How long a submission stays on the front page is almost entirely human-controlled, not vote-controlled.
It's algorithmic in the sense that there are obviously different thresholds, but a human decides which bucket a submission is in. And the bucket determines how long it stays, and whether it's low or high on the front page.
Then there are corner cases. Ever notice people point out in comments when an article is misleading (not just the title) and then a short time later the article gets downranked or whisked off the front page? It happens more than randomly.
The point is that the simple theory of how submissions work is almost entirely mistaken. HN is a newspaper; the editorial staff decides what to print.
Votes still matter in the case of getting things onto the front page, sometimes. Much of the time it's a result of the second-chance queue described in the article though.
The theory that people will find 30 interesting articles for the front page each day is false. It's a huge amount of hard work. Very few do it for free. And those that do, collectively tend not to have good taste. (Some individuals do, but not random HN members as a whole.) Leaving it up to the community would result in an uninteresting front page.
I think it's techie nature to want to find an algorithmic solution for something as complicated as HN curation. I fell into the same trap. But an automatic solution doesn't exist. The early history of flight proved that it was foolish to try to design a plane that would stabilize itself, as opposed to giving a pilot manual control. HN is much the same.
Reddit shows that there are automated solutions, of course. But at HN scale, the problems with Reddit's approach become significant. And we only have one subreddit here.
It's algorithmic in the sense that there are obviously different thresholds, but a human decides which bucket a submission is in. And the bucket determines how long it stays, and whether it's low or high on the front page.
Then there are corner cases. Ever notice people point out in comments when an article is misleading (not just the title) and then a short time later the article gets downranked or whisked off the front page? It happens more than randomly.
The point is that the simple theory of how submissions work is almost entirely mistaken. HN is a newspaper; the editorial staff decides what to print.
Votes still matter in the case of getting things onto the front page, sometimes. Much of the time it's a result of the second-chance queue described in the article though.
The theory that people will find 30 interesting articles for the front page each day is false. It's a huge amount of hard work. Very few do it for free. And those that do, collectively tend not to have good taste. (Some individuals do, but not random HN members as a whole.) Leaving it up to the community would result in an uninteresting front page.
I think it's techie nature to want to find an algorithmic solution for something as complicated as HN curation. I fell into the same trap. But an automatic solution doesn't exist. The early history of flight proved that it was foolish to try to design a plane that would stabilize itself, as opposed to giving a pilot manual control. HN is much the same.
Reddit shows that there are automated solutions, of course. But at HN scale, the problems with Reddit's approach become significant. And we only have one subreddit here.