An interesting tech-support query came in a few days ago. Someone who
had been using MS-DOS Kermit for years and years, even on 32-bit Windows
(tsk tsk) found, when trying to move the application to Windows XP, that
MS-DOS Kermit no longer worked right. Of course this was to be expected
and our advice, predictably, was to change to Kermit 95 for all the
well-known reasons:

http://www.columbia.edu/kermit/msk95.html

Yes, said the user, but Kermit 95 does not have a REMOTE MESSAGE command,
and my application depends on it. Wow, I had totally forgotten about that
command, and never knew anybody had ever used it. It sends a short text
message from the Kermit client to the Kermit server, presumably for display
on the server's screen. Evidently some sort of device exists, manufactured
long ago, that uses this protocol creatively: the client uploads a file and
then sends a remote message in a specific format to tell the server to
install the file as some kind of upgrade or patch to the server-side
application.

It would be easy enough to add this command to C-Kermit, but the user needed
it to be in Kermit 95 and putting out new K95 releases is a rather big deal,
what with all the packaging, publishers, distributors, and expense involved.

But Kermit 95 (like C-Kermit and MS-DOS Kermit) includes its own built-in
programming language. Can a small piece of Kermit client/server protocol
be implemented in this language? Yes indeed, and here it is:

ftp://kermit.columbia.edu/kermit/scripts/ckermit/rmsg

This is far from a real Kermit protocol implementation, but shows that it
would be possible: encoding, formation, and transmission of packets,
checking for acknowledgement and retransmitting if a valid acknowledgement
doesn't arrive in time, etc. Similar techniques were used to implement
TAP (alphanumeric paging protocol):

ftp://kermit.columbia.edu/kermit/scr...rmit/alphapage

- Frank