home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
kermit.columbia.edu
/
kermit.columbia.edu.tar
/
kermit.columbia.edu
/
newsgroups
/
misc.19950929-19951130
/
000044_news@columbia.edu_Fri Oct 6 19:00:32 1995.msg
< prev
next >
Wrap
Internet Message Format
|
1995-12-25
|
3KB
Received: from apakabar.cc.columbia.edu by watsun.cc.columbia.edu with SMTP id AA14598
(5.65c+CU/IDA-1.4.4/HLK for <kermit.misc@watsun.cc.columbia.edu>); Sat, 7 Oct 1995 20:45:10 -0400
Received: by apakabar.cc.columbia.edu id AA15488
(5.65c+CU/IDA-1.4.4/HLK for kermit.misc@watsun); Sat, 7 Oct 1995 20:45:08 -0400
Path: news.columbia.edu!sol.ctr.columbia.edu!news.kei.com!news.mathworks.com!tank.news.pipex.net!pipex!howland.reston.ans.net!torn!newshost.uwo.ca!usenet
From: David Maw <maw@gaul.csd.uwo.ca>
Newsgroups: comp.protocols.kermit.misc
Subject: tech info on kermit?
Date: Fri, 6 Oct 1995 15:00:32 -0400 (EDT)
Organization: The University of Western Ontario, London, Ont. Canada
Lines: 12
Message-Id: <Pine.SOL.3.90.951006145708.19592A-100000@obelix>
Nntp-Posting-Host: obelix.gaul.csd.uwo.ca
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Sender: maw@obelix
Apparently-To: kermit.misc@watsun.cc.columbia.edu
I am looking for techical papers/docco on how kermit works.
I have to write a paper on how Kermit handles errors, and have had little
luck finding information on WWW and Gopher.
If you can point me in the general direction please post to this group or
email me at: maw@obelix.gaul.csd.uwo.ca
Thank you,
----
David Maw
-----------------
"Errors" cover many sins and the responses vary depending on the
particular transgression involved. The formal statement of the Kermit
protocol is the book "Kermit, a file transfer protocol" by Frank da Cruz,
paperbound for under US$40, ISBN 09-32276-88-6. Within that framework each
implementation has room to select whether to respond to some conditions or
be patient, and how strongly to complain (sliding windows heurstics).
A general soft rule in networking is damaged packets are not packets
and are disregarded entirely as if they never occurred. Timers then step in
to recover from outages. Thus damaged packets become empty time on the wire.
Kermit is an ACK/NAK protocol and the intelligent and selective use of NAKs
can shorten waiting times (NAKs are timeout preventers). Another soft rule
of protocol stacks is there is room for local inventiveness on each end as
that end selects what to do when things are not in good shape. These are
heuristics that obey the formal rules and keep state machines sane, but
generally try to shorten recovery sequences. A last soft rule of stacks is
never try to outguess or even "help" the other side, because this side does
not have all the information and thus may well make matters worse for a while.
I implement this last soft rule as barbarian style actions in MS-DOS Kermit:
do it successfully or everything fails and we could care less why it failed.
Joe D.