This is indeed a challenge. At the moment TechMail-PEM deals with this in the
following way:
When a message is first displayed to the user, it is shown with all of its PEM
headers and other overhead intact. Across the top of the message a series of
buttons are displayed (for refiling, replying etc.). One of these buttons
invokes the PEM processing. Upon successful completion of PEM processing a
dialog box is displayed with the DN of the originator. This dialog cannot be
faked by anything in the message itself.
If the PEM message was nested, the nested message will now appear in
the message window (if there is no nesting, the de-enhanced message is
displayed with the PEM headers gone). The user can again hit the "PEM"
button to do the next level verification.
However, I have received comments from people telling me that first
displaying the enhanced message is "ugly" and that I should
automatically do PEM processing when I know I have a PEM message in
the window (I do know). Of course this raises the question of how to
deal with nested messages in this circumstance.
-Jeff