I like the idea. But what's to stop that user from just creating a new JID?
We might just see a lot of one-time JIDs pop up as happens with email spam
now.
If Jabber really takes off, someone will create a special Jabber server for spammers, which just sends every message from a different randomized fake JID. Although at least, with server-server dialback, they won't be able to fake the server name on their messages (right?)
For this and other reasons I would rather see trust as "opt-in", i.e. people will be blocked unless there is a reason to consider them trustworthy. Or if not blocked completely, then at least downgraded in importance, i.e. messages from them don't pop up on your screen but just get added to a list in a window that you can open once in a while and look through. Sort of like the "Junk?" mailbox that my mail rules throw suspicious messages into.
—Jens
