The page "Kermit File Transfer and Management as an SSH Subsystem" at

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

recommends doing Kermit over SSH by setting up Kermit as an SSH
subsystem on the host and using SET HOST /PIPE to connect to the Kermit
SSH subsystem.

However, my host (an embedded device) only has the dropbear SSH daemon,
and cannot be configured to support Kermit as a subsystem.

However, I have found that using the following works just fine (from
"2.7. Using C-Kermit with External Communication Programs" on the page
http://www.columbia.edu/kermit/ckermit70.html )

set host /pty ssh -e none user@host

followed by scripting to take care of logging into the host and
invoking Kermit as a server.

Besides the work of creating the extra scripting, is there any serious
drawbacks to this approach, versus the subsystem setup? It seems solid
to me, and frankly I really don't want to bother with getting OpenSSH
working on my device just so I can run Kermit as a SSH subsystem, in
light of this approach which seems to work just as well.

Any comments or suggestions are appreciated!