[1] http://news.ycombinator.com/item?id=2434333
Edit: Seems that PG answered a similar question recently http://news.ycombinator.com/item?id=2465002
1. From my experience, I stopped voting because I am not being educated on the fact that we need to vote (well I can't see the rating, so why vote). I have a feeling that other users act similar. Please confirm.
2. The voting process shows the values of HN. Comments that are ok in techcrunch would be downvoted here within a second. And I feel that voting system needs to be open to educate new hackers & founders about what's acceptable and what's not. I know there are rules but we learn by practice.
3. When voting process is open, good comments get even more votes, and irrelevant or bad comments get even more downvoted. I personally don't want to read irrelevant posts. I have limited time but now I feel like going through comments that really don't have enough information.
PG, I hope you take my personal opinion into consideration. Thanks
Something simple like (for example) a comment needs 1 upvote per day to remain unhidden, until some number of days pass or some number of points are achieved (4? 5?), and then the comment remains visible in perpetuity.
I like the new no-points system, but I do agree #2 is a concern and could used some tweaking.
On the side note, I got more votes right now than ever before. Go figure how people think about voting o_O
In the past, there were 2 signals of the comment-quality in the thread.
1. Vote count.
2. Positioning in the thread.
Now, that Vote count is not visible, There is only one signal left.
1. Positioning in the thread.
It makes it very difficult to segregate good comments from the bad ones, and decide which comments add to the topic, and hence worth reading/discussing, especially on a large thread, on a new subject matter, reader does not understand very-well.
Few suggestions to improve the comment readability of the thread.
1. Would it make sense to hide the comment count of a thread till a particular threshold (say votes < 7, comment time < 10 mins), or of the entire thread comments (thread time < 30 mins, number of comments < 30) and the likes, and make them visible to add to the readability/visibility of the thread?
2. Would it help if the up-votes are capped?
It's right for good comments that are read by more people to receive more votes. Also that the first person to make a particular comment should be higher voted as repetition of the same insight isn't improving the thread.
I don't agree with this. Just b/c someone was first to post a mediocre comment doesn't mean it's good or worth reading.
Also, early posts are often knee-jerk and among the least thoughtful of the thread.
When I feel clueless about a topic I look to comments with up votes to get a general feel for the direction of the discussion and hints how to educate myself on the topic.
On your first suggestion: Maybe it is hard to work with absolute thresholds as the amount of votes is probably proportional to how long something stays on the front page. Maybe there is some measure to determine when a thread has begun to "settle" down and that would be the time to make votes visible.
I'm not sure if positioning in the thread is much of a signal. Comments with a low vote replying to a comment with a high vote will get displayed above comments with a medium vote, and will more likely be read. Before, the reader could choose to ignore such comments, and keep on scrolling down to check for other comments with a high enough vote.
In sizing up what this means, let's all review together the stated reason for the experiment, expressed in a thread-opening post by pg titled "Ask HN: How to stave off decline of HN?"
http://news.ycombinator.com/item?id=2403696
He wrote, "The problem has several components: comments that are (a) mean and/or (b) dumb that (c) get massively upvoted."
How do the highest-voted comments visible in the bestcomments list
http://news.ycombinator.com/bestcomments
look to all of you recently? Are there fewer mean comments than before? Are there fewer dumb comments than before? Are the comments that are "massively upvoted" since the experiment began mostly comments that are reasonably kind and well-informed, helpful comments on the whole? In most of the treads you visit, do helpful, thoughtful comments seem to rise to a position of prominence, while mean or dumb comments gray out?
A link and comment in the thread referred to by the edit on this submission largely sums up the back-and-forth about visible comment scores as a signal on comments in active threads:
http://news.ycombinator.com/item?id=2465357
>> Please bring back the comment scores. It helps a lot in parsing the comments and assigning a proportional weight to each when reading them.
> I had to think about this a bit, and I disagree so far. I'm finding that I'm not pre-judging comments as much. It's nice to be able to read someone's comment without knowing first that 70 or 80 or 3 other people thought it was worthwhile.
My impression too is that even with comment scores not visible, it is still convenient to browse threads to find thoughtful, informative comments, but now there is less anchoring bias
http://www.sciencedaily.com/articles/a/anchoring.htm
of most votes on a comment converging to one score level that shows up early in a thread's development, and more engagement by readers of HN in actively reading comments and upvoting (or downvoting) based on each comment's characteristics in light of the context of the thread.
But the main criterion of the experiment is to "stave off decline of HN," and that is what will decide if the experiment was successful. For that every reader can help by actively upvoting informative, helpful comments, and also by downvoting comments that are either mean or dumb--and especially comments that are both. As I recall, the experiment has also involved some changes in the effects of flagging, so flagging inappropriate comments is also helpful.
but now there is less anchoring bias
How do I know what's important without vote counts?
Maybe you get twenty-three down votes and forty-six ups; the point count isn't going to tell you that. I know that I have comments where the score hovers around one but have maybe ten votes that cancel each other out. What to make of that?
A voting system should have modest goals. (While the Slashdot community hasn't been my cup of tea for a while, I feel like they've muddled their way to a system that works pretty well.)
Meanwhile, the useful/useless votes are smaller or perhaps harder to access, like the current "flag" system for comments where you have to click through to the comment's own page first. Perhaps voting either agree or disagree automatically registers your vote for useful, and to vote useless you have to click through like with "flag".
It's even useful for topics I'm already familiar with. If a comment I don't necessarily agree with has a high score, I'll wonder why. It's a great opportunity to walk down new mental paths or explore different angles about the topic without feeling vulnerable. If any insights emerge while doing this, I'll then follow up with my own comment.
In either case, it's impractical to analyze or reflect on every comment, so points are a pretty good heuristic. Even if it is a broken system, it does help a lot of the time.
Actually, the goal is for the comments to be interesting, but they will be because there are interesting people on HN. If interesting people are leaving, then that's a problem.
I feel that the tone of comments in general has improved significantly. Maybe that's confirmation bias, but I feel much more "at home" on HN now.
My experience was what I think common sense dictates. Since I couldn't see a score I couldn't let that color my opinion of a given comment and had to read it. This point alone has significantly increased my engagement in the threads I care to read and made me more carefully consider the top level comments on a story. Since pg turned off comment scores I've wondered if the primary reason people have reacted negatively is due to losing a positive status indicator. This is to say the score might have nothing to do with one's own engagement with HN, but with one's concern for other user's perception of their comments.
One other change I experimented with was moving the score of stories and comments to the far right top of the story/comment so that it wasn't the first thing in my line of sight. This helped with my bias as well.
At the end of the day this all anecdotal, and your mileage may vary.
One idea is to give each user only a finite number of comments they can make in a time period. As their karma increases, so does their comment allotment.
This way comments really can be voted on their merits rather than the fan club of the commenter.
Base 10 for example: the most-significant digit would be visible and the rest be zero's: '5' -> '5', '42' -> '40', '256' -> '200'.
That way, you could still see the impact of a comment while hiding the voting trends and other numeric details that keep you from unbiased (err, less biased) voting.
For example, asking for feedback or feature suggestions on the recently developed product to the community, will not be much helpful, if comment votes are not shown up.
Again, just ideas, I'm sure they can be improved or refuted (and I'd much enjoy the discussion). If you feel a comment is inaccurate or misleading, reply to it and explain. The reply will tell more than a number will.
I do find it interesting that the comment score has been hidden. That's kind of novel, and it might be the kind of solution needed in this case.
1) Removing "by" at the start of each comment stops me from trivially finding every other comment by the same poster in a given thread. Searching for just the name adds noise to the results. Usually not a lot, but I find it annoying.
2) Removing comment scores from our thread feeds makes it much more difficult to skim through all the things a given user has said recently looking for gems. Instead of a nice pre-filter (skim for posts exceeding some minimum score) you have to read every comment, no matter how inane.
Maybe I can give something in return for the great source code. :) Thanks PG.