The problem statement in the submitted public discussion post:
"The really interesting comments, however, remain forgotten somewhere below, because too few people take the time to scroll down, find them and read them."
The solution proposed:
"This should solve the above-mentioned problem:
"The computer counts how many people have read each comment, and takes this into account, when it sorts all comments."
Ladies and gentlemen, please check my reading comprehension. Do you see what I see here? The person posting says "too few people take the time to scroll down, find them and read them" and then says "The computer counts how many people have read each comment, and takes this into account, when it sorts all comments." How does this give any more prominence to comments that few people are reading (as compared to comments that more people are reading) than any other way of sorting comments? If the problem is that some people aren't reading certain comments, how can how often those comments are read be used to draw more attention to those comments?
Perhaps I am too tired after a weekend day of teaching followed by research to understand what is being proposed here, but I don't think this makes sense.
Anyway, in threads here on Hacker News, there are other ways to find good comments. First of all, there is the bestcomments view of the community,
which, while not a perfect technical solution either, sometimes does promote sub-sub-subcomments to visibility far greater than the visibility of the original greatgrandparent comment in the same thread. Some readers of Hacker News also follow people who post good comments by looking up the links to their comments from their user profiles, for example:
We can also use HN search to search up comment threads by keyword, and evaluate them for ourselves rather than by how they are placed in a thread. Anyhow, I don't worry about this. Sometimes I think the top comment in a thread is the most interesting and informative, by far, and other times I read far down into a thread to find the comments I like best and need most. Either way, there is plenty of good stuff here. The best way to bring about more good stuff here is to read a lot of the comments thoughtfully, and to upvote all the good stuff you find. Emphasize the positive, and upvote early and often.
Instead of ranking by net points, or points over time, the comments could be ranked by points per view. This essentially gives a close approximation to "What percentage of people who viewed this comment found it worthwhile?"
So the problem is that sorting by upvotes alone creates a list of comments where the first comments to the pack get the most reads, and therefore the most upvotes, such that the first comment will have 200+ upvotes if it is posted first, but may never even see the light of day if it's added later, once a majority of the votes have already been cast.
The solution is to sort instead by (upvotes/total views). That way, the comment at the top that 1000 people have viewed and 500 people have upvoted falls back behind one that 50 have viewed and that 40 have upvoted, a comment that typically would go unseen by the masses, but which is signficantly more likely to be upvoted (or enjoyed) by any given person.
"The really interesting comments, however, remain forgotten somewhere below, because too few people take the time to scroll down, find them and read them."
The solution proposed:
"This should solve the above-mentioned problem:
"The computer counts how many people have read each comment, and takes this into account, when it sorts all comments."
Ladies and gentlemen, please check my reading comprehension. Do you see what I see here? The person posting says "too few people take the time to scroll down, find them and read them" and then says "The computer counts how many people have read each comment, and takes this into account, when it sorts all comments." How does this give any more prominence to comments that few people are reading (as compared to comments that more people are reading) than any other way of sorting comments? If the problem is that some people aren't reading certain comments, how can how often those comments are read be used to draw more attention to those comments?
Perhaps I am too tired after a weekend day of teaching followed by research to understand what is being proposed here, but I don't think this makes sense.
Anyway, in threads here on Hacker News, there are other ways to find good comments. First of all, there is the bestcomments view of the community,
https://news.ycombinator.com/bestcomments
which, while not a perfect technical solution either, sometimes does promote sub-sub-subcomments to visibility far greater than the visibility of the original greatgrandparent comment in the same thread. Some readers of Hacker News also follow people who post good comments by looking up the links to their comments from their user profiles, for example:
https://news.ycombinator.com/threads?id=patio11
https://news.ycombinator.com/threads?id=raganwald
https://news.ycombinator.com/threads?id=jgrahamc
We can also use HN search to search up comment threads by keyword, and evaluate them for ourselves rather than by how they are placed in a thread. Anyhow, I don't worry about this. Sometimes I think the top comment in a thread is the most interesting and informative, by far, and other times I read far down into a thread to find the comments I like best and need most. Either way, there is plenty of good stuff here. The best way to bring about more good stuff here is to read a lot of the comments thoughtfully, and to upvote all the good stuff you find. Emphasize the positive, and upvote early and often.