Hacker News already has a flaw that conversations tend to be too short-lived to put any real thought into a comment. To describe what I mean by this: in order to do some real research, or to perfect the wording of a comment, might require a couple hours of time: right now, unless you are willing to spend all of that time frantically and immediately when a thread is posted, by the time you have managed to contribute something of real meaning the activity on the thread has died down because everyone has "already read it" (and won't return) and it may, in fact, have already dropped off the home page (and so will get no new traffic).
I thereby feel the need to point out that using a system where the comments are promoted by being viewed inside of the thread has the potential to worsen this effect, as the pool of people available to promote comments will decrease over time, even though the potential value of the comments in question increase over time (as, again, they will be able to have been better perfected, with more research, better wording, and careful thought). I might, thereby, put more weight onto your idea for putting the comments on a page like /newest, in order to better force them to be seen after-the-fact (which, I understand, may be one of the factors you are already thinking about: I'm just attempting to show that someone else might be thinking about that sort of thing).
(This issue has an interesting effect when combined with the Eternel September + Evaporative Cooling interaction effect that this site seems to have: I have found myself "taking a pass" on the first time something is posted, instead working on a comment that I post when the same thing eventually comes up later to a sufficiently-different set of people that it can again accumulate enough votes to hit the home page. FWIW, I would then make the argument that this problem of "good content takes time, and HN is setup in such a way as to discourage things that take time" is a more fundamental cause of the "misinformation due to comment blight by people who don't really know what they are talking about" issue.)
There seem to be a number of features in hacker news which presume longer threads tend more towards noise than signal. The threaded format as implemented, for one, makes it difficult to follow conversations in long threads, particularly in replies to the root post if you're linking from the latest comments page. The further into the nesting you get, the more crunched the text becomes. And, apparently, posts take longer to show up.
Also, there not being a 'notify' feature more or less ensures that anything not on the front page is almost certain to die, although hnnotify.com does a good job of filling this feature in, that the service exists to me suggests that it could be useful as an actual feature for the site. And I know it's been discussed and passed on, but still.
I also worry about how the effect of weighting votes based on karma will bias conversations (though that would kind of be the point I suppose.) I guess we'll be seeing a lot more crypto stories bubbling up in the future...
This is so true. Karma and other scoring metrics just magnifies the problem as the number of people and therefore also the potential karma points decrease the later the comment is posted. It's not an easy problem to solve because people naturally gravitate towards the newest items.
as i was not quite able to articulate below, it feels to me that forum comments, email and todo-lists lie on the same protocol manifold but not quite at the optimal point. (RE: your yesterday's comments... i use browser tabs as a to-do lists.) The obvious axes are synchronicity (temperature) and information (related to entropy) but probably this is still not the best frame to think about this problem.
I thereby feel the need to point out that using a system where the comments are promoted by being viewed inside of the thread has the potential to worsen this effect, as the pool of people available to promote comments will decrease over time, even though the potential value of the comments in question increase over time (as, again, they will be able to have been better perfected, with more research, better wording, and careful thought). I might, thereby, put more weight onto your idea for putting the comments on a page like /newest, in order to better force them to be seen after-the-fact (which, I understand, may be one of the factors you are already thinking about: I'm just attempting to show that someone else might be thinking about that sort of thing).
(This issue has an interesting effect when combined with the Eternel September + Evaporative Cooling interaction effect that this site seems to have: I have found myself "taking a pass" on the first time something is posted, instead working on a comment that I post when the same thing eventually comes up later to a sufficiently-different set of people that it can again accumulate enough votes to hit the home page. FWIW, I would then make the argument that this problem of "good content takes time, and HN is setup in such a way as to discourage things that take time" is a more fundamental cause of the "misinformation due to comment blight by people who don't really know what they are talking about" issue.)