<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>
Date   : Mon, 12 Aug 1985 22:13:00 MDT
From   : Keith Petersen <W8SDZ@SIMTEL20.ARPA>
Subject: Ward Christensen comments on protocol wars

The following is a message from Ward taken from Compuserve's CP/M sig.

#:  143664 S0/Communications
    10-Aug-85 16:05:46
Sb: #143226-#protocol wars
Fm: Ward Christensen 76703,302
To: Irv Hoff 72365,70 (X)

Then you said 'The new "K" addition opens the door for expanding into
tomorrow's technology, by replacing the "K" with say a "N" for 9600
baud use with say 2k or even 4k packet sizes.".  Sorry, I frowned so
much reading this that my eyebrows touched in the middle.

Thank goodness Charlie Strom jumped in saying " Since these characters
are not in the error-detection routine ... don't you have any concern
that adding still more possibilities will make the protocol more
fragile?"

You blew me away with your reply: "None whatsoever, Charlie, you (and
Ron Fowler) are likely overlooking the fact those protocol characters
are never sent once the first record is sent.  There is no more
liklihood that would be "more fragile" than the current method of
requiring a "C" or "NAK" prior to the first record being sent, in my
evaluation."

Gad, I'm getting to the point where no matter how much I want to do
all this via CIS, I'm tempted to call you on the phone and scream!
How can you not see, that the addition of characters that aren't
checked, is not opening the door to more and more errors!  I curse
myself for initially doing single-char ACK/NAK/EOT.  Then the opening
<nak> was invented by someone (who?) - and not a bad idea.  "C" naking
was a hack, and obviously it doubled the chances of having the
"protocol selection" hosed up, i.e. a 'glitch' that looks like a C or
a NAK; now a glitch that looks like a C or nak or K, is bad.  You seem
to think adding "N" and more chars is "perfectly OK". If you'd said "I
realize 'K' naking is a HACK, and while it opens the door to more
problems, the benefits far outweigh them", THEN I'd be on your side,
and Fowler and Homchick wouldn't be calling me to beg me to take the
"anti-K side".

I am willing to "bless the K-nak", as yet another hack whose side
effects are not significant enough to warrant throwing it away.
Please just don't smoke whatever you smoke when you rationalize that
it is GOOD, and extendable into the future!  It may not make YOUR
brain hurt, but it sure makes MINE hurt!
<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>