<<I agree that FTP should have better process handling error messaging.
 At the same time, lets look at what FTP is made up of.  Remeber the old
Net/IPC stuff that let you set up transmission networks.  Do a listprog
on FTP.ARPA.SYS sometime.  You'll find some old friends.  May be if the
messaging or running FTP from a program is not good for you, try Net/IPC.
 I mean, you are already using it, in FTP.>>


I agree-let's show some appreciation for that "old friend". Our primary
production software (at least the HP3000 end of it) is based on NetIPC.
After all, NetIPC was "sockets" long before BSD sockets came to the 3000.
And NetIPC has long had capabilities, like conditional acceptance of an
incoming connection, that are just starting to show up in the "other"
kind of sockets.

FWIW, I'm convinced that the problem with ftp last year (year before
last?) where it would drop the connection if a minute went by without any
data being received was directly attributable to the author of the ftp
software not thoroughly researching the NetIPC sockets on which s/he was
basing the software, and thus overlooking the default 60-second timeout
on IPCRECV.

Steve