communities need is a up arrow (agree), down arrow (disagree), and a flag button (troll/spam/unconstructive). You then stop hiding on unpopular opinions, and instead only hide flagged ones (e.g. if flag > (agree + disagree) then hide).
You can try to stop people using flag as "super-downvote" by making it so that after you flag a post you cannot either agree OR disagree with it, and additionally you cannot reply to it (effectively making you choose, to either "disagree" or flag, and since "disagree" is public and flag is private you're more encouraged to "disagree" than flag).
Further to this instead of giving people a literal count of "karma" (e.g. 10,000 karma) you give them a pie chart with % of popular, controversial, and removed comments (e.g. 30%, 68%, 2%).
Essentially the goal is to remove all downsides to saying unpopular things, but to keep some of the downsides of saying trolly/spam/unconstructive things. Which will create a better back and forward dialog.
I disagree. What communities need -- if they value contributions of substance rather than agreement -- is community members willing to upvote opinions that are substantively presented and add to the discussion even they disagree with them. Adding more user choices won't stop people from using the one intended to suppress inappropriate comments on comments they disagree with if they, in fact, want to suppress disagreement.
Fortunately for HN, I think HN has a sufficient number of those type of members -- most of the complaints I've seen about things being downvoted because of unpopularity (often including, in the same discussion, things on both sides of the issue) have been presented in a way which presents alternative explanations for downvotes than to the popularity of the opinion, and most of the substantive comments, whether voicing the dominant opinion or not, that I've seen that get downvoted subsequently get upvoted. (Certainly, I know that I've frequently upvoted comments that I disagree with.)
But as long as people have a tool that lets them suppress opinions they disagree with, you aren't going to prevent people who would be inclined to use it from doing so by presenting other alternatives.
> You can try to stop people using flag as "super-downvote" by making it so that after you flag a post you cannot either agree OR disagree with it, and additionally you cannot reply to it (effectively making you choose, to either "disagree" or flag, and since "disagree" is public and flag is private you're more encouraged to "disagree" than flag).
I think this rests on a false assumption -- private use of flag doesn't reduce the incentive to use it, it just reduces accountability. And, anyway, with your proposed hiding algorithm (hide if flag > agree+disagree), it would be stupid to use the UI option to "disagree" after you flagged anyway, since that would negate the effect of your flagging on its only purpose -- suppressing the opinion you disagree with, rather than publishing your disagreement.
The status quo prove that that is a failed strategy. It has been tried and tried again, it doesn't work. Nobody has been able to make it work, I don't see that changing without any other variables changing. People innately want to downvote to disagree, that's how people have proven they work.
> Adding more user choices won't stop people from using the one intended to suppress inappropriate comments on comments they disagree with if they, in fact, want to suppress disagreement.
It will for the reasons I set out above: Flagging stops them interacting with the comment (replying, disagreeing) and a "disagree" is a public disapproval rather than a private one which likely suits many people's ends better.
Additionally you now have more metrics (data) to measure a comment's quality on (e.g. agree, disagree, and flag) rather than two. More data might make it easier to detect flag-button abuse.
> I think this rests on a false assumption -- private use of flag doesn't reduce the incentive to use it, it just reduces accountability.
The upvote and downvote arrows have no more or less accountability, so I don't follow what you're getting at. The flag arrow just wouldn't be recorded in the comment's public scores, accountability remains identical (in that there isn't any).
> it would be stupid to use the UI option to "disagree" after you flagged anyway, since that would negate the effect of your flagging on its only purpose -- suppressing the opinion you disagree with, rather than publishing your disagreement.
It would on paper, it wouldn't for real people. While there are people who sit there and figure out how to best game the system, the majority just knee-jerk react to things they find "offensive." So they would hit flag, then disagree, then reply with "you are dumb." If we take away 2/3 of their toys if they hit flag incorrectly, they're less likely to utilise flag and instead just disagree and reply.
Plus it re-focuses flag's purpose. If you flag a comment you literally don't think it should exist, therefore you should have no need to interact with it. Heck you could go one step further and auto-hide it for that individual user.
I disagree, I think it works quite well on HN, and the vast majority of the specific instances of it supposedly not working I've seen pointed to are nonconstructive comments -- some of which supported viewpoints I agree with a non-constructive way -- which I agree should have been downvoted for non-disagreement reasons.
I do think there will always be people who are to prone to use suppressive options when they disagree with a comment, but I think that HN currently has a community where that isn't the dominant model and the people that complain about a problem of use of downvote-for-disagreement being a real problem with the community are people that are a selectively blind to the negative qualities of comments that happen to support points of view that they like.
> It will for the reasons I set out above: Flagging stops them interacting with the comment (replying, disagreeing) and a "disagree" is a public disapproval rather than a private one which likely suits many people's ends better.
For the people with which a public interaction suits their ends better, responding already exists. For people for which suppression suits their ends better, flagging will be superior even if it isn't the intended use of flagging. I don't think people who do use downvote-for-disagree in a community with clearly articulated norms of what downvoting is for generally do so out of ignorance or lack of options, they do so because downvote suppresses and that is exactly what they want to do.
Or because they genuinely can't separate whether something is non-constructive (and thus should be downvoted) from whether it is a viewpoint they agree with -- much the same as the people who have an excessive tendency to perceive any downvoting of a comment they agree with as the downvoter using downvote-to-disagree. And people who can't support "inappropriate" from "I disagree" are going to keep flagging as inappropriate based on disagreement, no matter what options are available.
OTOH, feel free to set up your own community site with the features you discuss and prove how those options solve the supposed problem.
Allowing people to interact negatively with ads (in the sense of "paid messages") in a way that affects public perception of the message reduces the value of purchasing advertisement; since FB uses a model where essentially content of all types can be ads (that is, paid messages) -- since "free" messages of any type of limited reach which you can pay to increase without changing the type of message -- allowing negative interactions would interfere with their revenue modelb reducing the value of paying for reach.
And that's why there is no Dislike button on FB. You can choose not to see things, and you can interact with things that make other people less likely to see them. But you can't choose absolutely to get things in your feed, and you can't interact with them in a way that reduces other people's likelihood of seeing them to less than it would be if you didn't interact with them at all.