I agree that it goes a long way to tell people how you prefer email. However, enforcing those preferences (through something like shortwhale) is different than expressing your desire and then hoping that users follow those desires while emailing through ordinary means. In other words, there are two ways of forcing a user to follow those desires - technologically or through force of implied disapproval (with whatever outcome that disapproval would hold, such as non-response). Might this difference be a good opportunity for A/B testing, and perhaps thereby helping Dan's research?

I suspect that the force of disapproval would be just as effective as the technological force for 'meaningful' emails, but would lose effectiveness for less meaningful emails (where Dan's disapproval doesn't mean much to the sender).

If so, then the real issue may be that Dan is getting too many emails from people who don't care what he thinks. The classic way to deal with this is to have one address for people who know you (and care what you think) and another for the unwashed masses. The problem with the classic solution is that it requires someone to monitor the unwashed email, or you suffer the opportunity cost of ignoring possibly important messages from people you don't know (yet). I wonder if Dan is already using this method, and shortwhale is only for the unwashed email address, thus reducing the cost while keeping the opportunity?