I'm inclined to believe that the normative case for these kinds of documents is not intentional harassment, but simple miscommunication. Starting with encouraging introspection and moving through mediation before getting to outright banning makes sense. Moving straight to ejection for comments that could easily not be intentionally abusive, is just giving ultimate authority to anyone who takes offence. Better to put both sides of a dispute on neutral ground, and allow community leaders (or the community at large) to sort out the disagreement. But a key step before that should be to encourage people to consider how their words and actions could be interpreted to hopefully avoid offending in the first place.