Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!lll-lcc!pyramid!pesnta!phri!roy
From: roy@phri.UUCP
Newsgroups: comp.os.minix,comp.mail.uucp
Subject: uucp mail inefficiency
Message-ID: <2629@phri.UUCP>
Date: Thu, 2-Apr-87 12:09:08 EST
Article-I.D.: phri.2629
Posted: Thu Apr  2 12:09:08 1987
Date-Received: Sat, 4-Apr-87 19:48:05 EST
References: <7326@boring.mcvax.cwi.nl>
Reply-To: roy@phri.UUCP (Roy Smith)
Organization: Public Health Research Inst. (NY, NY)
Lines: 27
Keywords: uucp uucico kermit
Xref: utgpu comp.os.minix:486 comp.mail.uucp:398
Original-Subject: Re: uucp source copyright status - IMPORTANT

In article <7326@boring.mcvax.cwi.nl> jack@boring.UUCP (Jack Jansen) writes:
> one of the gross misfeatures of most uucp's is [...] for small files, the
> overhead is enourmous (two files transmitted for each tiny mail message)

	What does uucp have to do with mail?  Uucp is a transport protocol
which only knows how to transmit files and perform remote execution of
commands.  Uucp itself doesn't know from mail at all.  It would be trivial
(or at least straight-forward) to design a batched mail system which ran on
top of uucp -- take all the messages bound for a given site and batch them
into a single data file and a single XQT file for transmission.  You could
even compress the files before sending them (just like news does).  I'm not
sure if it's worth the effort, but it could be done.  Come to think of it,
you could probably use kermit as the mail transport protocol instead of
uucp; modern kermit servers can transfer files and execute remote commands
as well as uucp can.

	My guess is that the single most cost effective way to increase the
efficiency of uucp is to make sure the comm port you use has DMA input and
output.  It doesn't matter much if you're running kermit, zmodem, uucp,
SLIP, or just plain "cat > file"; a steady stream of input, even at 1200
baud, on an interrupt-per-character port is going to swamp most machines.
-- 
Roy Smith, {allegra,cmcl2,philabs}!phri!roy
System Administrator, Public Health Research Institute
455 First Avenue, New York, NY 10016

"you can't spell deoxyribonucleic without unix!"