Peer-to-Peer Accountability Enforcement/mechanism/refinements

''This needs to be updated a bit. PCR needs to be defined.''

Ratings of Posts
For the sake of brevity, I'll use "post" to mean anything posted by one user for others to see, regardless of whether it's a root-level post or a comment on a post.


 * Each post has its own PCR that is separate from the PCR of the user who posted it.
 * The PCR for any given post will default to the poster's PCR.
 * Any additional post-specific ratings will modify that post's PCR by some sort of weighted average.
 * Experimentation will be needed to determine the best algorithm, but we could start by weighting the poster's PCR by how many individual ratings went into it.
 * Each user X only sees posts whose PCR exceeds X's Minimum Required Credibility (MRC -- see below).
 * Each user X may adjust the at various levels above their MRC.
 * Each user X starts with a default global MRC, but they may modify that default and they may also set specific MRCs for individual threads or groups.
 * All users can rate each post's credibility.
 * Optional: ratings of users' posts could have some small, cumulative influence on their overall PCR. Perhaps the influence of any given user A's ratings of B's posts should be overridden whenever they revise their overall rating of B. Experimentation needed.

"MRC" refers to a setting that each user can adjust to determine the strength of post-filtering. The lower it is, the more posts they'll see -- because they'll be allowing less-credible posts through. The higher it is, the higher-credibility a post has to have in order to be visible. (There will presumably be other user-options to determine how "hidden" posts are indicated -- e.g. a summary showing how many, a list of usernames, etc.)

Frequency of Notification
Rather than simply blocking those whose PCR falls below a threshhold, users can "distance" themselves from other users by selecting how often they wish to be notified of posts based on the poster's PCR (and possibly how often they wish for their posts to generate notifications to that other user).

For example, if I see user A with a high PCR and user B with a low PCR, I might want to be notified immediately whenever user A posts, but only every week regarding user B's posts. (Similarly, I might want to avoid notifying user B right away if I respond, in order to further reduce the amount of time I spend interacting with them.)

This would promote higher engagement between higher-value users without completely excluding others from the dialogue.

Note that this is different from the random-selective notification of social networks such as Facebook and Google+, where some posts are simply ignored. (The latter system may require substantially less computing power; a data design for a FoN system has not yet been worked out.) All posts are eventually included in a notification, but may be time-delayed and grouped with other posts by the same user.