Ask HN: Wouldn't it make sense to calculate karma on comments, not submissions? | Hacker News Reader

Ask HN: Wouldn't it make sense to calculate karma on comments, not submissions?

6 pointsUltramanoid16 comments

It seems irrelevant and unrelated to a specific user that an article that is picked at random and heavily discussed gives the submitter a large number of karma points while the submitters not 'chosen' but who have contributed in equal measure ( and often earlier than the one upvoted ) receive nothing.

One would think, this being a community, upvoted comments would be a much more useful measure for karma, possibly the only one needed ?

View on HN

Completely disagree. The articles are the most valuable part of the site.

Regardless of personality or politics, most can rightly agree that things like "Fancy Euclid's “Elements” in TeX" or "Text rendering using multi channel signed distance fields" are noteworthy accomplishments.

What some asshole thinks about those things, and what a bunch of other assholes think about his thinking, are neither here nor there--this comment included.

I've found that articles, even those that hit the front page, vary widely in quality, and the same is true for comments.

Some commenters are experts in their field, and their comments are worth reading. For example, I've learned more about security by reading tptacek's comments than I have by reading articles on security.

Indeed. And that is value only found in HN's comments. Weirdly less valuable ( by karma measure ) than just posting links anyone can find online.

That's not my point. Of course articles are the most valuable part.

But how does it make sense to give 400 points to a submitter out of 5 who posted the same thing ? And often to an article which may not be neither the first, nor the best about the issue. That's what I mean by entirely random.

Furthermore, without threads and comments HN would just be another link aggregator from some relatively fixed sources.

And you prove my point about comments mentioning a-holes, in that useless ones would be downvoted while useful or informative ones would be upvoted, and quite rightly bestow karma and benefits to contributors deserving it. That would make sense. You wouldn't see actual a-holes, they'd be downvoted to oblivion.

Giving the big points to submitters encourages people to keep submitting. Maybe one of the other five will get lucky next time.

I upvote every submission of the same article that I find interesting or useful, but it gets ridiculous after a while. And not to do so seems utterly wrong from my point of view, since points are given to submitters, and receiving points grants them benefits.

Maybe if HN consistently registered the first submitted link and marked all others as dupes automatically. Even so, still seems a flawed reward system.

  I upvote every submission of the same article that I find interesting

... then you are intentionally rewarding submitters of dupes?

So I should vote only the first submitter and to do so hunt down all the instances of the article posted before, to find out who was the first ? And that'd be the value in karma -- who submits first ? Sure, it'd make at least more sense than it does now, which is randomly giving hundreds of karma points to anyone.

Still wrong in my view, and worth giving more value to comments. I've already commented anyway, not sure why HN does not automatically register the first time a link is submitted and redirect the rest to the first one.

That would depend on how karma is used, right? I'm not sure if it is well known how various HN algorithms depend on karma.

I thought the point of HN is to share intellectually interesting stories and points of view, with karma etc. just being a (maybe) necessary by-product.

I feel your proposition would not affect this goal in a good way.

What is the point of karma then ( which is another and perfectly valid question -- why not just get rid of it... ), and why it grants benefits as it goes up when it is ( seemingly ) given entirely at random ?

To get karma, you need to have had some level of participation in the community (posting useful articles or making useful comments).

By putting a karma threshold on actions like flagging or downvoting, HN ensures that these actions can't be done by someone who has just created a throwaway account to harass people.

Karma is given to reward good contributions. There is perhaps an optimal scaling of each type of upvote, but without further study, a 1:1 ratio seems pretty reasonable.

Upvoted comments do get karma, just as upvoted stories do.

Upvoted comments are directly related to the user's comments, and thus the user, so that makes sense. Articles are not related to the user* posting them. It's basically random unless I'm missing something.

Like I said, I upvote all submitters of the same article because otherwise it seems wrong given how karma works now. But we shouldn't have to do that.

Edit : * With the exception of Ask HN or Show HN, I suppose. But the vast majority of links are from external sites and submitted several times by different users.

Stories are the life blood of HN. Without stories, no comments, and no Hacker News. So it doesn't make sense to reward them less.

I agree that there's an issue with randomness about which submission of an article gets the karma. It's on our list to do something about that. But it does even out in the long run for anyone who submits many good stories. And I don't think most HN users care about karma too much.

Reply on news.ycombinator.com