Past polls show that a substantial portion of readers (me included) think the points are very useful. They draw our attention to the best comments.
Clearly, you think this reduced usefulness is worth it to reduce fights; fair enough.
So, is there any way of getting back the utility, without increasing the fights?
Maybe show points, but with a time lag?
Show points, but just in bins (maybe logarithmicly binned?)
You trust users above a certain karma threshold with downvote, because they've learned the ropes; maybe display points to these users? People that get in fights, and hence are downvoted, won't see points?
Beware of selection bias. I saw these submissions when they were posted, but voted in neither, because I didn't mind the status quo of no points, and I'm wary of meta-discussion. If I had felt differently, I'd definitely have voted.
So, I actually think dlss is making a very valuable point here: the argument that many people are making with relation to hiding point totals, is that if you can see the number of points a post has, it may affect whether you decide to vote at all: you may go "oh, that's already high enough, I won't bother voting".
In essence, people have a mental concept of how high or low they feel posts they are looking at are, and adjust, possibly very subtly, possibly without even realizing it in a way that they would explain if asked, their behavior to cause the post to hit that target, as opposed to showing their true interest with an up/down.
(For a reference on users discussing this effect, and some commentary from PG about having seen this effect on the high end of comment karma scores, see this example thread from another post: http://news.ycombinator.com/item?id=2465271 .)
How peoples' behaviors change is complex. On the one hand, you may see "your side" heavily lower than "the opposition", and decide "ugh, it's pointless". On the other, you may go "ok, that's just not fair", or "help out the team", and make extra certain to vote on that specific poll.
Due to these attitudes, which are pretty much the same things that are driving the changes in behavior with respect to comment voting, you don't get accurate information with public live polls. This really is the same effect, and should be thought of and treated the same.
Now, making the entire poll /permanently/ only visible to the poster would mean, as you say, that the poll becomes meaningless. However, I would say that's a strawman: there are numerous other ways to handle the situation that still make sense.
Example: polls could be private until they "time out", at which point the poll is no longer votable on, and the data becomes public. Frankly, this seems correct, and I'm glad dlss brought it up (although apparently people disagree with the way he worded it or something, which seems silly).
(I also think that this behavior would dovetail very well with "show the comment scores after people can no longer vote on them", which came up elsewhere in the comments for this post: http://news.ycombinator.com/item?id=3122362 .)
It's an analogy. Without publicly viewable karma, posts lose a lot of their information content.
Polls without results lose more, but some would argue[1] that there is no royal road to understanding what others think, and that the best way to use polls is to simply read the possible responses and ignore the results.
For example, [1] believes that if you saw a poll on "which database do you use?", you should ignore the poll results and researching each option on your own. The person he is replying to is advocating researching databases in HN popularity order.
Social proofs are bad. You should be reading the comments and then deciding what is interesting and noteworthy, not looking at comment scores and then reading and then "deciding" after your opinion has already been influenced by the group-mentality of popular opinion.
> Clearly, you think this reduced usefulness is worth it to reduce fights; fair enough.
I don't think that's a fair comparison, see above.
> So, is there any way of getting back the utility, without increasing the fights?
This is not the utility you are looking for. waves hands
In all seriousness, this kind of short cut seems nice on the surface but in reality is a bad habit: social proofs are misleading and bad, e.g. leading to groupthink. Hiding comment points was, in my view, a very good move. I certainly noticed an improvement after it happened and I think it's something that is needed in order to maintain a richness of quality in the comment threads.
>Social proofs are bad. You should be reading the comments and then deciding what is interesting and noteworthy [...]
Respectfully, I disagree.
Its a core idea of this site, that the community does a good job of surfacing interesting interesting content; what appears on the front page is what has been upvoted sufficiently. I presume you wouldn't argue that HN users should read all submissions, and decide what is interesting and noteworthy for themselves.
You have a point, in that there may be feedback effect with comments.
Personally, I don't think that a comment being highly rated seriously effects my judgement of its content. It does, however, draw my attention to that comment.
I would argue that comment scores should not be optimised to give the most accurate rating to comments, so much as to filter comments - e.g. draw attention to comments that are interesting, and might otherwise be overlooked.
As such, if there is positive feedback in good comments being 'overvoted' I don't see a problem.
I accept that this is a subjective perspective, determined by what I want from HN. I want to use the smart community here to quickly learn important information, and gain insight; as such, I don't really care whether comment scores are 'correct', so much as whether they are useful.
As others have said, the ability to see the wider community's aggregated opinion on a particular set of conflicting comments, is also useful, where I lack the expertise to evaluate them myself - but I wouldn't be casting any votes, in such a situation.
> It does, however, draw my attention to that comment.
There is a subtle feedback loop in there. After all, a comment that draws your attention will then be more likely to be upvoted by you and people acting like you.
I'd appreciate having the bins that you mentioned. My issue is that often times I see a parent comment being corrected by a child comment. However, due to no point values, I often dont know who is correct and who isnt. Here's an example:
http://files.jjcm.org/example1.png - Here we see a comment as it is now on HN. At the moment, I dont know if my page has a legitimate problem or not.
This is an issue that I find I encounter time and time again on HN. While comments are sorted by best score at each level, between indentation levels there is no way of discerning which is more highly ranked. I think having a system such as feral's logarithmic bins would be a nice middle ground here.
>Past polls show that a substantial portion of readers (me included) think the points are very useful. They draw our attention to the best comments.
That is not the primary reason that I want points back. I believe that allowing people to signal their opinion with public upvotes discouraged redundant posts, and that visible upvotes encouraged participants to let their arguments stand rather than creating unnecessarily deep threads.
A secondary issue is that votes were a valuable part of the signal in many "Ask HN" threads.
Just to note, they don't necessarily draw your attention to the best comments, just the ones that have been visible the longest. If you come across a post with 800 comments and the top post has 300 upvotes, how much less likely are you going to be to post a comment or thought? Removing the score both makes it less intimidating to post, and also increases the likelihood that people are going to find your post-- each user has to trawl through the top level comments on their own to figure out what's good, as opposed to stopping as soon as they hit the 1- or 2- upvoted comments at the bottom of the page.
I can categorically state that the removal of publicly-viewable numbers has not made me more likely to read the 2-liner comments by latecomers, and I don't think I'm alone in this.
I don't think those polls captured intensity, either. After several months to get used to no points, I feel very strongly they should be brought back. These threads are much less useful without them, especially when I don't have time to read/judge every single comment, or am not familiar with the discussion topic.
E.g., if there were points in this thread, I'd have a much better idea what the 99% of non-commenters thought about points...
+1 for the time delay idea. Once a post has been up for a certain amount of time, points are displayed so we can sort out which comments to immediately go to. No points makes HN more time consuming for me and less interesting. If there is a fight due to points, I'm just not going to read it.
In general, useful is a good thing. A tool that washes your clothes for you is useful, and good. A tool that hammers nails for you is useful, and also good.
Now, a tool that judges what you are reading is useful, but is it good?
The points don't judge if what I'm reading is useful, they merely indicate how useful others found it. This helps give me context in situations where I am a) inexperienced or b) short on time.
I'm curious, what has happened to the average time spent on comment-pages since removing the points?
I personally don't care about fights, they're usually confined to a single nested thread which is easy to skip right over. What I can't stand is not being able to judge consensus on topics in which I'm inexperienced. In that regard Hacker News has gone from being my main resource for diving into new topics, to being near the bottom of my list of resources.
Removing points hasn't merely reduced HN's usefulness for me, it's put it in an entirely different category. This used to be a place for me to skim and read an entire thread, top-to-bottom, and provide insight where appropriate. Now, if anything, I read the first couple comments and move on, because I just don't have the time.
I have to totally second this. Reading this comment made me realize I got the same behavior too. Even worse, sometimes I avoid reading comments altogether, because to me they lost almost all the value they had when displaying votes.
I am really curious to know how the number of fights is measured, though. Is there some sort of sentiment analysis on the comments, comment statistics or it's just PG feeling?
What you were doing before was a bad idea. HN talks about a lot of things I don't know much about --- Haskell, high frequency trading, scaling Cassandra and Riak, AI search. But it also talks about some things I know very well. And it was not my experience that comment scores generally tracked correctness on those threads. Generally, they track what people want to hear, or what the "cool kids" think.
You misinterpreted my comment. I wasn't looking to comment scores as a measure of "correctness". Hell, I can find out correctness on my own by actually researching the topic. I used comment-scores as a measure of consensus. That is, after all, exactly what it is.
Simple example:
> Ask HN: What's the best way to perf-test my server?
> comment X: I use httperf, it has great features and is very robust.
> comment Y: Check out Apache Benchmark (ab).
There is a huge difference between whether comment X is marginally higher than comment Y (i.e. it has slightly more upvotes), compared to if it's a factor of 2-5 or more. In the former scenario, I know to spend an equal amount of time evaluating both, while for the latter scenario, I know one project has much greater user-adoption among the HN community than the other (or at least more passionate adoption).
Likewise, if a comment thread has a question that has 50 upvotes, then many more people seem to be interested in that particular facet than another question with 3 upvotes.
My point is that comment points provide an additional form of context. I actually want to know "what the cool kids think", because that in itself is useful info. In most cases, I take it with a grain of salt, but that's really up to me. If other people misinterpreting comment points, or abusing them, means I can't have them... well, that's a shame. HN will continue to be the fraction of its former self for me.
The problem with your logic is the sample population. What is the "HN Community", if we define it as "the set of people who click a vote button within a given topic"? It's a nonuniform cross-section of God Knows Who Is Using HN At This Hour.
It isn't even a representative sample of "what the cool kids think". Even cool kids get bored with repetition. For example, once upon a time almost every XKCD got linked on the HN home page, but due to tacit agreement (and, perhaps, intervention by Our Invisible Friends The Moderators) this almost never happens anymore. And yet everyone I know continues to routinely read XKCD. It's not uncool or anything. It's just no longer new.
Similarly, Lisp and Erlang make far fewer guest appearances on the HN homepage. Should I conclude that these things are now uncool? No, just that time has marched on, and HN has grown and changed, and the makeup of its increasingly gigantic audience has shifted.
If you want to take a poll you should try to find someone who takes polling seriously. Otherwise you're better off with un-numerated anecdotes. At least an anecdote doesn't masquerade as data.
Your comment is missing the point. I know about sampling bias, and I'm aware that the points are representative of the subset of people who read that particular comment (or didn't) and decided to click a button. These are all facts that go into my interpretation of the data. But it's still data!
And the data was useful to me, in my own interpretation. Why is everyone so determined to convince me that comment points weren't useful to me? You're not qualified to determine that, because you are not me. And I'm telling you, they were useful to me. I can't make it any clearer.
Frankly, how I interpret the data is none of your concern. If you're afraid that other people will misinterpret the points and assign some false value to them, I can't do anything about that. It's simply a shame that the solution is to take away data from everyone, because you're afraid some will misinterpret it.
But hey, if making HN less useful for me makes it more useful for everyone else, then I guess you can disregard everything I've said.
Why bother having a voting system at all, then? Karma that can only be seen on the user's page is simply just point-scoring. Sure, you gain minor functionality increases with karma, but you could do exactly the same thing with comment count for comments that aren't flagged (since you need to comment to get karma anyway)
Just give people a way to flag inappropriate comments, and then you've got pretty much the same system as we currently have. Early comments stay up the top, and there is no way to tell which has engaged the community apart from volume of response. Inappropriate comments get flagged and with enough flags they die. Get enough unflagged comments under your belt, and you're 'participating' and are then allowed the escalated priveleges.
I second this. At first I was against the hiding of points for the reasons others have stated: to show good comments.
In reality you would get situations where someone would post a thoughtful and well-reasoned comment (2 points) and someone would reply with a very minor correction (12 points).
It also seemed like "wars" were fought over points too, purely anecdotally.
My own experience is that voting now seems to be more rational. People seem to upvote when something should be upvoted (my guess is that this topped out before if someone thought a comment had too many votes). When something is negative, it's indicated so it doesn't tend to get downvoted into oblivion.
As for curating interesting comments to the top, this happens now and seems to work quite well.
I just wish there was a way to:
- Look at comments you haven't seen yet; and
- Filter comments by "quality", sorta like Slashdot has. The meaning of this can be vague and "quality" doesn't have to be a straight point measure but I'd like to see a way to just see the best comments, particularly when there are 300+ comments.
Also, collapsing comment trees would be great so as to avoid having to scroll past dozens of comments in a long-winded discussion of something only tangentially-related to the submission.
(See I have to do this, otherwise there is no way to indicate assent. For PG to know this is popular, if he just looks at the regular site page, he would need a huge string of such comments.)
Perfect, thanks. It'd still be nice if it was a site feature so I wouldn't have to install that script on every computer I browse HN from, but it's a great band-aid solution. Thanks again.
As an alternative, I think it would be a good idea if some people could detach such branches, moving the entire branch into a separate topic.
I would leave the top comment both as introduction for the new thread and at its original location, with a link to the new topic.
I also think it might be possible to somewhat automatically detect such branches. That would make it easier for the 'some' that get this power to detect them.
However, if one has such software, it might be better and easier to 100% automatically collapse all candidate threads with a generated remark "this appears to start a new topic", ideally with keywords describing what the topic is about)
To better illustrate this effect, let me present an example. When I'm reading a comment and it makes an initial impression on me (good or bad), I have to decide whether I'm going to vote on it (if I'm unsure, I'll just leave it unmoderated). When I see the score, it gives me more confidence in my opinion so I vote more.
If others fall victim to this similar pattern of behavior we'll wild variation in comment votes and we'll see scale free behavior of votes (i.e. preferential attachment, comments with higher votes will tend to get more votes).
According to PG (I can't find the exact comment right now) the exact opposite has happened - posts with a lot of points get more points, not less.
The reason probably being that when you look at a commentscore you assess whether it has the right score, and if it doesn't you vote it up (or down). If you can't see the scores you can't do this.
Generally the community here was pretty good at voting based on worth of the comment in furthering the conversation and not on the opinion expressed therein.
Why vote? No one but you knows you did. The comment owner gets a number showing only assent or dissent which usually is irrelevant to their purpose for commenting.
I'd like to also add that it would be nice to get a show-comment-at-full-opacity-on-mouseover for downvoted comments, like SomethingAwful's Forum Cancer. As it stands, having to highlight someone's comment just to see what they're getting blasted for (and my system highlight color is a light green, which makes it even more fun) seems like poor UI design.
I actually wish reddit would get rid of displaying karma; gaining karma has just become a game and is honestly causing more problems over there than not.
I honestly hope that this happens. I got a Chrome extension that displays the number of times I've up/downvoted the comments of a given user wherever their name is displayed, and I never realized how many posts on the front page are posted by the same people.
>> As for curating interesting comments to the top, this happens now and seems to work quite well.
The comments still form an N-ary tree. Useful top-level comments also take not-so-useful comments on them to the top. With the points visible, I could tell just by looking. Given the lack of time, I now find myself reading just the top-level (not top-on-the-page) comments.
The problem is that a simple up/down vote system is too crude.
- low effort for drive-by downmods (there are stupid numbers of comments blasted to nothing despite being insightful or informative)
- upvote doesn't distinguish between 'I agree' and 'hey, this is really informative'; the former is mostly meaningless, the latter is useful
- single downmod diminishes what you said and makes it harder to read (on an already low-contrast site). Hardly fair given the ease of 'drive-bys'.
The forum I spend most of my time on has a reason required for the mod - it's not this brainless +/- system that the web is moving towards. You get an idea why people are up/downmodding, and it raises the bar ever so slightly so that it takes actual effort to mod (that is, critique) a comment. The displayed point value is also capped - while the user gets the karma for the mod, it's not displayed beyond a certain cap.
Mostly I see the benefit of this system as stopping mindless up/downvoting - requiring the tiniest bit of effort produces better quality results. Still not perfect, but much, much better.
The up/down voting system is indeed crude -- but it's both simple and relatively well understood (and popular).
Also, what you characterize as "drive by voting", I think of as being "lightweight feedback". Sure, we could make things a wee bit harder so a (much) smaller fraction of people would vote with with more deliberation, but I'm not sure that actually improves things.
Seriously though, this is a problem I think a lot about. One solution that keeps popping up in my head is implemented in an OS forum somewhere (don't recall specifics), so that each comment can be marked as informative/funny/shocking/etc instead of mere upvotes.
Downvotes should go along similar lines so downmodded people don't have to beg for reasons. If "it makes boring reading" [1], don't make people write it!
Requesting a reason for a mod via an associated dropdown is not hard for anyone to understand - and I would expect users of "Hacker News" to be able to figure it out even if it were slightly difficult.
The "lightweight feedback" is "throwaway feedback", because you really only see how unpopular a comment is (is it dead yet? oop! there it goes!). Only the owner knows how popular it is, and even then, doesn't know if it's just a bunch of 'yeah, me too!' or 'that was informative or insightful!'
The most egregious example of 'why is this modded so' is this comment, so much so that I've stuck it in my profile:
The thread was about RMS and what an eccentric he is, and about his comments subsequent to Steve Jobs' death. The comment is defending RMS using an oft-quoted verse of Jobs', where he defends mavericks making their own rules and violating the status quo. The comment makes a very insightful link, and is not offensive or malicious. Yet it's downmodded to dead status, something that other comments in the same thread didn't have happen despite being much more fractious. Why was it downmodded? Who can say? It's probably fanboys, nothing else makes sense from a mature point of view.
Which reminds me of another issue with the mod system: it's totally anonymous. There is no accountability. You can diminish someone else's statement without giving a reason why nor standing behind it with your own name/pseudonym. Again, the forum I come from gives you the option to see who modded a comment whichever way - which while there is the occasional problem, is overall a significant net gain for an online community that strives to keep things mature.
Can we at least have it so that the arrows don't disappear when you click on them? I can't tell you the number of times I've clicked on the wrong arrow and been frustrated by not being able to correct my mistake. This is especially bad on a small touchscreen (i.e. phone).
>People seem to upvote when something should be upvoted //
How do you know?
All I see is some fading, which presumably means a comment has been heavily downvoted? But for all I know it could mean the commenter is new, or they're generally downvoted, or something else (perhaps it's cold in their city?).
I agree that it has improved the civility of discourse.
Have you considered allowing users to turn karma display back on once they have reached a set amount of karma, much like other features (such as downvoting)? I imagine that, as long as the threshold is high enough, users would have learned how to be a benefit to the community by that point and, in addition, it might remove some of the psychological tendency to "score points off of someone" if only a subset of users can see those points.
I found myself to be fairly susceptible to reactions based on other people's karma scores; I don't see why it's a good idea to encourage an "elite" of HN users to write dumb comments.
It would be pretty easy to add a feature such that once you've seen the karma scores of comments in an article, you can no longer comment on them. That seems like a fair trade to me.
What about alt accounts? Sure, you can try to track that via IP, but by the time you start thinking about that, even this "pretty easy" feature starts being real complex and prone to problems (like misidentification behind NATs, etc).
I, too, found myself susceptible to reactions based on karma score and, personally, it wouldn't bother me at all if the scores never come back. If given the ability to view karma scores, I would leave them off, since I'm self-aware enough to realize that -- when karma scores are enabled -- my comments are of a lower caliber.
However, I'm not against allowing users with a certain level of karma the choice to view karma scores on threads if they really feel that HN is better with karma on, which it seems like a fairly large subset of users (or at least a significantly vocal one) wish for. I think that StackOverflow's successful moderation policies (and, to a lesser extent, HNs own) have sufficiently demonstrated that extraordinary members of a community can handle the responsibilities that come with increased privilege.
For example, I suspect that, if pg allowed individuals with high karma, such as yourself, to choose to toggle an option such as "view karma scores", no flood of dumb comments from high karma individuals would suddenly manifest. Further, I think that the transition from default karma displayed to our current situation has allowed anyone who has invested significant time in the community an opportunity for introspection that wasn't previously available. I'm now able to evaluate, personally, how my comments and behavior were effected by the display of karma and I'm less susceptible to the psychological effects as a result.
Of course, even if there isn't a significant negative aspect to allowing high karma individuals to view karma scores, it's still hardly worthwhile unless there's some kind of positive impact as a result. Perhaps mollifying the community would be enough of a benefit? Maybe a more pro show-karma individual would be better qualified to comment on the potential upside. My feeling is that it's possible that some subset of HNers exist whose experience with the community is improved by having karma scores enabled and, as such, it would be beneficial to allow them the possibility of a choice.
Old pages are valuable since the quality of discussion is very high here. I, for example, did a search on "NYC" and read all the discussions (pros and cons) of doing a startup in NYC vs SF and, just in general, life for a hacker in NYC vs SF. And I do searches on tools like VIM and see what everyone is recommending to do and all the good tutorials people submit. For certain topics, I prefer to search here than Google. A number of discussions on here are timeless.
I'm not arguing to show points. It's fine without the points. Just saying that old pages are valuable to me.
While this might be true "by volume", I would be very surprised if it were true "by weight". The first thing I do when I'm looking at a new library, daemon, service, company, or really almost anything at this point, is go to Hacker News to see if anyone has mentioned it in the past, and what their take on it was during that discussion.
As an example, when I got interested in GitHub last year (note: I mean as "what are they doing in this space and where are they going", not "what does it do"), I seriously went to searchyc and went through the Hacker News archives chronologically, skimming every single thread that has ever talked about it; it took days, but it was incredibly worth the time.
This has become such a resource for this sort of thing, that I've been trying to "get to know" the people I interact with here, building a tagging framework (browser extensions) that I can use to add little notes to users, allowing me to get a better feel for who they are, rather than some anonymous user: if nothing else, it helps me remember at all times that these are real people whom I am interacting with ;P.
So yeah, I guess I'm rambling now, but what I'm trying to say is: the archives are important. They may not be something you use every day, but when you do find a reason to use them it is one of the most valuable resources I currently know of, and I have to see it labeled as something that there is no point in improving due to "lack of use".
I agree that points on old stories would be useful as it helps in search results - I find myself searching for stuff on HN quite frequently. Today, I was looking into a way of publishing software contracts to clients online so they could sign them digitally and found out about RightSignature, Tractis, EchoSign and SignNow all from searching HN, and some of the stories that were coming up were from 2 years ago.
Agreed. When I'm researching a new topic, I always search hacker news for anything on point. Sometimes that's enough to find exactly what I'm looking for. The collective knowledge here is astounding.
(my first post. Thanks to everyone who contributes, it's been an incredible learning experience since I learned about hacker news 8 months ago. And I say this as a non-hacker. This site has been immensely practical.)
Statistics if nothing else. Maybe more people would look at old pages given the incentive. Maybe someone will come up with a "best of hacker news" website, with the highly rated or popular stories and a selection of the most prominent comments.
Given the demographics, I'm sure the ratings will be used by someone for something. :)
I do frequently return to old conversations when looking for advice about everything from tools to strategy. I think having the points visible would make it easier.
I understand the argument for not having points but I think I get less out of HN without them. I'm not able to quickly find gems within subcomments without actually reading thoroughly and I don't always have time for that.
I'm not able to quickly find gems within subcomments without actually reading thoroughly
I find, however, that reading subcomments thoroughly is the one and only way to know which are gems and which are just popular cheap shots. There is no royal road to learning mathematics,
and there is no way to get good information out of written material except by reading it carefully.
After edit: for readers who haven't seen all this before, here is historical background on why Hacker News changed the interface to not showing publicly the karma score of each comment about half a year ago. The site founder, pg, asked for suggestions on how to "stave off decline of HN" in a thread he opened 197 days ago.
and now today he has said that he is satisfied with the results of the experiment. He runs the site, and I defer to his judgment about how to run the site, but I must say that I agree that the site has improved since visible comment karma scores were turned off. I can still see MY OWN comment karma scores, and so there is still a reason for readers to vote--to guide the behavior of other people posting comments so that the site guidelines
With points I might have to read the odd 'popular cheap shot' when finding gems.
Without, I have to read everything, including these 'popular cheap shots'. And as many have stated (but i can't really quantify-- because i can't see how many people are voting them up!) there isn't the time or the inclination to read everyone's posts.
But you are showing that with points shown you have two tiers of users:
-Those who read every post & up/down-vote each one on its merit.
-Those who scan the page quickly & read the comments with the highest votes.
The problems is, how do you tell which kind of user is voting when comes to ranking a comment?
I would argue that a vote from someone who just scans the top rated comments is less useful than a vote from someone who has read the whole thread. The chance of gems being left languishing in a corner decreases with many eyes hunting for good content.
By being the second type of user you are essentially out-sourcing your opinion of a comment to the crowd and assuming everyone else is not doing the same thing.
All of us are using imperfect heuristics to select good content to read. Just the fact that we're reading HN instead of some other source right now is an imperfect heuristic. There's just too much potential reading for anyone to read everything equally.
That is a good point & I don't know what the perfect solution would look like, however I do think it is understandable that PG has decided to adjust the system with a view to improving comment quality.
So you think all internet comments need to be read in order to learn something?
Do you also believe that employers should spend a few days with each applicant before making a hiring decision? Would Amazon be better without product ratings? You could apply your argument to this situation and say that Amazon is better off without ratings because the best products may not get rated.
Services like Google, Reddit, Facebook, and Amazon owe much of their tremendous success to the fact that they provide users with effective screening mechanisms for the internet. We all benefit from lower information processing costs.
It seems as though they still are weighed for sorting. Isn't that enough? I think it looks fine without the points, and hopefully it will reduce the number of people that say something a certain way just to get points. I just say whatever the hell I want to most of the time, though.
That's great for comments that are direct replies to posts but when points are displayed it's much easier to notice excellent replies to mediocre comments.
The problem is that there are some rather heavy casualties that come from not showing points.
- It's very hard, if not impossible to scan a thread for good comments. Busy people don't have time to read through a whole thread to pick out the good comments.
- I did a back of the envelope calculation here: http://news.ycombinator.com/item?id=2569997 that shows that there is no way to assess the scores of around 50% of the comments. If there isn't a good way of assessing which comments are deemd good by the community is the system working properly?
- When a poster postulates something and is refuted by a sibling it's impossible to know who is right. Points give you a gauge of what HN as a community thinks. For instance, if a user comments that you should always use Bcrypt and a sibling replies that this is simply wrong I don't know what is right and wrong, even though it may be obvious to someone knowledgeable in encryption. Points give you a good header.
- When you don't show points people tend to vote less. Without having seen the numbers I'm confident that voting has decreased significantly. If noone is voting (or even worse, only a subset of opiniated users vote) then we're back to a threaded discussion ordered by randomness.
- As I remember it one of the original reasons for hiding points was to stop the "piling on" where already popular comments got more upvotes. You posted at some point that comments which had a lot of points before actually have more points after the don't-show-points change.
Hiding scores to reduce arguments seems like a very crude and inelegant way of solving the problem. You solve one problem but effectively render the scoring system useless in the process.
>> Hiding scores to reduce arguments seems like a very crude and inelegant way of solving the problem.
Instead of having just the sum of all the votes (points := up - down), it may help to keep both up and down values as they would encompass information about both usefulness of a comment as well as how controversial it is.
Having ruminated on the subject of voting I think if it were my site I'd have several options including no vote display, up+down as seperates, comment coloring, user-based comment promotion, etc..
Then I realised that slashdot was probably closest to what I was after ... then I started going back to slashdot.
Well, it's your site and you're doing what you think is best for the community.
Fair enough.
I feel pretty unhappy about it though. As something of a data-geek, knowing there's data I used to get (and that I personally found really useful) that you have chosen to withhold... I don't feel good about that at all. Especially when you did it to solve a problem that I didn't even notice.
Interesting observation. I almost never post now. I never used to fight, intensively or otherwise. Nor do I now. I spend far less time on HN now than I did when there was a mechanism to easily determine the quality of the debate.
So crippling utility in favour of civility is a fair trade?
Like others, I found it very helpful to be able to gauge the general consensus on a topic by the number of points each comment had. If the price for that is having to skip over a fight or two, I'm pretty okay with that, personally.
Utility is better served in the long run by ensuring continued civility. Reduced civility frequently results in the retirement of the best contributors to utility.
Have you read "Evaporative Cooling of Group Beliefs" by Eliezer Yudkowsky[1]? It's the same principle: when civility diminishes and hostility increases, the first members to leave aren't the ones being uncivil and hostile, but the ones who are civil and gregarious.
All you have to do is note that the majority of quality contributors here are civil and polite to see who's going to leave if the civility of discussion decreases.
I am generally civil and polite (though I may not be a quality contributor), but one serious problem I have with HN is the cognitive dissonance between "No humour - we're a mature, civil site" and "Random yahoos can make your comments unreadable, with no reason and no accountability - we're a mature, civil site".
There are several comments in this very post where the user is being civil and polite, yet their commentary is faded out. One of them is even dead.
It leads to a feeling of randomness, and the sense of popularity being more important than insightfulness. On a site dedicated to the mavericks, it's a curious mechanism to enforce toeing the popular line.
To buy in to it I'd have to see some evidence that the hypothesis is supported in groups similar to the population of active users of HN. What I see is that a lot of main contributors are bought in heavily and so are unlikely to leave. It is these users here that are like the group in the article of whom it is said:
"The challenge to their belief presents an immense cognitive dissonance; they must find reinforcing thoughts to counter the shock, and so become more fanatical."
I warrant that those who are most civil, in this context, are those who are effectively 'fanatical' about maintaining the group as a useful communication tool.
Basically the evaporative metaphor requires you to decide who the fanatics are and you can spin this in different ways according to the result that you wish to forward.
Utility is not crippled except by people who, for some reason, cannot skim the content of a post to determine whether it's worth reading without having a number next to it.
I find upvotes to be an incredibly poor indicator of comment quality. It's barely useful as a sorting metric and fails almost completely as a filter. Worse, it's information that distracts attention from the actual content of the comment.
These "fights" do you have any sort of metric that backs up your contention?
IMO people downvote on opinion a lot more. Now the only time I vote is to upvote someone who's faded out (I'm assuming that's what fading means) and who is usually making a point I disagree with but nonetheless a good point clearly made. From a UX perspective the fading of comments sucks.
Vote blindness makes the whole site of far less worth to me.
I haven't been here that long, but in the few discussions about this I've witnessed, I haven't see the suggestion to try breaking voting into agree/disagree and good/bad. Have you considered something like that before?
Could we perhaps have powers of 2 shown?? ie where points are 1,2,4,8... 1,2,3,4.. are shown.
This would still allow some way to see the relative opinion of other HN users, but hopefully with less problems than you saw with the direct points system. (personally i was ok with how it used to be).
Whilst here could i request a minimize thread button (like reddits); without it allows anyone to hijack the top comment's popularity (and overly/improperly focuses discussion around it), at the cost of the guy who properly started a new thread.
Reddit does "ninja bans" where a banned user can see his submissions, but others can't. Could you do that with arguments and long threads in general? If there's a lag of 10*2^n minutes before seeing replies in a thread n messages long, arguments will fizzle; nobody can stay specifically mad that long.
HN has that too, AFAIK. Check out the "showdead" option in your profile - it lets you see people who've been ninja-banned (aka. hellbanned). It's not the same as delaying replies, though - that seems like something that might be helpful.
A better way to do it would be to cut off responses to any thread once there are a certain number of flags in the parents.
The way it's currently set up, there's been a couple of times I've been having a valid conversation with someone, but then run into the wall of not been able to respond directly to them - no fight, no flagging, no reason given.
He's suggesting a cross between a hellban and the reply delay, where you can reply but your comment has to sit in purgatory for a while before anyone can read it. That way people think they've got in their two cents in the fight, but in reality they have to wait just as if they couldn't post.
It would presumably solve the problem of fights turning into a series of sibling replies or spreading to other threads where the participants see each other.
Out of curiosity, did you measure that quantitatively or is it just a qualitative observation? If quantitatively, what did you use to measure it? (Totally just randomly curious...)
Would you consider putting comment scores on a user's Comments page or perhaps on the permalink page to an individual comment? That way the truly curious can look?
not sure if I can share this observation. I did like the point system as it took out the noise level and I was able to get the good feed in an instance, now is it search and read over and over again like the internet.
It also decreases the level of engagement, readability of a thread, and because weighting of comments is now invisible there is no reason to upvote or downvote anything, or read any particular comment over another. It's all just one big blob of text that I have to manually sift through. Might as well just remove the arrows because they serve no purpose now.
except for the fact that in a threaded discussion where points is only one of the parameters (time and user karma being the other tow as far as I know) this isn't true for most comments. I did the numbers for a long thread some time ago here: http://news.ycombinator.com/item?id=2569997
>My points received per comment seem to be higher after that change. //
One often gets a lot of votes for a low value comment, something like a clarification. It's clearly valuable but before people would see that the vote on it was say at 2 points and wouldn't vote it up as that would be the right level for such a comment.
Comments that would have been -1 before now appear to be faded to almost nothing. I've no idea what that means WRT downvotes but presumably it means most everyone hates the opinion in that comment.
I would like to see points displayed after voting. I think they have more information content when hidden before the voting takes place, but I think it's useful information.
I've never seen "fights" over karma scores, especially over the difference between +2 vs. +20 (downvoted posts are visible by color). Usually posts about upvoting or downvoting (i.e. "Why is this crap at +20?" or "Stop downvoting me" or my favorite, "I know this will get downvoted but...") get downvoted hard and people learn that the community dislikes them.
