You can test how well this would work by collecting a corpus of whatever sort of comments you think HN could do without, and then checking to see how many of those commenters would have run out of karma using whatever combination of initial karma and cost to post you have in mind.
I'd be open to experimenting with charging karma to comment if someone did a study of that type and showed that it would work.
(I realize it's not a perfect test, because people would make different comments if they knew comments cost karma.
But that is a desirable direction to have error in.)
I'd be open to experimenting with charging karma to comment if someone did a study of that type and showed that it would work.
(I realize it's not a perfect test, because people would make different comments if they knew comments cost karma. But that is a desirable direction to have error in.)