A model could be Stackoverflow.com which has an option to mark a post, thread or single comment as “karma neutral” (in LW terms; they call it something else). The up- and downvotes still work, they’re just not applied to the author’s karma.
There’s some slight potential for abuse if that were transposed here, insofar as someone could take advantage of it to troll with impunity; but I don’t see someone being amicable enough in the first place to amass much karma, and then turning around and becoming a troll.
I’m personally uncomfortable with endorsing another voting system that functions just like karma but with a different name. Having more than one set of numbers that can fluctuate from positive to negative and are differentiated in name alone sounds like a headache, as well as an interface that would scare me away were I not already familiar with the site.
Instead, I propose simple “I Agree” and “I Disagree” buttons. Rather than just tallying the amount of votes for either, it should simply attach one’s name to the post via a pulldown tab, just like hiding and exposing child tabs in the comments section. Ideally the lists of those who agree or disagree, as well as how many votes are for either, would not be revealed until one clicked to reveal it. Or possibly make it so that one would have to vote before being able to see the results. Hopefully that would prevent people from being primed to agree with the majority (or being overly contrary).
This would reveal which discussions are actually contentious versus which are being supported by a vocal minority. It could also function as an all-purpose karma-neutral voting system.
Some possible problems: While I want to know who agrees with what, and from a Bayesian perspective this is going to help me be more accurate, I’m fearful that I would be too inclined to agree with prominent posters. I may be unable to discern whether I truly agree with something or am inclined to agree because I see that Eliezer supports an argument. Handled incorrectly it could turn into an in-group out-group situation.
edit: I forgot to include the purpose of having a list of names instead of an anonymous number. I want to know who’s judgement I am relying on in matters that I don’t have time to do research myself. Also, should I disagree, knowing who I am refuting would help to motivate me. Knowing that I am arguing with a practicing rationalist instead of an anonymous number helps encourage me to construct better arguments.
That’s why I suggested “no net positive karma”. It would be the exact same system as you just described, except that votes would be applied to an author’s total karma if the cumulative votes for that comment were negative.
I’d prefer a “karma neutral” setting, going both ways: one thing that strikes me as an issue here is that an “open” thread isn’t quite as open as it might be if you could declare it safe to post there without worrying about losing karma.
A model could be Stackoverflow.com which has an option to mark a post, thread or single comment as “karma neutral” (in LW terms; they call it something else). The up- and downvotes still work, they’re just not applied to the author’s karma.
There’s some slight potential for abuse if that were transposed here, insofar as someone could take advantage of it to troll with impunity; but I don’t see someone being amicable enough in the first place to amass much karma, and then turning around and becoming a troll.
I’m personally uncomfortable with endorsing another voting system that functions just like karma but with a different name. Having more than one set of numbers that can fluctuate from positive to negative and are differentiated in name alone sounds like a headache, as well as an interface that would scare me away were I not already familiar with the site.
Instead, I propose simple “I Agree” and “I Disagree” buttons. Rather than just tallying the amount of votes for either, it should simply attach one’s name to the post via a pulldown tab, just like hiding and exposing child tabs in the comments section. Ideally the lists of those who agree or disagree, as well as how many votes are for either, would not be revealed until one clicked to reveal it. Or possibly make it so that one would have to vote before being able to see the results. Hopefully that would prevent people from being primed to agree with the majority (or being overly contrary).
This would reveal which discussions are actually contentious versus which are being supported by a vocal minority. It could also function as an all-purpose karma-neutral voting system.
Some possible problems: While I want to know who agrees with what, and from a Bayesian perspective this is going to help me be more accurate, I’m fearful that I would be too inclined to agree with prominent posters. I may be unable to discern whether I truly agree with something or am inclined to agree because I see that Eliezer supports an argument. Handled incorrectly it could turn into an in-group out-group situation.
edit: I forgot to include the purpose of having a list of names instead of an anonymous number. I want to know who’s judgement I am relying on in matters that I don’t have time to do research myself. Also, should I disagree, knowing who I am refuting would help to motivate me. Knowing that I am arguing with a practicing rationalist instead of an anonymous number helps encourage me to construct better arguments.
That’s why I suggested “no net positive karma”. It would be the exact same system as you just described, except that votes would be applied to an author’s total karma if the cumulative votes for that comment were negative.
I’d prefer a “karma neutral” setting, going both ways: one thing that strikes me as an issue here is that an “open” thread isn’t quite as open as it might be if you could declare it safe to post there without worrying about losing karma.