*BSD News Article 5538


Return to BSD News archive

Newsgroups: comp.unix.bsd
Path: sserve!manuel!munnari.oz.au!uunet!snorkelwacker.mit.edu!bloom-picayune.mit.edu!athena.mit.edu!tytso
From: tytso@athena.mit.edu (Theodore Y. Ts'o)
Subject: Re: Catch What They're Saying About Us...
In-Reply-To: wjolitz@soda.berkeley.edu's message of 23 Sep 92 00:34:11 GMT
Message-ID: <TYTSO.92Sep23235422@SOS.mit.edu>
Sender: news@athena.mit.edu (News system)
Nntp-Posting-Host: sos.mit.edu
Organization: Massachusetts Institute of Technology
References: <19oe23INNqh0@agate.berkeley.edu>
Date: Thu, 24 Sep 1992 03:54:29 GMT
Lines: 43

In article <19oe23INNqh0@agate.berkeley.edu> wjolitz@soda.berkeley.edu (William F. Jolitz) writes:
>Check out the "Re: Linux or 386BSD: neither or both ??" thread in
>comp.os.linux. It's pretty amazing what's being said about 386BSD and
>the people who've worked so hard on it.

>Since I'm not one who likes to let disinformation be the order of the
>day, I've challanged the Linux people to put their money where their

A couple of points.  First of all, note that the few people that have
engaged in that thread do not represent the views of the majority of
Linux users/developers out here.  Most of us have just ignored most of
the thread as a waste of time.

Secondly, I don't think "disinformation" is a fair way to characterize
those postings.  They were merely observations by a number of people
that subjectively, on their machines, Linux felt faster than 386BSD.
Bill then fired off a completely unprofessional "You must not know what
you are doing", which kicked off personal attacks on both sides, and the
thread denegerated from there.  However, all this brouhaha doesn't
change the fact that a number of people have reported that when they
tried the two operating systems, they felt that 386 BSD was slower than
Linux.  Unless you have reason to believe that these observers were
out-and-out telling lies, calling their observations "disinformation"
is just not fair.

I will note that most of those people had relatively small amounts of
memory, and so perhaps Linux's use shared libraries were decisive.
Linux also supports copy-on-write when forking, which I understand 386
BSD does not do (my information on this may be dated; please correct me
*gently* if I am wrong), and that may have helped as well.  Linux also
has fewer abstraction layers than BSD does at this point, so that may
have helped a tiny bit as well.

It would be nice to get some hard performance data on the two operating
systems.  Unfortunately, I don't have room on my hard disk for 386 BSD,
so it would be hard for me to do the testing myself.  (Due to all of the
different variances of hard disk speed, amount of memory, speed of
processor, etc., you can probably only do fair tests by running the two
operating systems side by side.)
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Theodore Ts'o				bloom-beacon!mit-athena!tytso
308 High St., Medford, MA 02155		tytso@athena.mit.edu
   Everybody's playing the game, but nobody's rules are the same!