I think I have failed to make my point why a pull system would be effective
against spam and viruses. The main thing in my proposal actually is that a
pull system (where the mail client displays outgoing mails) will be a visual
proof that something is very wrong if the owner can see strange mails in the
outgoing arena. The user will see that he or she is hacked or infected and
Sure. But you could equally well say "if everyone runs software that
tracks outbound port 25 connections they'll be able to see that a
spammer is using their machine for spam". In fact, I'm sure such
software exists.