I like the idea of eliminating crap from a public forum.
And there's a lot of crap in public forums. I'm sure you (like everyone else who doesn't work in advertising or watch Mad Men religiously) hates advertising. That's all over public forums like TV, the Internet, radio, periodicals, well basically everything that gets people's attention in any real way has advertising.
And spam. Most spam is advertising, good to get rid of that too.
And trolls. They're the worst. If it were up to me we would put those people to death - go talk about something you like, nincompoop. (love calling people that) You're wasting your own time for the explicit purpose of angering strangers and wasting their time. Way to do something of value to anyone, even yourself.
There's a bunch of reasons. Let's figure out the best way to do it.
First of all, we need to have a strong user ratings system. It should be simple enough for any user to use like the popular +/- 1 system (also royalty-free) but requires some tuning for accuracy.
Every post or thread should begin with a value of 0, adding or subtracting a score based on the accumulate votes that post or thread has received. Additionally, threads should add or subtract 10% of the average value of posts in the thread to their value.
Every user should begin with a weight of 1 per +/- vote, a state we will refer to as equality (EQ). Users who are below a value of EQ*.5 will lose posting privilege and a moderator will be flagged to investigate that user.
Moderators, at one end should have votes that carry the most weight The more trusted they are in the community, the more power they should have, and moderators are the police. Just like in real life, moderator rule should not be absolute - there should be crowd-sourcing moderation for the sake of simplicity, democracy, accuracy, and efficiency. Just like the police, they should not be allowed to simply "kill" you(r account, thread or post).
But a moderator's vote should count for let's say 10*EQ. This may sound high, but remember that this is the top of our curve. This also makes it simple for mods to quickly flag new topics or posts as they see them to other mods and remove them from the general stream of discussion.
Moderators would lose the ability to delete threads individually, but retain the ability to move threads off to a separate "incarceration" category (in our police metaphor), only threads in incarceration could not be moved by their original poster, only by the mod who placed it there or by a separate tally of moderator votes alone to move it out of incarceration, to prevent abuse by an individual moderator.
The process here would involve the disagreeing mod calling for a vote, and then other mods having a week to vote, with individual mods being allowed to abstain (the default vote). Once an incarcerated thread received more than half of the moderator votes on the site for it to be released, it would be automatically returned to the control of it's original poster in it's original position, as well as symbolically linked in a separate "released" category. This is designed to prevent individual mods from attempting to remove a thread
For a thread or post to be deleted, it must have received more than half of the mod votes negative, more than two thirds of the user base votes negative, marked at least 3 times as spam or offensive, and have been incarcerated for a week.
If a thread or post's score ever drops below -10, any logged in mods should be flagged until the thread is incarcerated or a moderator responds with a "do not incarcerate" order to the system.
If a thread has been incarcerated, that thread's original poster should lose weight on future votes they make, but this should be a small but significant decrease of.1 votes per +/-. Past votes should receive a smaller penalty of .01 per +/-, resulting in long term degradation of the voting weight of malicious users. For users with a vote weight below equality, this penalty should degrade at a rate of .01 per week towards equality. This prevents genuine mistakes from stacking on a user and resulting in their banning from the system.
To scale this theory to individual posts (the finest granularity of the message board), simply gray out low-rated posts, with posts rated below -15 flagged for deletion by mod vote. The combined user's vote score for this post should count for two thirds of the total score of all mod votes in the post deletion process. This vote closes in an hour from the time the flag is thrown to the system, and to prevent abuse by mods.
In order to prove themselves as newbies, in addition to post count, newbies build up their rating by voting on the site. Newbie posts are automatically grayed out. Once they cast 1,000 points of votes positive, and 1000 points negative they gain full membership. This encourages newbies to read and participate in the site, post well early, and not troll, spam, or harass.
If for the first 50 posts a user can stay above .9*EQ, there is a high likelihood they are not a malicious user.
Furthermore, let's reward the approved user base for reading and filtering newbie posts for us - that's the whole theory behind crowd sourcing. Every vote on a newbie's post or thread by an established user should add .001 to the established user's voting weight for the next 24 hours. This will make users happier to deal with the spammers and trollers - users who care about voting on newbies (the source of new minds in the environment, and the source of spam. In order to prevent it from being abusive we will place a cap on this bonus weight of (.1*user's base weight) and so maintain the relative power of the moderators.
This system also makes it easy for users to opt-out of the newbie filtering process - simply don't view grayed out comments or click on the incarcerated category.
But most importantly, it solves our fundamental issue here: designing a system which will facilitate rapid entry of new users to the forum, while filtering malicious users both at an introductory and long term level, in a way that removes most of the effort from any one individual, both lightening the workload and democratizing the moderation process.
I'm sure there are weaknesses and problems in this theory, please discuss them here.