Just what would we like to accomplish with an e-mail security protocol? Let's list some goals:

1. We would like to enable correspondents to send private messages to us without requiring separate security procedures for key exchange (and we would like to write back to such correspondents with the same ease).
2. We would like to allow correspondents to "sign" messages and thereby provide a reasonable assurance about the true origin of messages.
3. As a corollary of the first goal, we would like to have a reasonable assurance that the keys we believe to correspond with a certain person really are associated with that person (no spoofing of identities).
4. We would like the whole protocol to make as many limited computational demands as possible while obtaining these other goals.
5. We would like the whole application or system that implements our protocol to be transparent and user-friendly.

The last goal falls outside the scope of this tutorial, but it is not something to ignore when one gets to the actual programming and design.