2 ms·
So, I get their broader point: It doesn't scale; too much human time and effort to read the comments and apply policy correctly. What would scale, then? You'd
by SilasX 10y ago
So, I get their broader point: It doesn't scale; too much human time and effort to read the comments and apply policy correctly.
What would scale, then? You'd want to a) use free labor from readers, and b) have a way to punish trolls.
I imagine something like:
1) Some central identity/commenter rep provider (CIRP) that shares rep across forums. (privacy issues discussed below)
2) If you're registered with a good CIRP at all, you get a big bump in visibility -- maybe even a trashed account appears before anons. The higher a rep your CIRP vouches for, the more visible your comment.
3) Readers upvote and downvote as before, but this doesn't directly affect visibility; it just decides probability of the comment being placed before a human reviewer, who can pass a more focused judgment on it.
4) If multiple, non-communicating human judges agree on good/bad, that users' rep is adjusted in that direction.
5) Privacy: the CIRP just signs a token indicating your rep (within a range), not your ID itself. (They don't even need to store your comments, just your login and rep, though comments are nice for later auditing.)
The advantage are:
A) If you establish yourself as having good comments on other forums, that carries over to the NYT (or whatever) comment section, creating an bigger incentive for good comments.
B) Humans can take the time to carefully judge your comments, since they only need to see a sample of them.
C) It has a nice hierarchy of checks and balances: forums can discount bad CIRPs; CIRPs can check that their judges are doing it right; judges validate other judges.