*BSD News Article 14882


Return to BSD News archive

Xref: sserve comp.os.linux:35384 comp.os.386bsd.questions:1836
Newsgroups: comp.os.linux,comp.os.386bsd.questions
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!haven.umd.edu!uunet!world!jrs
From: jrs@world.std.com (Rick Sladkey)
Subject: Linux and 386BSD
Message-ID: <JRS.93Apr21200843@lepton.world.std.com>
Sender: jrs@world.std.com (Rick Sladkey)
Organization: The Internet
Date: Thu, 22 Apr 1993 00:08:43 GMT
Lines: 84

There has been a lively cross-discussion within the Linux and 386BSD
groups largely concerning distribution licenses.  The focus has been
the difference between the Linux GPL-style license and the 386BSD
BSD-style license.  I feel that this discussion has been for the most
part counter-productive to both groups and I want to see if we can
learn something from it and move on.

People who feel that the "other" license style is absurd, stupid,
brain-damaged, etc. are denying the fact that many intelligent free
software developers have thought long and hard about the issues and
have often come to different conclusions about what license is best.

If it were so clear cut that one could change someone's mind by
ridiculing their position or restating the obvious then everyone would
be using the same license.  Since they are not, it is clear that
developers have differing goals and that many of the issues are still
vague or depend on premises which have not yet been tested by time.

The BSD-style license has many advantages.  It preserves copyright.
It ensures attribution.  It provides great flexibility to the
developers of derived works.  Many intelligent people have carefully
chosen to distribute their works under this license.  They knew what
they were doing, they knew about the potential for abuse, and they
would do it again.

The GPL-style license also has advantages.  It too preserves
copyright and ensures attribution.  It greatly increases the chance
that improvements will also be free and incorporated back into the
distribution.  Many practical non-zealots have chosen to distribute
their works under this license.  Usually they chose it not because
they are looking forward to a post-scarcity society but because they
believe it makes sense for their software today.

What I believe are misconceptions abound in the discussion.

Misconception: The 386BSD developers do not understand the potential
for abuse with the BSD-style license.

Of course they understand.  Their goal is not to ensure that all
derivatives of their software are free but to make research
contributions to the Operating Systems field.

Misconception: The Linux developers steal free code from 386BSD, make
one or two changes and then attach the GPL to it.

In rare cases where the author of the changes has explicitly stated
this and his/her changes are clearly marked, even then the GPL only
applies to the changes.  However, in the vast majority of cases,
non-Linux software and the Linux changes are distributed using the
unmodified license of the original.  This is just common courtesy.

Misconception: The 386BSD developers hate the GPL and scorn software
that uses it.

Hardly.  Without GPL software there would have no 386BSD at all.
However, they would prefer that the kernel and C libraries contain
only BSD-style licensed code.  This is a natural consequence of the
fact the all of the 386BSD kernel and C libraries currently have this
license.

Misconception: Commercial programs and binary-only free programs
automatically violate the Linux library license.

Since Linux uses shared libraries, commercial developers are free to
sell fully-linked binaries which liberally use the C library as long
as the binaries don't use the static libraries.  The only piece of
Linux library code that is included in such a binary is crt0.o and a
static version of load_shared_libraries().  These are not subject to
the restrictions of the LGPL.

The main reason there is friction between the 386BSD and Linux
communities is because they share the same audience: PC owners looking
for an inexpensive Unix-clone.  It is rather like the perpetual war
between `vi' and `Emacs' users.  However, this is no reason for
name-calling and flippant remarks.

The 386BSD group is a talented and colorful collection of people
working on a fine software product.  I would like to think the Linux
group could be similarly described.  There is less cross-fertilization
than might be imagined because both groups enjoy doing things from
scratch as a learning experience.  Independent reinvention is valuable
because the opportunity for real change occurs very early in the
discovery process.  It may sound inefficient but it sure is a lot of
fun.