Date   : Thu, 10 Mar 1983 00:06:00 EST
From   : Keith Petersen <W8SDZ@mit-mc.arpa>
Subject: Rainbow 100 and file transfer

The "compatibility problems" between various programs which say they
use the Christensen protocol are probably caused by timing errors in
the program.  Some well-meaning, but misguided, person updating MODEM2
and MODEM7 changed the 10 second timing to 2 or 3 seconds about a year
or two back.  There was a VERY good reason for it being 10 seconds -
to allow for slower disk systems' time it takes to write to the disk.
This is particularly critical on mini-floppies - some of which have as
slow as 30 milliseconds track-to-track seek time (I know, I have one -
the Micropolis Mod II).  The result is (in some cases) that the
end-of-file byte from the sender is lost by the receiving end.  The
solution is simple - restore the timing back to the 10 seconds that
Ward Christensen and I agreed was optimal.  Those commercial programs
that don't work were probably based upon the public-domain versions
that were available at the time they were written - probably the ones
that had 2-3 second timing values.  If you contact the authors of the
commercial programs, they should be able to tell you the addresses of
where to patch the delays.

Your comments about MODEM7 having so many versions are well taken -
however (flame on) if the program had been written right in the first
place (and had been "clean" source code instead of the poorly
formatted version that was released originally) many of the revisions
would have been unnecessary.  It is inexcusable that the author of
MODEM7 REMOVED ALL OF WARD CHRISTENSEN'S COMMENT LINES in the file!
This made it almost impossible to service the program for updates/bug
fixes, and in my opinion, caused updating it to be more prone to
having bugs. (Flame off).

The present version of MODEM7 (MDM705) is finally in reasonable
condition and has MANY new and useful features - a lot of which are
especially useful here on the net (i.e., the terminal mode now allows
ALL control characters to be passed to the modem, allows toggling the
printer on and off, and has a buffer for the printer in case it's
slower than the incoming data on the modem.  Mind you, I'm not saying
it's perfect, but it finally in good enough shape that I've stopped
recommending MODEM2xx to newcomers.  With the configuration files it
is now no longer necessary to download the .ASM source file.  I
consider that a major step forward and I would hope that people will
not talk down this program simply because it has gone through so many
versions.  That would be throwing away a lot of good work!
--Keith <W8SDZ@MC>