home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
kermit.columbia.edu
/
kermit.columbia.edu.tar
/
kermit.columbia.edu
/
newsgroups
/
misc.19941221-19950208
/
000033_news@columbia.edu_Tue Dec 27 20:24:32 1994.msg
< prev
next >
Wrap
Internet Message Format
|
1995-07-31
|
3KB
Received: from apakabar.cc.columbia.edu by watsun.cc.columbia.edu with SMTP id AA14857
(5.65c+CU/IDA-1.4.4/HLK for <kermit.misc@watsun.cc.columbia.edu>); Tue, 27 Dec 1994 15:42:39 -0500
Received: by apakabar.cc.columbia.edu id AA11304
(5.65c+CU/IDA-1.4.4/HLK for kermit.misc@watsun); Tue, 27 Dec 1994 15:42:38 -0500
Newsgroups: pdx.computing,comp.protocols.kermit.misc
Path: news.columbia.edu!panix!news.mathworks.com!hookup!swrinde!howland.reston.ans.net!ix.netcom.com!netcom.com!kientzle
From: kientzle@netcom.com
Subject: Re: Kermit / xyz modem showdown results?
Message-Id: <kientzleD1HKox.Ktw@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
References: <3d4olb$hju@sun.lclark.edu>
Distribution: or
Date: Tue, 27 Dec 1994 20:24:32 GMT
Lines: 46
Apparently-To: kermit.misc@watsun.cc.columbia.edu
In article <3d4olb$hju@sun.lclark.edu>,
John Miller <miller@sun.lclark.edu> wrote:
>Awhile back Chuck Forsberg announced such a showdown, and I saw a followup
>article in the Oregonian. The Oregonian's article was more entertaining
>than informative, and Forsberg did not seem to publish any hard results
>on the net.
Chuck does have a gift for generating publicity, doesn't he?
;-) Unfortunately, I've never seen a `showdown' that produced any
really useful information. As Mark Twain said: there's lies, damned
lies, and statistics.
If you read between the lines of the glossy graphs and charts
used to support the `my protocol is better than yours' wars, you can
get some real laughs. For example, one chart I saw proclaimed one
protocol `best' due to an advantage over a competitor of only .3%!
Another chose a very particular file that had certain properties that
caused it to transfer unusually quickly under one particular protocol
(but not a competitor's, of course). Both of these comparisons were
done by generally well-respected individuals.
>Anyone have any comparative results between latest versions of these packages?
>We are running the latest kermit on a DEC Alpha, but have the older versions
>of sz, etc.
Latest versions of which packages? I've seen 10-20 percent
variations in speed between different implementations of the same
protocol on the same equipment transferring the same file. (In one
case, a data scope was used to verify that the implementations were
functioning identically. Apparently, one was just 10% less efficient
at managing the serial and disk I/O on that hardware.) If that kind
of difference is important to you, then simply specifying which
_protocol_ to compare is pointless.
My two cents: Kermit and ZModem are sufficiently close in
speed (given good implementations of each) that ease of use and
reliability are THE most important issues. Speed is no longer a major
consideration. If you want faster transfers, buy a faster modem and
a faster computer. Still not fast enough? Rent a T1 line.
In fact, I regularly use a protocol that's much slower than
either Kermit or ZModem simply because I have an implementation that's
extremely easy to use. The minutes I save not having to remember
which settings to fiddle more than make up for the seconds of extra
transfer time.
- Tim Kientzle