I'm starting to really dislike the whole format file approach. But
I can't think of anything better at the moment.
The format file approach is painful to me because it's a lame little
programming or extension language in an era when there are many better
choices available.
I had often thought of replacing that mechanism with a Tcl interpreter
so that the functions performed by the format files could be implemented
in a more clear way with simple Tcl scripts. Having it be part of the
the (n)mh command means easy access to the properties (headers, etc)
associated with the original message.
Adding this capability really opens up a wide range of possibilities,
and making the "Right Thing" easier for many end-users, without getting
tied up with the arcane syntax of the existing capability.
IMHO,
louie