Publication System

For the past year I've been getting all sorts of ideas about how to apply point to point software for all sorts of things. One such idea is a direct replacement of the old Usenet bulletin board system; you know, the alt.*, comp.*, misc.*, and other such message groups. It's sort of like a combination of Usenet and the web, but it's more of a direct publication system. The other application is a scaled down version of it used for personal ads.

The only problem with such a system and why you haven't seen them is all the spam they would attract. But I believe the solution to that is an automated trust group based system. Basically everyone catalogs and flags everyone else to various degrees. When you read a personals ad and it isn't categorized you flag it. If it's spam you flag it as spam, Everyone does this and their lists get swapped and you compare other people results with your own and the computer automatically establishes a trust factor as far as how accurately a person catalogs ads. You can also have hierarchy's of catalogers that more of less vouch for each other and that's when it becomes really effective. It would be pretty much automatic unless you decide to look at unreviewed ads or improperly cataloged ads. I'd imagine at first there would be quite a bit of spam but it would die off once it's realized how effective the system is. People establish anonymous reputations based on what they publish, how long they have been publishing, and how often they publish.

The only weakness is a minor one, if computer hijackers hack into people computers and assume their identity to post spam. But that can probably be mitigated by noticing the number of posts a person makes, stammers would sent tons of messages and it would be flagged immediately as suspicious. The system is redundant so there really isn't any cheating. Say someone chooses to sell out and allow certain spam though, you'd notice this and that persons reviews are trusted less and you start relying on other people reviews more. Other than power users that volunteer that take on more complicated reviewing and sponsoring tasks, it's really an automated system a user can choose to ignore and just know that it will work 99.9 of the time.