I just found out from ReadWriteWeb that someone was messing up data on OpenStreetMaps. This is similar to the vandalism that occurs on Wikipedia. Mentioned in this article was an article from 2010 about poisonous people in open project communities by a member of the OpenStreetMaps community.

This is a separate issue from vandalism but it still hinders a project from accomplishing anything. Basically, there are people in communities who will nitpick on minor details and slow down the process so that nothing gets down, or they will do something else that deters others from making a contribution.

I'm thinking that mailing lists, IRC, email, and other forms of communication absolutely need trust filters. You set up how much you trust a person on a mailing list and if you label them a troll, then their messages won't be received by you at all or they will at least be hidden from view. The trust values that you set could be shared among friends on the mailing list or shared with the whole mailing list.

I'm not sure how much this will help since the real problem is that poisonous people are allowed to join the community at all. Maybe having a tiered system where contributors whose contributions have been verified as good have their own forum but are also part of the public forum would be helpful. It may appear as if an hierarchy will form with a tiered system but anyone can contribute and if no one verifies their contributions as good, they still have the public forum to complain about this (the contributors group has no control of the public forum). If it becomes necessary, a revolt may happen and a new community will form with different standards, basically like a fork of an open source project.