Surely one of the potential solutions has to be verifying accounts with a formal ID (even if you choose to keep the public facing part anonymous)? Then give users the ability to hide(/block? "You can't @ this person if you are not verified") everything from accounts that haven't been verified?<p>It wouldn't solve <i>all</i> the problems, but would surely go some way? And you still have the ability to live in a Twitter wild-west if you want to.