Is there any reason you chose to compare these rules to HN's rules? HN's moderation is one of the best I've ever encountered, I'm not sure there's much of a need to improve it.
It is. But compared with FB, HN has two distinct advantages:
a) The scale is small enough that most content can be manually reviewed, or reviewed after it reaches any significant number of views.
b) People in HN are expecting a professional discussion forum and are largely ok with a relatively active moderation. People posting in FB often assume that any moderation by Facebook the company must comport to freedom of the press or freedom of speech laws. This is of course legally nonsense, but it might not be an unreasonable expectation over all. In some ways, because of the scale of online platforms, the societal effects of FB/Twitter/Youtube content policing norms is closer to say, a code about what is allowed or not on-air in radio/TV in a country than a particular set of editorial policies for a particular newspaper.
HN's moderation works only because Scott and Dan patrol threads flagging and banning people who are (for instance) openly anti-Semitic on threads. It would be better to have simple, coherent rules against these things, rather than unstated rules that require enormous amounts of effort to enforce.
Again I want to be clear that most of these rules seem silly or even puerile. Just the beginning of it made sense.
Frankly I think it's starting to fray a little around the edges. Downvotes and flags are supposed to be used to suppress non-constructive comments and submissions but I see them being used more and more to express disagreement (or, to be more precise, I see more and more comments being downvoted and stories being flagged that seem to me to be constructive and respectful but which express unpopular points of view.)
For the downvoters of the parent, here is the citation:
PG: "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."
He wrote that back when downvoted comments didn't get greyed out the way they do today. Because of the grey-out, a downvote today is effectively a vote for, "No one should read this." IMHO, disagreement should not have that effect. There really ought to be a way to distinguish between "I disagree" and "This comment is not constructive and should be suppressed" -- and there is. If you disagree, you can post a comment to that effect. But whether PG intended it or not, a downvote today is suppression because of the way HN works.
Thanks for pointing out the context of when the pg comment was made. I wasn't aware of this, and the comment makes more sense to me now. It also reinforces—for me at least—that having separate "agree/disagree" and "constructive/unproductive" button pairs may be a good idea. Open questions:
- how they'd play into karma (probably just the constructive/unproductive points, maybe agree/disagree displayed as a pair but not contribute to karma)
- whether this is would be too complex from a UX perspective
- while I think "disagree" and "unproductive" can be usefully distinguished, whether "agree" and "constructive" would be useful. That said, if we want people to comment for disagreement (substantively, I would hope), I wouldn't want to see or encourage a bunch of "+1" or "me, too", or "this" comments. I think users want to show support for good comments, and up-votes are currently the only way to do this if you have nothing else to add.
Actually, HN already has a separate "flag" link. The only real change needed would be to dim based on flags rather than downvotes. Then downvotes could go back to their original meaning (disagree).
I would abolish karma. It turns HN into a silly game.
I agree that dimming should be a function of constructiveness rather than agreement. If flag is to be used for that purpose, there's currently no way to express "I think this is a constructive comment", no way to balance "I think this is not a constructive comment". Having such a button is effectively the same as having "constructive/unproductive" buttons: the flag link would be subsumed.
As for karma, I think it serves two purposes: it acts as a gating function to ensure members have contributed usefully before allowing them to behave in a potentially unproductive manner (e.g., creating throwaway accounts to down-vote or flag). It also encourages continued constructive contributions. Yes, it's artificial, but that doesn't necessarily mean that it can't serve that purpose. I don't think karma is perfect, but I think it is, on balance, a constructive feature on HN.
If there are other methods of serving those two purposes more effectively and as simply as karma does (and I admit to ignorance here), I think they'd be worth exploring. I think there needs to be some pressure over the lifetime of an account to encourage constructive comments, non only on a per-comment basis.
"Agree" generally implies "constructive" so these are not orthogonal. You really need three options: agree, disagree but constructive, and non-constructive. Only the last one should count towards suppression.
As for karma, you make a good point. Have an upvote :-)
Back at ya! Very good point regarding {agree|disagree but constructive|non-constructive}. Which really just leads us back to up/down + flag, doesn't it? Then both up-voting and down-voting would count towards "un-dimming" while flagging would count towards dimming, I think. Thanks for working through this with me...er, bringing me around to your opinion :)
- It really should be a ternary selector, not up/down + flag. You should only be able to select one of the three.
- Having both up and down contribute to undimming would actively discourage using disagree as "punishment". It would very likely change how "flag" is used, though the proper use of "flag" is unclear to me right now, as from what I can tell its effect is not visible until the comment is dead. It would clarify the effect between "down" and "flag".
I can't think of any comments 'Dead' because of disagreement.
I've seen feedback from users that they feel that this is exactly why a comment was down-voted, whether it gets killed or not. I admit I don't have any dead comments at hand, and I don't think the search API returns dead comments. Granting the anecdatum, here's one that I came across:
I think users blame downvotes on disagreement, but in reality it's often badly formed written viewpoints lacking in value or OT.
I think that's often (but not always) the case. That said, I don't know how to convince those who think otherwise. One person's weakly argued comment might be someone else's disagreement. Compare with the Political Detox Week experiment: there was quite a bit of expressed disagreement as to what "politics" means. Or what constitutes civility. Or what's off-topic. It's not only whether the down votes are due to disagreement; it's also whether the commenter perceives the down-votes as such (which is the case of the example I provided).
Edit to add: There will always be some percentage of members who won't agree, and they're more likely to be vocal. I'm not sure how to measure when the optimum is reached (most users feel the system is working as expected).
If anyone is still paying attention here, can you help conduct this experiment by down-voting aaron695's comment to see what happens? I can't do it myself because you're not allowed to down-vote responses to your own comments.
"Constructive" is interestingly subjective. The higher the raise the bar, the more likely it is to be controversial because e.g. you require posters to share and build on the same assumed knowledge. You are welcomed to challenge assumed knowledge from a position of expertise but not from a position of ignorance.
I give "I disagree" downvotes on topics I am very familiar with and the poster is simply insufficiently experienced to be broadcasting their bogus opinions however polite they are. I can surely imagine even more inexperienced posters would see that poster as constructive and helpful but its not the case.
Unfortunately, "more coherent" doesn't mean "better", it just means "more fleshed-out". The problem with formalizing rules of human behaviour is that the latter is generally too fuzzy, so you either end up with a very complex system of rules, or with a simple system that has tons of edge cases (like the one in the post).
I think you were attracted to the improvement of writing rules down, but that has a host of other problems, including the fact that you are then forced to enforce the rules, even where they don't make sense (e.g. you have to remove some comment because of the letter of the law, even though it wasn't actually inflammatory, it just ran into an edge case).
I moderate(d) a large reddit community (millions of subscribers?) and you pretty much can't win. If you don't have rules, you're "murky" and "wanton". If you have rules, you get "these rules suck" or "you're only enforcing the rules whenever you feel like it" when you try to go by the spirit of the law rather than the letter.
To illustrate your point, Facebook censored the Pulitzer Prize-winning photo taken June 8, 1972 with crying children, including 9-year-old Kim Phuc, after an aerial napalm attack on suspected Viet Cong hiding places.
That photo violates a several of Facebook "ban" rules, but is an exception to the rules. After much outcry, Facebook reversed their rule-based ban.
Why? What makes you certain of that? I'm not suggesting that Dan and Scott stop moderating, just that it's possible to make a clear, coherent statement about bigotry in the moderation guidelines.
My experience with moderating large communities and, mainly, my inability to define bigotry to any satisfactory degree. I certainly urge you to give it a go, maybe you'll be able to solve the problem for all of us.
Hell, I can't even ballpark where your freedom to speak ends and my freedom to not feel bad starts.
It would be better to have simple, coherent rules against these things, rather than unstated rules that require enormous amounts of effort to enforce.
Would it?
It would certainly require less effort, but we know from other places that blanket rules are pretty easy to game (eg, see the games people play on Wikipedia with the 3RR rule).
Once people know them, the rules themselves become weaponized.
