kenton@space.mit.edu (Kenton C. Phillips) (06/23/89)
I would like to make a couple of suggestions about handling signatures,
remote and local. First, I would like to express my dislike for the idea
of a system-wide database of "local" vs. "remote" machines. I do *not*
want to have to deal with yet another table of machines, which has to be
changed every time something on our network changes. And anything which
encourages users to complain about "why doesn't it treat machine X the
way I want it to?" is not going to be installed on systems under my control.
Also, I would like to suggest that expanding aliases to include
such things as local vs. remote is a bit excessive. ("But for some mail I
want to use emacs, and for some I want to use vi. Why won't it let me
define whose mail gets which editor?" :-) )
There. Now that the flames are out. Second.
I vote in favor of having a headers-edit-screen option to allow selecting
one of three options, (i) local, (ii) remote, or (iii) edit signature.
(Actually, you don't need that last - just put a signature in your
messages, and don't append a signature file.)
Lastly, I would like to offer a possible solution (which could be used in
conjunction with headers option solution mentioned above) which was
suggested to me by a colleague (Don Alvarez), to wit:
If the first line of the .signature file is of the form
#!/bin/sh
then treat it as a shell script, with the To: and Cc: addresses as standard
input, and take the standard output as the signature to be appended
to the message.
Why is it that this seems to make so much sense? Am I missing something?
I would still want to be able to easily select my standard local or remote
signature, though.
--
Kenton C. Phillips
kenton@space.mit.edu