zlacker

[parent] [thread] 51 comments
1. strlen+(OP)[view] [source] 2011-04-03 20:41:13
Cap the score that is displayed with a comment e.g., past 10 points, just display "10+". Don't display karma and average scores of users, again, past a certain point: this prevents (subconscious) game incentives which lead to e.g., posting comments that say something stupid or mean but which tend to agree with general tendencies of the site.

For example, I can post a comment decrying Blub with a snide remark (e.g., "You wrote a 1,000 line Blub program? Was it 500 getters and 500 setters?" in a thread discussing software projects) that is both information free and mean (perhaps Blub wasn't the author's preferred choice, but chosen for him or required in order to build an application for the iBlubber). People on this site generally dislike Blub, so the comment will get upvotes without adding any value to the discussion (an example of adding value would be saying you were able to do this in 100 lines of Flub using its cool new hygienic macros with a link to a paper on hygienic macros in Flub).

That's not to say all comment score data should be gone. Comment scores can still be kept and comments could be displayed on stories in the other in which they're displayed now (a mix of comment score and how recently it was posted). Generally, what I've found is that comments showing up _first_ tend to be of higher quality i.e., overall algorithm works more often than not.

[NB: I work at LinkedIn and we do this for connection counts-- we want users to network with each other, but we don't want to make it a "who has the most connections" game, that's why when you have over 500 connections (which is perfectly legitimate and allowed), only "500+" is displayed as the count on your profile]

replies(6): >>jerome+D2 >>tptace+P3 >>Tycho+U3 >>mixmax+c4 >>_delir+s5 >>crassh+gh
2. jerome+D2[view] [source] 2011-04-03 21:15:11
>>strlen+(OP)
Many people are suggesting replacing numeric point totals. That could help, but let me throw out another idea. Why would someone write "You wrote a 1,000 line Blub program? Was it 500 getters and 500 setters?" Simple, they seek to build community reputation/acclaim by coming across as clever, even if it means being mean. But it's easy to take that incentive away. I propose leaving everything the way it is, except two changes: hide all usernames (make visible upon hover/click) and reset visible karma points daily. Truly interesting/insightful comments would still rise up, but with even more clarity as people would be reading for content before casting their vote, and incentive to be "clever" while adding no value would be gone.

Edit: I like my later thoughts on this better (http://news.ycombinator.com/item?id=2404267)

3. tptace+P3[view] [source] 2011-04-03 21:32:05
>>strlen+(OP)
The blub setter/getter comment really is the archetypical "dumb comment", isn't it? Nicely done.

Users should live or die by their votes on that comment. If you vote up the blub comment, you should personally get the downvotes for it too. Upvotes should expose you to the karmic downside of superficial comments.

Especially because the really good comments, the ones most deserving of upvotes, don't seem to get a lot of downvotes; watch the scores on a 'patio11 comment closely sometime to see an example.

replies(3): >>nyelli+r9 >>strlen+Ec >>dpk+U59
4. Tycho+U3[view] [source] 2011-04-03 21:33:04
>>strlen+(OP)
Actually, you just gave me a great idea (I think). The problem is mainly snide remarks, right? So make it impossible to make snide remakrs. How?

Minimum comment length of ~50 words.

This would A) get rid of casual snideyness, as those sorts of people wouldn't put in the effort to formulate a longer post; B) discourage crowd-pleasing one liners, which while enjoyable have a long term negative effect; C) still allow jokes, they'd just have to be asides to actual substance; D) encourage longer, better thought-out posts in general, and backing up of claims.

replies(9): >>jnovek+O4 >>tptace+85 >>bmelto+i5 >>dolins+N5 >>rbanff+d6 >>cabala+Na >>crassh+Di >>ww520+0j >>run4yo+nW7
5. mixmax+c4[view] [source] 2011-04-03 21:38:42
>>strlen+(OP)
You, and other users both in this thread and in similar ones before it, are only looking at one side of the equation.

The problem is only partially too many bad comments, it's also too few good comments. Your proposal is targeted at lowering the amount of bad comments, but it might do so at the expense of the good ones.

Most of the users that have a high karma count have so because they always have something insightful to say (mechanical_fish (http://news.ycombinator.com/user?id=mechanical_fish) is a good example), and their incentive for posting interesting stuff is marginalized by this. Almost all people, whethere they'll admit it or not, are incentivized by other peoples approval, eg. karma, and downplaying their contributions will make them more prone to not submitting great comments.

replies(2): >>bvi+R8 >>strlen+za
◧◩
6. jnovek+O4[view] [source] [discussion] 2011-04-03 21:49:51
>>Tycho+U3
This reminds me of the Slashdot lameness filter, which IIRC just caused (causes? I haven't been there in a while) lots of posts ending in, "La la la la, adding some content to get past the lameness filter." and the like.
replies(1): >>jerome+v7
◧◩
7. tptace+85[view] [source] [discussion] 2011-04-03 21:57:56
>>Tycho+U3
You just banned 10 out of 12 of the 'pg comments on this thread. :)
replies(3): >>Tycho+u5 >>jerome+W5 >>agscal+6h
◧◩
8. bmelto+i5[view] [source] [discussion] 2011-04-03 22:00:15
>>Tycho+U3
Short and snide are by no means mutually exclusive.

I can spend all day writing a post that has no substance, garners upvotes (in today's HN climate) and is as long as the day, but that doesn't make it good, or worthwhile.

Many of my short remarks are actually more to the point, or are asking a relevant question, or are politely correcting errors. It takes considerably fewer words to do this sort of thing than it does to be mean, while trying not to look like it.

replies(1): >>Tycho+F5
9. _delir+s5[view] [source] 2011-04-03 22:01:48
>>strlen+(OP)
The Kuro5hin solution of displaying averages plus number of ratings might give information in that direction, factoring out into two different numbers the average quality people thought the comment had, from the number of people who rated it. Not quite the same as a threshhold maximum, but it produces an asymptotic maximum instead. It's also, imo, useful to have them as two separate numbers, because they convey semantically different information.
replies(2): >>strlen+ec >>crassh+Fi
◧◩◪
10. Tycho+u5[view] [source] [discussion] 2011-04-03 22:02:01
>>tptace+85
Ha. It would certainly lead to less comments overall, but maybe that would be a welcome timesaving. Who knows, people might even start pre-emptively addressing counter-points for their arguments and the whole debating process would be greatly contracted. (am I at 50 words yet? maybe 30 would be better...)
◧◩◪
11. Tycho+F5[view] [source] [discussion] 2011-04-03 22:05:58
>>bmelto+i5
Valid points, but I think for any worthwhile comment it's easy to 'extend' the content beyond a single line. For instance, if you're asking a sincere question, you could provide some hypothetical answers, or state your reasons for asking - basically increase the level of mutual comprehension in the entire discourse.
replies(1): >>bmelto+j8
◧◩
12. dolins+N5[view] [source] [discussion] 2011-04-03 22:07:22
>>Tycho+U3
I downvoted you because I have found this to not be the case in practice, and I find that setting some minimal comment length is not correlated to increasing the quality of submission. Casual snideness will still occur, and one-liners are not a problem in and of themselves. The context of that one-liner is, and a minimum comment length should not considered a judge of quality or context.
replies(1): >>Tycho+k6
◧◩◪
13. jerome+W5[view] [source] [discussion] 2011-04-03 22:08:54
>>tptace+85
I was thinking the same thing which is why I didn't mention the length idea. However... what if users over say 1 year and with a high comment average were exempted?

To add to that, lionhearted posted a comment about inevitable site decline because of open membership and equally weighted voting. What if only accounts over 1 year could vote (or diminish their vote weight)?

Wouldn't instituting a min. length requirement and taking away new user voting improve the problem significantly?

replies(1): >>Tycho+97
◧◩
14. rbanff+d6[view] [source] [discussion] 2011-04-03 22:13:03
>>Tycho+U3
All the times I wrote reprehensible, un-HN things, I did not strive for conciseness. Quite the contrary - if someone can cause disruption with 10 words, 50 are a nuke ;-)
◧◩◪
15. Tycho+k6[view] [source] [discussion] 2011-04-03 22:14:42
>>dolins+N5
Have you seen this actually implemented somewhere? Note that I'm not contending that none of the short comments are valuable: far from it, sometimes they're the best thing in the thread. But having a limit would not necessarily deter that content from appearing. The authors would just share a bit more info/insight with us before hitting 'reply.' Meanwhile the timewasters really would be deterred. And while it's great to have sharp, witty remarks now and again, the problem is that loads of unfunny jokes appear, trying to emulate.

However, I suppose you could argue that valuable content would be blocked to some extent, it a knowledgeable contributor was put off because s/he didn't have time for a longer comment. Still, 50 words isn't much, or 30, 25... anything above the average sentence length I suppose would start to have the desired effect.

replies(2): >>Mz+S6 >>dolins+98
◧◩◪◨
16. Mz+S6[view] [source] [discussion] 2011-04-03 22:23:31
>>Tycho+k6
I used to actively participate somewhere with a 10 character minimum. It caused people to add remarks like "10 chars" after a perfectly valid, meaningful short reply to get it past the filter. It also caused savvier members to come up with creative ways to fool the computer into accepting their shorter post without making it appear longer to readers, which caused less savvy members to wonder out loud if some folks had special posting privileges.
◧◩◪◨
17. Tycho+97[view] [source] [discussion] 2011-04-03 22:26:27
>>jerome+W5
But there is something unsavoury about that sort of elitist/privilige system I think. Perhaps, however, the word-count restriction could apply only to the 'top-nodes' of the comment threads, ie. if you're replying to the submission itself. That way, things are started off on the right foot, and snidey one-liners never get upvoted to the top of the whole thread.
replies(1): >>jerome+R7
◧◩◪
18. jerome+v7[view] [source] [discussion] 2011-04-03 22:33:10
>>jnovek+O4
And we could make it policy to immediately downvote such cases.
replies(1): >>gridsp+1d
◧◩◪◨⬒
19. jerome+R7[view] [source] [discussion] 2011-04-03 22:37:29
>>Tycho+97
I know what you mean, but is it really elitist if it's only a time requirement? It's like admission to a club, anyone can get in but everyone past the current site size pays their dues (in this case by sticking around). If the goal of the site is to not die from popularity, this may be understandably necessary.

On your other point, I'm thinking that short unwanted quips can rear their ugly head anywhere along threads.

replies(1): >>hugh3+C8
◧◩◪◨
20. dolins+98[view] [source] [discussion] 2011-04-03 22:41:23
>>Tycho+k6
We implemented this on zootoo.com during a contest where posting rewarded the user with points (I understand 100% how ridiculous this is as a concept, how awful it is at achieving any kind of 'quality' by itself - and if I was the one who had been allowed to make the decision on this then it never would have been implemented - but karma is a kind of never-ending contest). Not only did the minimum length limit not work DURING the contest, but even after the contest was over and we had kept the comment length requirement in place, the quality of comments did not increase, and if someone wanted to say something that was below the minimum threshold they would simply add 'foobar-like' content to their post to meet the threshold.

The problem with time wasters is that it's no big deal for them, they're already wasting their time. It's always going to be up to users who care like you, me and others to try to mitigate the exposure of those time wasting comments while not penalizing legitimate contributions by users who care, and that's why I don't think a hard comment length should be implemented.

replies(1): >>Tycho+u9
◧◩◪◨
21. bmelto+j8[view] [source] [discussion] 2011-04-03 22:44:28
>>Tycho+F5
Forgive me, I didn't mean to imply that your suggestion doesn't have merit. It does.

I just think that it's far too easy to game.

Another big issue with comment quality, of course, is in the general vagueness of it all. Jokes, for example, are almost always discouraged. Almost. On some occasions, it's quite appropriate.

◧◩◪◨⬒⬓
22. hugh3+C8[view] [source] [discussion] 2011-04-03 22:52:05
>>jerome+R7
The thing about elite prestigious clubs is that a large portion of the discussion and activity inside the club turns out to be about membership -- who wants it, to whom we should give it, who is unworthy of it, how worthiness should be decided, and how awesome we folks inside the club are compared to all those folks who aren't inside the club.

(Or at least, that's what I assume it's like -- I don't know, I've never been invited into any elite prestigious clubs...)

Even if you say "Oh, it's just a time requirement", what are you gonna do if Big Name Smart Person shows up wanting to comment? Surely you're going to let them in. So now you have an elite line and an unwashed line and you're arguing about who should be in the elite line.

replies(1): >>jerome+A9
◧◩
23. bvi+R8[view] [source] [discussion] 2011-04-03 22:59:59
>>mixmax+c4
One way around this could be through badges (a la Stack Overflow). You could automatically get badges based on average comment karma and so on, instead of tying your contributions to the HN community to a mere number.

That way, you're rewarded when you contribute meaningful, valuable comments, plus from a reader's point of view, your comment will still have 10+ votes.

replies(3): >>mixmax+Z8 >>noahc+9e >>pg+rj
◧◩◪
24. mixmax+Z8[view] [source] [discussion] 2011-04-03 23:02:52
>>bvi+R8
That's actually a great idea!
◧◩
25. nyelli+r9[view] [source] [discussion] 2011-04-03 23:11:36
>>tptace+P3
My initial reaction was to disagree with your suggestion: Your method would train people to predict how well a comment is going to be received and to vote based on that. I disagreed, because I don't want to read comments rated by the hive mind, in order to please the hive mind.

However, at a second glance, your idea could work. If people see diverse and insightful comments being voted up, perhaps the hive could learn to encourage creative and interesting comments.

Then again, this would incentivize upvoting comments with an existing positive score, and vice versa.

Perhaps the solution is to not display comment scores at all until you vote on a comment. (But order comments the same way they are ordered now.)

replies(1): >>Elbert+7d
◧◩◪◨⬒
26. Tycho+u9[view] [source] [discussion] 2011-04-03 23:12:51
>>dolins+98
To be honest, you're probably right. Unless comment padding was universally down voted, there'd be lots of it going on, defeating the point. And sometimes valuable things like corrections are at loggerheads with a minimum word count.
◧◩◪◨⬒⬓⬔
27. jerome+A9[view] [source] [discussion] 2011-04-03 23:17:26
>>hugh3+C8
The restriction isn't for commenting, only voting. "Big Name" would have to stick around 6 mo. to a year to vote. Shouldn't be a problem if site quality (the key draw) remains high.
◧◩
28. strlen+za[view] [source] [discussion] 2011-04-03 23:37:24
>>mixmax+c4
> The problem is only partially too many bad comments, it's also too few good comments. Your proposal is targeted at lowering the amount of bad comments, but it might do so at the expense of the good ones.

Very interesting point about not discouraging high value contributions: you don't want to create a situation where users have no desire (or are afraid to) comment.

> Most of the users that have a high karma count have so because they always have something insightful to say (mechanical_fish (http://news.ycombinator.com/user?id=mechanical_fish) is a good example), and their incentive for posting interesting stuff is marginalized by this. Almost all people, whethere they'll admit it or not, are incentivized by other peoples approval, eg. karma, and downplaying their contributions will make them more prone to not submitting great comments.

I think we can all agree that users with high karma generally have it for a reason. However, you can't always deduce that if user A has higher karma than user B, than user A has higher karma because his contributions are always more insightful than user B's. It's not a total order.

The marginal incentive ("get more karma by writing a comment that gather N+1 up votes") is not aligned with the over all goal of the sight (produce consistently insightful content).

◧◩
29. cabala+Na[view] [source] [discussion] 2011-04-03 23:44:48
>>Tycho+U3
> Minimum comment length of ~50 words.

This would backfire because many useful comments aren't that long, and therefore people will be forced to pad them, this decreasing the s/n ratio.

For example, my paragraph above has 24 words.

◧◩
30. strlen+ec[view] [source] [discussion] 2011-04-04 00:19:34
>>_delir+s5
I like this idea. As another commenter pointed out, we seek out external validation (whether we like it or not). However, there's good and bad kinds of external validation: looking for popularity (total number of upvotes) vs. more direct feedback from people you respect (Kuro5hin model: a score with a name attached to it) works better. The fact that a comment would be rated rather than voted on Kuro5hin (with a maximum rating) also has the capping effect.
◧◩
31. strlen+Ec[view] [source] [discussion] 2011-04-04 00:30:25
>>tptace+P3
> The blub setter/getter comment really is the archetypical "dumb comment", isn't it? Nicely done.

It's actually based on a real comment I saw: the discussion was about migrating a 10 MLOC (iirc) enterprise Java system to git. One comment said that this system must have been "5MM getters, 5MM setters". That struck me as particularly mean and below the belt strike against the programmers who worked on this system: it's very likely there is a good reason why it had to be in Java in the first place (and other JVM languages may not have been available when it was created) and even so, it didn't mean the programmers working on it would have chosen Java as the language themselves (but they were not there when the architectural decision was made). Further more, it added nothing to discussion.

This is not unlike poking fun of somebody for wearing the wrong kind of clothes on the school yard: cheap way to score social points with the plurality of others present, mean and ignorant (may be they can't afford the right kind of clothes, may be they are going hiking right after class).

An insightful comment would have been something like "That's great that you were able to get this into Git, changing a VCS is a painful task. Have you considered using Scala in some of the modules? Functional objects, case classes and implicits could help you model your business domain better, write thread safe code, and get rid of much of the boiler plate."

◧◩◪◨
32. gridsp+1d[view] [source] [discussion] 2011-04-04 00:39:54
>>jerome+v7
Sorry, your comment is too short (as is mine)
replies(1): >>gridsp+fj
◧◩◪
33. Elbert+7d[view] [source] [discussion] 2011-04-04 00:41:14
>>nyelli+r9
Reddit does this pretty well, they often feature a submission with a dot instead of the number of votes. Perhaps HN could randomly put a thread at the top without showing votes.
replies(1): >>cmontg+IU1
◧◩◪
34. noahc+9e[view] [source] [discussion] 2011-04-04 01:06:20
>>bvi+R8
I like this idea as well and I think it supports how a digital community actually works. For example, some people are great at finding awesome, relevant, high quality articles to submit. They should be encouraged to do so by earning a particular badge.

Others are better at commenting, and maybe there are even different classifications of commenting or different badges.

You would be able to incentivize behavior (commenting) and then narrow it down to a particular type of commenting.

One interesting badge might be the efficiency badge, and that would be generating the most points or badges per minute on the site.

◧◩◪
35. agscal+6h[view] [source] [discussion] 2011-04-04 01:58:21
>>tptace+85
The minimum length of your comment is inversely proportional to the amount of karma you have
36. crassh+gh[view] [source] 2011-04-04 02:01:44
>>strlen+(OP)
That's not to say all comment score data should be gone.

I already like what HN does with the colour of posts. What if HN switched to colour only? Displaying points

* encourages "karma whoring" by posters or just makes people care about how well their comment was received

* biases the decision to upvote/not by subsequent readers. Subconscious, conscious, contrarian, consensus-seeking, whatever -- shouldn't people make up their own minds as to whether a comment was good or not, on its merits rather than on what the rest of the community thought?

The information filtering function would still be accomplished by colour and order only.

◧◩
37. crassh+Di[view] [source] [discussion] 2011-04-04 02:28:40
>>Tycho+U3
I generally try to keep my comments short. Verbosity isn't good per se.
◧◩
38. crassh+Fi[view] [source] [discussion] 2011-04-04 02:30:07
>>_delir+s5
Trimean or winsorized mean could be useful in this case.
◧◩
39. ww520+0j[view] [source] [discussion] 2011-04-04 02:37:39
>>Tycho+U3
Length of a comment is not an indication of insightfulness. Wall of text sometime is a turn-off. People who are busy tend to just leave bare minimum to convey their point.
◧◩◪◨⬒
40. gridsp+fj[view] [source] [discussion] 2011-04-04 02:43:17
>>gridsp+1d
It's telling that this comment got more upvotes than any of my other (much more constructive) comments in this thread
replies(1): >>spin+ao
◧◩◪
41. pg+rj[view] [source] [discussion] 2011-04-04 02:49:09
>>bvi+R8
I tried badges based on average comment score (the infamous orange dot experiment) and everyone hated it, including me, because it divided the community into haves and have-nots.
replies(4): >>mixmax+pk >>alanfa+Ll >>bootlo+vp >>jdavid+zF8
◧◩◪◨
42. mixmax+pk[view] [source] [discussion] 2011-04-04 03:09:33
>>pg+rj
If the badges are only shown on a users profile page I think you'll see a very different reaction.
replies(2): >>pg+Xp >>bvi+ay
◧◩◪◨
43. alanfa+Ll[view] [source] [discussion] 2011-04-04 03:36:12
>>pg+rj
That's when you start selling orange dots. And then orange dot removers.

http://www.youtube.com/watch?v=v3yJomUhs0g

◧◩◪◨⬒⬓
44. spin+ao[view] [source] [discussion] 2011-04-04 04:38:54
>>gridsp+fj
I think that gridspy's comment is worthy of some upvotes. He made a valid point, and in a meta-joke kind of way. I'm perfectly happy with flippant posts, so long as they're not mean, and they add something to the discussion (be it humor or insight or both).
replies(1): >>MrMan+SJ
◧◩◪◨
45. bootlo+vp[view] [source] [discussion] 2011-04-04 05:17:31
>>pg+rj
"... because it divided the community into haves and have-nots. ..."

Isn't this is a logical inconsistency? There seems to be no problem dividing the community into haves and have-nots when it comes to post quality.

◧◩◪◨⬒
46. pg+Xp[view] [source] [discussion] 2011-04-04 05:25:05
>>mixmax+pk
Hmm, maybe you're right.
◧◩◪◨⬒
47. bvi+ay[view] [source] [discussion] 2011-04-04 10:37:17
>>mixmax+pk
Exactly what I had in mind as well. I'd prefer if the current UI stays as it is, but if I click on a user's profile (or hover over his username), it would be neat to see the type of badges earned.
◧◩◪◨⬒⬓⬔
48. MrMan+SJ[view] [source] [discussion] 2011-04-04 15:06:22
>>spin+ao
Pithy, incisive remarks are to be celebrated, not supressed.
◧◩◪◨
49. cmontg+IU1[view] [source] [discussion] 2011-04-05 15:20:44
>>Elbert+7d
I believe Reddit does this for the first X hours of a submissions life - to discourage bandwagons, and so that newer submissions that may otherwise be overlooked (due to having less initial upvotes) will get a chance. I agree that it's a good idea and something that may work well on HN's new page.
◧◩
50. run4yo+nW7[view] [source] [discussion] 2011-04-11 21:15:57
>>Tycho+U3
There is no point in saying 100 words when you can make the point in 10.
◧◩◪◨
51. jdavid+zF8[view] [source] [discussion] 2011-04-12 14:08:21
>>pg+rj
i know many of us Use HN Plugins for chrome. For posters that I like, I follow them and their comments are displayed in a different color.

It's not quite badges, and it is a bit twitteresque but i really feel like it ads value to the site for me.

https://chrome.google.com/webstore/detail/mkdhfabjcebcgnpgnh...

◧◩
52. dpk+U59[view] [source] [discussion] 2011-04-12 19:46:31
>>tptace+P3
I see that I'm replying to a rather old thread, so apologies in advance if I'm doing it "wrong". (It doesn't appear that is specifically outlawed in the welcome page but that doesn't mean it's kosher.)

This is similar to an idea I was toying with a while back but never got around to nailing down. Upvoting or downvoting an item should result in a change on the personal account, but rather than it changing the account's "score" up or down, the system records the vote based on the "type" of item. If a series of items are categorized as "Gossip" and I vote them down, the system learns that I don't like "Gossip" items. An item could be in multiple categories ("Gossip" and "IT") and my past voting would determine whether or not I would see the item on the page (super-roughly "the item was categorized 50/50 Gossip/IT, dpk's Gossip score is -11 and IT score is 10, result is -1, don't show"). In effect, users would themselves be group-able by their votes, so if someone in your "group" posted an item, it would get an automatic bonus. If someone in your "anti-group" (someone nearly diametrically opposed to you) posted something it would get a negative bonus.

Categorization would need to be put to the community, and would be done while the item is on what is currently termed the "new" page. Once the item is categorized the various display scores (as loosely described above) will be computed and the result could be shown to the users.

One side-effect of this is that it allows users to "shun" spammers in to their own "group". They could spam all they want but nobody would see it unless they were really excited about seeing spam.

The idea has (at least) one major serious problem: It encourages group-isolation. This could be partially resolved by always showing a "best of group" block of links somewhere on the home page, which could encourage users to branch out a bit.

As I said, this idea is not fully fleshed out, hence the overuse of quotes and its nebulous, hand-wavy nature. It's a less punitive, more categorical system. It may not scale at all.

[go to top]