*BSD News Article 92943


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!news.mel.connect.com.au!munnari.OZ.AU!news.ecn.uoknor.edu!feed1.news.erols.com!news.maxwell.syr.edu!newsfeed.direct.ca!nntp.portal.ca!cynic.portal.ca!not-for-mail
From: cjs@cynic.portal.ca (Curt Sampson)
Newsgroups: comp.unix.bsd.freebsd.misc,comp.unix.bsd.netbsd.misc,comp.unix.bsd.openbsd.misc,comp.unix.bsd.misc,comp.unix.bsd.386bsd.misc
Subject: Re: *BSD Unification?
Date: 6 Apr 1997 10:08:28 -0700
Organization: Internet Portal Services, Inc.
Lines: 92
Message-ID: <5i8lac$9r2@cynic.portal.ca>
References: <860029226.1885@dejanews.com>
NNTP-Posting-Host: cynic.portal.ca
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:38577 comp.unix.bsd.netbsd.misc:5760 comp.unix.bsd.openbsd.misc:54 comp.unix.bsd.misc:2946 comp.unix.bsd.386bsd.misc:1219

In article <860029226.1885@dejanews.com>,  <prw@cyberwar.com> wrote:

>        I have heard people say that the presence of multiple BSD
>systems is comparable to the fact that there are multiple Linux
>distributions.  I disagree wholeheartedly with this notion.  The Linux
>distributions share the same kernel, while with BSD there are now
>three separate teams working on three separate kernels.  Is this not a
>waste of resources?

You make it sound like the kernel is the hard part, and putting
together the userland after that is easy. That's certainly not the
case; just putting together all the programs for a distribution,
much less testing and debugging them, is an enormous amount of
work. Now isn't having nine (or however many it is) different
organisations doing this a `waste of resources'?

And even in the kernel, we see things like to completely separate
and independently maintained drivers for the DEC Tulip Ethernet
chip. Is this not just as much a `waste of resources'?

Having all of these different distributions makes life difficult
for commerical software developers as well. There have been three
NetBSD kernel releases in the last couple of years (1.1, 1.2,
1.1.2). FreeBSD has had perhaps a couple more. How many different
kernels have been released in the last couple of years in Linux
distributions?

And the same goes for userland software, of course. You can't ever
really be sure which version of bash or awk you're going to get
when your software is running on a `Linux' system. This leads to
statements like the following one in the ReadMe file for Adobe
Acrobat reader:

: Adobe develops and tests this Reader on the Yggdrasil 1.2.13 Fall 95 release
: configuration (1.2.13 Linux kernel, XFree86 3.1.2 and the following system
: libraries: libc 5.0.9, libm 5.0, libdl 1.7.10). However, we believe and have
: heard from users that it runs well on the following system configurations:
: 
:   Redhat Linux 3.0.3 and kernel 2.0
:   Slackware and the 1.2.13 kernel
:   XFree86 3.1.2G X server
:   XFree86 3.1.2 server, using S3 Trio64 hardware
:   Linux 2.0.4 ELF system
:   Linux 2.0.21
:         XFree86 3.1.2G (beta; X11R6.1)
:         libc 5.3.12
:         ld.so 1.7.14
:   Debian 1.1 with kernel upgraded to 2.0.21
:   Slackware 3.1
:   2.0.18 and a hacked-up Slackware 3.0 install
:   Red Hat 3.0.3 Picasso system (Linux 1.2.13 kernel)

Deciding which version of the dozens of versions of the kernel and
the system libraries you want to run is hardly any more friendly
than deciding which *BSD system you want to run!

And how much work has been duplicated between the Slackware and
Debian and Red Hat and Yggdrasil folks? How many times has a
sophisticated package manager been written? How many times has an
installation procedure been written? How many man-hours have been
lost testing these duplicated installation procedures?

I'm not taking the Gnu/Linux crowd to task for this, and I hope
nobody takes this as a flame. But a close look reveals that the
Gnu/Linux folks duplicate effort very frequently and at every level.
That Gnu/Linux is much more popular than BSD is due to other factors.

Now all that aside, I'd also like to point out, in I hope a
non-inflamatory way, that asking the current developers to champion
merge projects is probably not going to get anywhere. It's not as
if the *BSD developers are all sitting around saying to themselves
`What shall I do? I've nothing to work on.' If someone wanted to
actually start doing the legwork to see what could be merged, ran
around talking to the folks in the various camps, arranged for a
common CVS repository, and actually started merging code, we'd see
some progress. But I've seen very few people with a real interest
in doing this. Most of the really pro-merge crowd seem content to
post to the newsgroups every once in a while asking why we shouldn't
merge. Please don't take that as a flame, but as a guide to what
you can do to help merge things.

Oh, and if you don't have the programming skills to do that part
of the merge, there are plenty of other things you could work on.
Expanding the FreeBSD handbook to apply to NetBSD as well would be
an excellent start, and would probably bring a lot of the small
incompatabilities to light, where many could be fixed.

cjs
-- 
Curt Sampson    cjs@portal.ca	   Info at http://www.portal.ca/
Internet Portal Services, Inc.	   Through infinite myst, software reverberates
Vancouver, BC  (604) 257-9400	   In code possess'd of invisible folly.