>>1
It seems like the difficulty with federated forums is allowing different instances differing views of what is considered effective moderation. Just to free associate a little, imagine there was a public list of rules (even hierarchical rules) which any instance could select from or add to. Then moderation could be federated by having each moderator select the reason for the moderation action from the list and having this propagate to the other instances who would programmatically choose to honor it depending on if they have selected this rule for their instance. Moderation actions could propagate so that replies to posts by banned users are also hidden. Instances could also have a mechanism to flag moderation which is considered inaccurate, and would also need to be able to unpeer moderation from those who they deem to moderate excessively.