PE> Are zmodem send and receive two separate objects, both with one
PE> operation, perform()?  [this is what I think is the solution at this
PE> exact instant of time].

RS> The most obvious consideration is the question of bidirectional protocols.
RS> If you dont allow for that in that design stuff you are considering now,
RS> that can rule out allowing that in the future without a complete change.

PE> The zmodem send and receive have some code that is common to both.

RS> Not relevant to what it looks like to the higher level code.

Excellent points, both which radically change my original
intentions.  I am now going to be considering this as an
object that must be able to be used on a system where two
computers are linked by modems, both of them use internal
arrays, and both of them have agreed "nicknames" for these
arrays, and every now and again, they have a mechanism for
requiring updates of their own array (and possibly updating
the other person's array [different one] at the same time).

And of course I shouldn't have been trying to make the low
level commonality influence the higher level interface!!!
BFN.  Paul.
@EOT:

---