Subject: Re: Exploring the limits of free software: Cygnus, and GPL
From: "Stephen J. Turnbull" <>
Date: Tue, 25 May 1999 11:28:05 +0900 (JST)

First, let me say that I apply the usual disclaimer of warranty of
merchantability or usefulness for any purpose to the BSD license; my
purpose is not to advocate it.  I merely want to maintain it as a
useful example.

But I am an economist, not a businessman.  I am not interested, except 
academically, in how to write a license to optimally compromise the
goal of freeness with the goal of making maximum profit.  (NB:  the
phrasing presumes that freeness does to some extent constrain profit;
maximization takes place within that constraint.)

I am equally interested, academically, in how to write a license to
optimally compromise the goal of freeness with the goal of maximizing
social benefit.  Given the need and greed of human beings, _it is not
clear that purely free software is the mechanism to accomplish this_,
despite the advocacy of many brilliant and ardently public-spirited

It's obvious that BSD is not a way to make money, unless you are an
effective cartel (the X Consortium, which is evidently regretting its
mistaken judgement as to its effectiveness) or it's a contractual
requirement (as in the case where the NSF funds your study).  But it
is quite plausible that BSD, or something close to it, may be the
social optimum, given the spirit of fun and generosity that is also
part of the human condition.

So I need to make careful distinctions here, some of which (as Russell
advocates) are unnecessary for some purposes.  I also think that they
have ethical implications (a la GNU's deprecation of the word
"piracy"), but that's a problem for some other venue (mostly 'cause I
haven't worked them out yet).

Here's the kicker: if I figure out how to maximize social benefit, I
will probably be able to do the profit maximization (and any
combination in between) as a corollary.  And I'll GPL[1] it, if so; I
am afraid people will try to give me thousands of dollars for advice
they could read on my Web site anyway ;-)

>>>>> "rn" == Russell Nelson <> writes:

    rn> Stephen J. Turnbull writes:

    >> The BSD code incorporated in your program is _not_ your
    >> property in any normal sense.

    rn> Sure it is.  I control who gets a copy of it.  Ownership in
    rn> the intellectual property world means control.

No, you don't control who gets copies.  _Not the BSD code_.  Only your
"improvements:" the BSD code itself is presumably still available from
wherever _you_ got it.  You seem to be confusing ownership of a
particular copy of the code with ownership of the code.  Both are
rather slippery concepts.

Richard Stallman does not have the right to come and inspect the copy
of Emacs I am using to write this post.  _It is mine._
(Philosophically speaking; I am not a lawyer.  But clearly I have the
right and power to destroy it; he may not demand that I transfer it
back to him---that is a physical impossibility.  Of course, some
licenses do attempt to enforce similar provisions w.r.t. improvements,
etc.  But those apply to the code itself, not to any particular copy!)
He does have the right (assigned to the FSF) to insist (and have the
state enforce) that should I try to redistribute further copies I obey
the terms of the GPL.  _But that is a contractual right_, not
ownership.  He does have the right to insist that if my team of
monkeys should produce, not the works of Shakespeare, but the source
code contained in even one of the files of Emacs, that I not
distribute it in any form (except under the terms of the GPL, I think;
this is one of those "not Homer but another man by the same name"
cases).  _That_ is a property right conferred by the copyright; it
requires no contract between me and the FSF.

    rn> This has definite implications for a free software business.
    rn> I view a significant portion of the value of the code I
    rn> distribute as advertising.  If someone else gets to restrict
    rn> that advertising, I lose.

How are they restricting it?  I restrict your advertising because I
don't use any of your code (as far as I know, and currently; I used
to).  You lose.  Tough.

Are you telling me that you think you have the right to free ride on
their distribution activities?  Surely not.  You want the right to
take advantage of their activity, write it into the license so that
they can choose whether to allow you to do that (in return for their
use of the code, thus not a free ride).  But if you don't explicitly
contract for that service, I don't see how ownership of the code gives
you the right to it.

    rn> Worse, the BSD license says that I have to acknowledge the
    rn> source of the code that I'm using, but it doesn't say that I
    rn> have to retain the quality of it.  I don't want someone using
    rn> my brand on code I haven't written.

Both advertising and quality control are excellent reasons for using
something other than the BSD license.  Neither is relevant to the
issue of ownership.

It is theoretically possible that a software "seed" planted in
proprietary "ground" (via the BSD license) will be more fruitful in
_end-user_ value than the same seed planted in the less nutritious (in
financial resources) "soil" of copyleft (choosing an extreme example
for illustrative purposes).  If that should turn out to in fact be the
case, I might very well choose to donate my code in BSD fashion (I
have no commercial interests, because I am a hog snarfling from the
public trough: ie, a civil servant).  I would regret the loss of
advertising (I'm a bit vain), and also the losses incurred by people
using inferior versions of my code---but the assumption is that on
balance those are outweighed by having more people use my code in some 
form---not limited to future programmers.

_As the owner of that code_, I would be satisfied with my compromise,
because it maximizes social contribution.

[1]  This is a non-trivial promise; I won't be able to publish it in a
refereed journal if so.

University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
What are those two straight lines for?  "Free software rules."