Return to BSD News archive
Xref: sserve comp.os.386bsd.questions:6645 comp.os.386bsd.misc:1416
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!haven.umd.edu!darwin.sura.net!howland.reston.ans.net!usenet.ins.cwru.edu!eff!news.kei.com!world!ksr!jfw
From: jfw@ksr.com (John F. Woods)
Newsgroups: comp.os.386bsd.questions,comp.os.386bsd.misc
Subject: Re: [ANSWER] What is *BSD?
Message-ID: <34941@ksr.com>
Date: 9 Nov 93 12:51:47 EST
References: <jmonroyCG7wED.77x@netcom.com>
Sender: news@ksr.com
Lines: 129
jmonroy@netcom.com (Jesus Monroy Jr) writes:
> [A big long screed that, as usual, appears to refer to an alternate universe]
Here are some alternate answers to this question:
>>> What is
>>> freebsd, 386bsd, bsd386, netbsd, bsd4.4 ...
4.3BSD
The popular UNIX variant by the Computer Systems Research Group
(CSRG) of the University of California at Berkeley (UCB). Based
a long time ago on "UNIX 32/V" (Bell Labs' Version 7 for the VAX),
it had evolved into a substantially different creature, but still
had key pieces of AT&T code buried in it (chiefly in machine-dependant
kernel modules).
NET/2
Also known as net-2, or "Networking Release 2". This was a release
of a late version of 4.3BSD which had the AT&T-written code removed,
released chiefly to make public the latest version of the networking
code. NET/2 formed the basis of several ports, some of which are
seen below, and is available for FTP from many places, including
ftp.uu.net (in /packages/bsd-sources/). (NET/1 was a late version of
"4.3BSD-Reno", and did contain AT&T code.) Note that there is a suit
filed by USL against BSDI and UCB over this code, alleging copyright
infringement and trade-secret violations. The case has yet to come
to trial, but in pre-trial motions, USL hasn't generally fared as well
as they might have hoped.
BSD 4.4 (or 4.4 BSD)
The latest, and possibly last, release of UNIX from UCB. The CSRG
is being disbanded, as the money for OS research is drying up.
Depending on the outcome of the USL vs. UCB suit, there may be a
"4.4 BSD Lite" release which will be 4.4 BSD without AT&T code;
this assumes that any members of the now-disbanded CSRG group feel
like going to the effort then. Until then, you need an AT&T
source license to get 4.4BSD, just as you did to get 4.3BSD.
BSD386
A port of NET/2 offered by BSDI. Most of the founding members of
BSDI were members of CSRG; the initial BSDI port filled in the
missing pieces of NET/2 with work by them, William and Lynn Jolitz,
and several other people on the Internet. BSDI sells a commercially
supported OS with source for about $1000, with cheaper binary-only
licenses available. They have contributed some of their work to the
general BSD-porting community, but most of it they hold proprietary
(so they'll have something to sell).
386BSD
A port of NET.2 done by William and Lynn Jolitz, written up in a
series of articles in Doctor Dobbs Journal. The official release
was given a version number of 0.1; it was supplemented by a series
of patches, maintained by an independant group that came to be called
the patchkit coordinators. 0.2 has been promised for a long time,
but nothing has been heard about it for some time, and most people
have given up waiting; the Jolitzes became increasingly erratic in
public postings and then disappeared from the public eye.
NetBSD
The first group to decide the patchkit process was unwieldy and that
386BSD 0.2 wasn't going to happen in this life. The original 386BSD
port had a lot of needless dependancies on the 386 architecture in
areas of the code that should have been machine independant; most of
that crud has been replaced, and NetBSD currently runs on HP300s,
Amigas, SPARCs, some Macintoshes*, (a couple of other architectures
I forget offhand), and of course the 386. The first release of NetBSD
was numbered 0.8, the currently-available release is 0.9, and if
you're adventurous, the work-in-progress version (NetBSD-current) is
available for FTP or SUP from various places.
MacBSD *
A port of NetBSD 0.8 to some Macintosh models. Since they're directly
accessing hardware, it doesn't (yet) run on all Macintoshes (since
Apple moves around the hardware definitions freely). Its source
appears in the NetBSD-current tree, but I'm not sure if that copy
builds yet (it may not have tracked all the NetBSD changes); it's
also available separately (and buildable, of course).
FreeBSD
What the patchkit coordinators decided to call their work when *they*
decided the patchkit process was unwieldy and that 386BSD 0.2 wasn't
going to happen in this life. FreeBSD 1.0 is the current version,
after a sequence of four trial versions (alpha, beta, gamma, epsilon).
A Digression on NetBSD versus FreeBSD:
Innumerable attempts have been made to describe the difference between
NetBSD and FreeBSD; NetBSD is dedicated to a "stable" kernel, and
FreeBSD is oriented toward research -- or is it the other way around?
NetBSD incorporates and experiments with bizarre, new, untested
features before FreeBSD, except of course for the bizarre, new,
untested features that FreeBSD has tried first. NetBSD is lemony
fresh, FreeBSD leaves a fresh minty aftertaste. And so it goes.
The difference between NetBSD and FreeBSD is the people who work on
each of them. There are some people in the NetBSD working group who
are polar opposites to some people on FreeBSD (and not even on any
particular issues, that I've seen), but there are also a substantial
number of people working in *both* groups. These people all work on
what they want to work on, and if they only work on one of the two
big variants, that's the one that gets their work first.
There are people in the NetBSD camp to whom machine independance is
very important; hence the NetBSD source base is ahead of FreeBSD in
having 386 dependancies ripped out. There are people in the FreeBSD
camp to whom ease of installation is important, so FreeBSD has a
nifty installation tool and NetBSD just sort of clunks along.
On the question of who wants stability ("I do, I do!"), I suspect
this "distinction" comes from a series of historical accidents:
one impetus behind the NetBSD camp was the "patch of the day"
circus the patchkit scheme had become; instead they preferred to
have occasional checkpoint releases that were reasonably stable
(hence 0.8 and 0.9); hence the misperception that "stability" was
an overriding goal for NetBSD as opposed to 386BSD/FreeBSD. However,
the NetBSD camp was making more substantial changes to the kernel
(for machine independance and general fixes) and those people who
regularly pick up the NetBSD-current sources often find that the
instantaneous sources don't build, which happened (somewhat) less
often with the patchkit releases, hence the misperception that
NetBSD was less stable than 386BSD/FreeBSD.
The above is written from the viewpoint of someone who uses NetBSD and who
has contributed a couple of minor fixes but isn't involved in the internal
politics of the development group; nor do I regularly post multi-kilobyte
screeds about driver software I can't release because aliens haven't used
the radio receivers in my head to give me permission yet.