I read about bits of this concept here and there, but I haven't seen it implemented, though it could be very useful, revolutionary even.
Lots of information reaches us via the net. Lots of news, ideas, comments, and one usually has to spend lots of time cleaning up trash until some gold is found.
People come up with clever solutions to mitigate this problem. For example, the Digg community filters individual items, so that trash is buried and better content is endorsed.
This is very useful, I'm also a Digg reader, but it has the drawback that others still find things interesting which I don't, so the problem is not as bad as before, but it's not solved completely.
In order to solve this one could use the relations between the diggers (using Digg as an example, but the method is not limited to it), so that if one scores articles similarly to me in the long run then the system should give more weight to his or her scoring when showing news for me. The italicized part is important, because it has the implication that there is no common front page which is the same for everyone, there is no common denominator, but rather there is a personalized front page for each person showing articles which are relevant for him or her in the first place.
Why should there be a common, generalized view in case of a news site which relies on community scoring and has no group of editors who approve articles to appear on the front page? It doesn't make much sense, because the data is available to bring things to the next level.
If the above scoring method is used then very dynamic groups or bubbles are created from users with similar interests and values. These people are bound together by their scoring history and they help each other to find the best stories and articles within their interests.
Note that these groups do not exist in the algorithm, the word "group" is used only to illustrate the effects of this method in the long run.
If one changes his taste and starts scoring articles differently then he leaves his current "group" and is connected with other people who share that different taste.
This method would be very effective against spammers too, because they can send in their links and give them high scores, but if others - whose opinion based on their scoring history has more weight for me - score it down then I won't see it. So the spammer will be alone with his scoring history or he might get connected with other spammers by the algorithm, so that they can read each other's links.
This method would decrease the need for censorship too, because the dynamic group I belong to would work actively to show only those items which I want to see, therefore porn links are unlikely to get high scores in a "Christian group". On the other hand, a "porn group" would value these links highly and vote down religious propaganda.
In essence this algorithm would result in a personal Internet where (if I actively vote for or against the items I see) I would rarely see content which I consider trash, and much more of those which I like.
One could choose to go deep and score in the raw, unscored news flood. These people would be the adventurers who are not afraid of exposing themselves to different ideas or simple trash. They would dig up the promising items from the swamp.
Others could choose to play it safe and score only items which are already discovered by their peers. These would be the people who would decide the fate of an item for other safe players.
I hope something like this will be the next step in the evolution of sites like Digg which rely on user participation.
I hope to see it implemented soon.