By the way, suppressing these metrics is pointless. We are intelligent people who notice the difference between two comments with very similar upvotes and one that far outpaces the other and find these proportions to be of interest. The new portrayal of comment rankings keeps interesting information from users and will end up stunting the community's progress. If you're just trying to ward off users, maybe remove the nav and make the site even slower.
There are far more interesting ways to differentiate. I'd like to see this site be as light and fast as possible (would appeal to our hacker ethic) and allow users to customize the hell out of it given knowledge of Arc or similar (good chance to promote your language as well).
It's true that comment-score information is interesting, but is it interesting for the right reasons? People are hardwired to have a morbid interest in the thinking of the tribe. Even if hackers consciously try to avoid paying undue attention to the opinions of the tribe, the subconscious instinct is still there.
I think that the dangers of majority-conformist thinking and other groupthink behaviors are very real and should be discouraged. I don't think hiding comment scores will have that effect, though it might make it more difficult to recognize when it happens.
The one thing in this experiment that might (I intuit) help discourage groupthink related behaviors is hiding one's own score from oneself, because people will then be less emboldened when they're being idiots just for expressing popular opinions. Unfortunately, this will also (I intuit) discourage people when they're being smart from as readily taking another look at whether what they've said is idiotic when others recognize the idiocy.
This strikes me as something of a parallel to when people sometimes claim that programming languages should be made "safer" at the expense of power provided for the programmer, and does not seem like a good direction to take the site.
>I don't think hiding comment scores will have that effect, though it might make it more difficult to recognize when it happens.
Comment scores make group beliefs far more obvious. Common knowledge of group beliefs is a necessary condition for groupthink.
>The one thing in this experiment that might (I intuit) help discourage groupthink related behaviors is hiding one's own score from oneself, because people will then be less emboldened when they're being idiots just for expressing popular opinions.
The thing we need to be afraid of is public praise or shame for writing a well-received or poorly-received comment.
>Unfortunately, this will also (I intuit) discourage people when they're being smart from as readily taking another look at whether what they've said is idiotic when others recognize the idiocy.
I disagree. When people are called out publicly for unpopular beliefs, they tend to commit to them even stronger.
Consider the following two stories. In Story A, Joe is downmodded on a comment with no replies and his low score is public knowledge. Joe becomes unhappy and stubborn. In Story B, Joe is downmodded on a comment with no replies and his low score is private knowledge. Joe wonders what might have caused so many to individually downmod his comment. He replies to his comment and says "Just curious, I noticed that a lot of people seem to be downmodding my comment. Is there some sort of flaw in my argument that I'm not seeing?"
> Comment scores make group beliefs far more obvious. Common knowledge of group beliefs is a necessary condition for groupthink.
True -- but obviousness is not a prerequisite of common knowledge in a knowledge-oriented community like this. Maybe if it was Digg the scores themselves would be a lot more necessary for groupthink.
> The thing we need to be afraid of is public praise or shame for writing a well-received or poorly-received comment.
I think you're underestimating the power of a pat on the back.
> I disagree. When people are called out publicly for unpopular beliefs, they tend to commit to them even stronger.
See above, re: "being smart". Were you not reading what I said?
I didn't realize how integral the numbers were to how I read HN until now. It's taken me a while to stop feeling lost without them, but with the orange dot present, I at least feel like I'm getting the insightful comments after a quick scan of the page.
Well, no. With the numbers I can get a sense of how quickly a comment is being upvoted by looking at the time it was posted. I can also compare a comment's total points to the those of surrounding comments for an indication of its relative worth.
These things aren't perfect measures, but I find them useful.
It's just a filter. By reading only high scoring comments you will get a combination of insightful posts and annoying witticisms. Only rarely do you get a false positive: a post that with a high score that is inane or simply wrong.
By reading every post you'll certainly find the "diamond in the rough", posts that don't get the recognition they deserve. But you'll also have to trudge through loads of posts of little merit. Given that there is an infinite amount of information on the internet, those 2 minutes can probably be better spent on wikipedia or reading the top post in a different HN thread.
Additionally: HN isn't just about opinion, it is also for a large part about fact. And insightful facts almost always rise to the top. On the bottom of threads you often find things that are wrong or subtly wrong.
Judging by the numbers may very well be the most efficient and reliable way to get to insightful content.
If it's a post with a lot of discussion, it'd help with one or two more levels of "good post" indicators.
Maybe making the default text color slightly more grayish and the "a bit better than orange dot" have #000 text, and for the exceptionally upvoted posts also make the "pg 26 minutes ago | link | parent | flag" header bold.
Goes to show that removing scores was a good thing. You are saying that you are more interested in following the herd than in evaluating the text, which is where the real information is, for yourself.
BillSwift your comment has all of the sniping dogma that you attribute to Herd Followers.
Have you tried to think of a more generous reason that people might look at the scores? perhaps people (like me) do a first pass to understand the herd reaction, and then do a second pass to delve into the specifics.
Simply saying score readers are herd following cretins is insulting and wrong.
I said nothing at all about anyone being cretins; sounds like you are being a bit defensive. The post I was responding to said specifically that he used scores as "which comments to keep an eye on" and to avoid "getting lost in the text". But actually reading the text is the only way to absorb the information. The biggest problem I have is with following a comment back to its reference when there are a lot of intermediate comments - allowing comment levels to collapse, like emacs outline mode, would help.
Don't higher rated comments get pushed to the top?
And of course these dots will help you locate higher rated comments that get pushed to the bottom (perhaps because they replied to a low rated comment).
I hope it lasts long enough to gather some stats on voting; whether or not more or less votes are cast with the new system, how spread out they are etc. I liked the idea of masking votes when it was initially suggested too. Bravo to pg for trying it out.
Yes, concerning the site speed I definitely upvote the idea of making the server faster. Maybe they really removed the comment scores to save bandwidth. ;)
By the way, suppressing these metrics is pointless. We are intelligent people who notice the difference between two comments with very similar upvotes and one that far outpaces the other and find these proportions to be of interest. The new portrayal of comment rankings keeps interesting information from users and will end up stunting the community's progress. If you're just trying to ward off users, maybe remove the nav and make the site even slower.
There are far more interesting ways to differentiate. I'd like to see this site be as light and fast as possible (would appeal to our hacker ethic) and allow users to customize the hell out of it given knowledge of Arc or similar (good chance to promote your language as well).