I took a one-day trip to Boston last week to visit the MIT Media Lab. 

While I was there I heard a presentation by Caroline Sinders about ways Twitter and other online services could offer controls that would help minimize abuse. 

She designed new prefs for Twitter that I thought were pretty clever. Here's a screen shot.  And a summary:

• Allow followers of your followers to tweet at you?
• Do not allow accounts with less than X followers to tweet at me, @mention me, or follow me.
• Make my account public, but my handle can't be tweeted at by non-followers. 

I'd of course add my favorite (that doesn't actually exist) -- mute-with-timeout, which would allow me to turn off messages from people who might have a reason right now to be angry, but who I'd like to automatically give another chance to communicate with me, after having a chance to cool down.

We should be doing more thinking about stuff like this, and of course it would be great if developers working outside Twitter could implement clients that were more abuse-resistant.