I hate StackOverflow's downvoting with the heat of a trillion simultaneous supernovas all going off at the same time. It's a blunt instrument applied thoughtlessly and indiscriminately.
Strange strange strange to hear StackOverflow lecturing others on the value of downvoting.
Can you elaborate on this? Your respone seems very emotional, but you don't provide any facts.
In my experience with SO, if an answer has a score below zero, it's because the answer is either irrelevant or wrong. And I do sometimes take time to read all the answers to a question trying not to judge them based on their score.
Could you provide some more information or maybe ideas for a better system?
From what I've experienced as a casual user on SO (<200 rep in over a year), most of the times I (or anybody else) am downvoted is because:
1. My question was too subjective. I have edited my questions to narrow down the scope to avoid this after somebody commented the reason for downvoting and the question did get answered then.
2. I did not do enough research. This is a tough one because facing this makes me feel inadequate about myself. It basically happens when I fail to provide any indications that I tried solving. Such questions typically look like homework questions instead of people stuck somewhere.
But yes, there have been some times when people downvoted me but did not provide an explanation but that has been very rare for me.
One idea I had was to be able to see who downvoted a post and be able to list of the last 50 or so posts a particular user has downvoted. Thus would hopefully shame users who downvote for reasons other then to keep the conversation constructive.
You could even allow downvoting of the downvote, so users could lose karma for downloading an insightful post that they simply disagreed with.
Of course, if you allowed downvoting the downvote of a downvote, and so on recursively, you could have downvote wars between two users, where they both drain each other's karma down to zero. I don't see much problem with this, because anyone willing to participate in such a battle seems childish enough that they deserve what they get.
The best variant of this idea that I've seen is - believe it or not - Slashdot. The key aspect of their system is that you don't get to choose which actions or users to metamoderate. You just get a random set, which breaks the endless downvote cycle you mention.
I don't think "they deserve what they get" is sufficient, because bilateral symmetric conflict is not the only scenario. It's easy enough for an ideologically aligned group to hammer an outspoken invididual with metavotes, just as they already do with votes. This can happen with or without explicit coordination, and carries almost no cost/risk for each individual in the lynch mob. There's also the scenario of a karma-rich individual continuing to reap the dividend of the "rich get richer" phenomenon common to all social networks (which this is, whether or not denizens like to admit it). Such a person can afford to spew out hundreds of undeserved downvotes a week. I've been targeted that way here many times, and I'll bet you have too. What SO prescribes is already one of the worst things about HN. Making it easier for those petty individuals to identify targets, and more ways to harass them, scarcely seems like it would help.
BTW, I'm not suggesting that the /. system is ideal. Their meta/moderation system has its flaws, as do other aspects of the site. I'm just saying that the random element specifically seems to avoid a lot of pathologies that are common to most such systems.
> ... and carries almost no cost/risk for each individual in the lynch mob.
I see a big cost if the person downvoted can downvote those of the lynch mob. Sure, the lynch mob could downvote the downvote of the victim but I don't see that potential vicious cycle being a winning strategy for suppressing an unliked idea.
Wrt the karma rich individual. This change wouldn't help the "rich get richer" because it only opens up the algorithm for additional downvotes. Therefore, the only result that could occur would be everyone, including the rich, having a zero effect on their karma, or a lessening of it.
Wrt slashdot, personally, I could never stand meta-moderating. It's so boring to read random comments about a topic that you are not interested in. Maybe it's different for you, but I'd like to know how popular meta-moderation is.
Also, the results speak for themselves. Slashdot has gone down in quality a lot over the years, the signal to noise ratio is currently very low. I hardly ever browse the comment section there, anymore. Maybe for you this is different?
> I see a big cost if the person downvoted can downvote those of the lynch mob.
When it's many to one, each member of the mob only has to absorb 1/N of the target's downvotes. For large N, the target will be exhausted before any of their attackers feel a thing.
> I don't see that potential vicious cycle being a winning strategy for suppressing an unliked idea.
Just because you don't see it doesn't mean it's not real.
> This change wouldn't help the "rich get richer"
That's not the point. The rich already get richer, because people with high reputation get more attention and therefore more new reputation. Thus, they have a steady inflow of new karma that they can burn on a whim without any real consequence. Even I, at my medium karma level, do this sometimes. I'll post a comment that I know will attract a few downvotes, because maybe some of the less ideologically-rigid folks here will appreciate it and I'll make up the difference on the next comment. Many of the highest-karma people here constantly do that to behave in far more egregious ways. What your proposal does is just give the bullies a new weapon.
> Also, the results speak for themselves. Slashdot has gone down in quality
Have you considered the possibility of there being other factors at play? Yes, Slashdot has gone downhill, but I think it would have gone downhill even faster with HN's system, and faster still with your proposal.
> When it's many to one, each member of the mob only has to absorb 1/N of the target's downvotes. For large N, the target will be exhausted before any of their attackers feel a thing.
They can do mob-downvotes today, only they don't get punished at all. This would be an improvement.
If a mob wanted to hurt only one target, your model makes sense. But a mob with an agenda would need to downvote hundreds of posts, at least, to make a significant impact. So having their karma drained while they do so would be a dissuading factor.
> What your proposal does is just give the bullies a new weapon.
It also gives people a chance to defend themselves. Right now, bullies can not only get away with their tactics with no fear of retribution, but they can also do so anonymously. This new system at least gives the victims a chance to "punch the bully in the nose", regardless of whether that means even a harder beating or not.
Also, that you could then "name and shame" even the most karma rich individuals as bullies, by directing people to their profile and "unfair" downvotes, this is another way of fighting back against them.
When you weigh that against a bully being able to simply add another negative point, and only if the victim has decided to fight back against the bully (by downvoting their downvote), it seems like a win in my opinion.
> Yes, Slashdot has gone downhill, but I think it would have gone downhill even faster with HN's system, and faster still with your proposal.
I disagree, but anyway bringing up Slashdot when you admit that its quality is subpar is not evidence that its system is a more ideal model.
In both cases you have the same data loss problem. Without downvotes you can't separate uninteresting from harmful and with downvotes you can't separate harmful from low quality which is why voting etiquette is recommended:
> If someone spends the time to make an honest effort to answer a question, but it's not that great an answer, just don't upvote them
That means voting etiquette is part of the vote mechanic itself; a critical piece that makes the technology work rests in the minds of its users. If you've spent any time on Reddit, it's obvious this is a common source for trouble as a person seeing an up/down binary will naturally think this is just a personal value assessment. This is amplified by the fact downvoting has actual effects, lowering the visibility of a contribution encourages downvotes to hide content the user thought was poor quality.
This is why FB reactions happened. A mechanism intended to reduce noise has to guard against data loss to some degree or the mechanism itself produces noise, as it is with lossy audio formats.
There's an interesting solution: showing upvote and downvote counters separately. I would much more enjoy to read somebody who has plenty of upvotes and a good share of downvotes than somebody who has mostly only upvotes. I'd suspect the latter to be a karma hoarder who tries not to post anything that any big part of a community disagree with.
13 comments
[ 3.2 ms ] story [ 32.7 ms ] threadStrange strange strange to hear StackOverflow lecturing others on the value of downvoting.
In my experience with SO, if an answer has a score below zero, it's because the answer is either irrelevant or wrong. And I do sometimes take time to read all the answers to a question trying not to judge them based on their score.
From what I've experienced as a casual user on SO (<200 rep in over a year), most of the times I (or anybody else) am downvoted is because:
1. My question was too subjective. I have edited my questions to narrow down the scope to avoid this after somebody commented the reason for downvoting and the question did get answered then.
2. I did not do enough research. This is a tough one because facing this makes me feel inadequate about myself. It basically happens when I fail to provide any indications that I tried solving. Such questions typically look like homework questions instead of people stuck somewhere.
But yes, there have been some times when people downvoted me but did not provide an explanation but that has been very rare for me.
You could even allow downvoting of the downvote, so users could lose karma for downloading an insightful post that they simply disagreed with.
Of course, if you allowed downvoting the downvote of a downvote, and so on recursively, you could have downvote wars between two users, where they both drain each other's karma down to zero. I don't see much problem with this, because anyone willing to participate in such a battle seems childish enough that they deserve what they get.
I don't think "they deserve what they get" is sufficient, because bilateral symmetric conflict is not the only scenario. It's easy enough for an ideologically aligned group to hammer an outspoken invididual with metavotes, just as they already do with votes. This can happen with or without explicit coordination, and carries almost no cost/risk for each individual in the lynch mob. There's also the scenario of a karma-rich individual continuing to reap the dividend of the "rich get richer" phenomenon common to all social networks (which this is, whether or not denizens like to admit it). Such a person can afford to spew out hundreds of undeserved downvotes a week. I've been targeted that way here many times, and I'll bet you have too. What SO prescribes is already one of the worst things about HN. Making it easier for those petty individuals to identify targets, and more ways to harass them, scarcely seems like it would help.
BTW, I'm not suggesting that the /. system is ideal. Their meta/moderation system has its flaws, as do other aspects of the site. I'm just saying that the random element specifically seems to avoid a lot of pathologies that are common to most such systems.
I see a big cost if the person downvoted can downvote those of the lynch mob. Sure, the lynch mob could downvote the downvote of the victim but I don't see that potential vicious cycle being a winning strategy for suppressing an unliked idea.
Wrt the karma rich individual. This change wouldn't help the "rich get richer" because it only opens up the algorithm for additional downvotes. Therefore, the only result that could occur would be everyone, including the rich, having a zero effect on their karma, or a lessening of it.
Wrt slashdot, personally, I could never stand meta-moderating. It's so boring to read random comments about a topic that you are not interested in. Maybe it's different for you, but I'd like to know how popular meta-moderation is.
Also, the results speak for themselves. Slashdot has gone down in quality a lot over the years, the signal to noise ratio is currently very low. I hardly ever browse the comment section there, anymore. Maybe for you this is different?
When it's many to one, each member of the mob only has to absorb 1/N of the target's downvotes. For large N, the target will be exhausted before any of their attackers feel a thing.
> I don't see that potential vicious cycle being a winning strategy for suppressing an unliked idea.
Just because you don't see it doesn't mean it's not real.
> This change wouldn't help the "rich get richer"
That's not the point. The rich already get richer, because people with high reputation get more attention and therefore more new reputation. Thus, they have a steady inflow of new karma that they can burn on a whim without any real consequence. Even I, at my medium karma level, do this sometimes. I'll post a comment that I know will attract a few downvotes, because maybe some of the less ideologically-rigid folks here will appreciate it and I'll make up the difference on the next comment. Many of the highest-karma people here constantly do that to behave in far more egregious ways. What your proposal does is just give the bullies a new weapon.
> Also, the results speak for themselves. Slashdot has gone down in quality
Have you considered the possibility of there being other factors at play? Yes, Slashdot has gone downhill, but I think it would have gone downhill even faster with HN's system, and faster still with your proposal.
They can do mob-downvotes today, only they don't get punished at all. This would be an improvement.
If a mob wanted to hurt only one target, your model makes sense. But a mob with an agenda would need to downvote hundreds of posts, at least, to make a significant impact. So having their karma drained while they do so would be a dissuading factor.
> What your proposal does is just give the bullies a new weapon.
It also gives people a chance to defend themselves. Right now, bullies can not only get away with their tactics with no fear of retribution, but they can also do so anonymously. This new system at least gives the victims a chance to "punch the bully in the nose", regardless of whether that means even a harder beating or not.
Also, that you could then "name and shame" even the most karma rich individuals as bullies, by directing people to their profile and "unfair" downvotes, this is another way of fighting back against them.
When you weigh that against a bully being able to simply add another negative point, and only if the victim has decided to fight back against the bully (by downvoting their downvote), it seems like a win in my opinion.
> Yes, Slashdot has gone downhill, but I think it would have gone downhill even faster with HN's system, and faster still with your proposal.
I disagree, but anyway bringing up Slashdot when you admit that its quality is subpar is not evidence that its system is a more ideal model.
> If someone spends the time to make an honest effort to answer a question, but it's not that great an answer, just don't upvote them
That means voting etiquette is part of the vote mechanic itself; a critical piece that makes the technology work rests in the minds of its users. If you've spent any time on Reddit, it's obvious this is a common source for trouble as a person seeing an up/down binary will naturally think this is just a personal value assessment. This is amplified by the fact downvoting has actual effects, lowering the visibility of a contribution encourages downvotes to hide content the user thought was poor quality.
This is why FB reactions happened. A mechanism intended to reduce noise has to guard against data loss to some degree or the mechanism itself produces noise, as it is with lossy audio formats.
- 1 year ago, 36 comments: https://news.ycombinator.com/item?id=10875619
- 8 years ago, 114 comments: https://news.ycombinator.com/item?id=507948
I could clean up the front page much faster and still see which articles are interesting.