The position RMS takes...

W. Yip weng at
Tue Mar 28 16:34:17 UTC 2000

Hi folks, I have attempted to consolidate my response to all your heartily
given contributions in one piece. Please do go through it and tell me what
you think of what I understand.

On Fri, 24 Mar 2000 17:53:45 -0800, David Johnson <david at>
>On Fri, 24 Mar 2000, Mark Koek wrote:
>> The GPL is dishonest, then?
>I didn't say that! Nowhere within it does it say it is not property,
>and plenty of places where it acknowledges it is guarding a piece of
>intellectual property. But I hear a lot of statements to the effect
>that "copyleft is not copyright" when in fact it is. People who don't
>believe in information as property, but turn around and use the GPL
>saying "no one owns it", are being dishonest, either with me or with
>themselves. I can only assume that they are being unintentionally

Copyleft is a method of copyright licensing. Thus 'Copyleft' is a method of
licensing. OTOH, Copyright is a proprietary (meaning property) set of
rights vested in the author (or employer) by law. These rights become
property because they are made *exclusive* to the holder, meaning no one
else can perform those listed acts without his permission. These are
property rights because the author can exploit these rights for their
market value by the transfer of these rights, whether through license or

Legally, licensing represents permission to the person. In this, OSS
licenses are no different from conventional licenses. However, the issue
with Open Source licensing is compounded by the following:

1) The (i) nature and (ii) extent of the terms/permissions in OSS licenses.
OSS Licenses are different from conventional licenses in that they 
(i) place *no* limits on the copying, and (ii) regulate/impose rules on
subsequent sublicenses by the licensee and (iii) seem to offer permissions
to *anyone* [unilaterally?].

2) The presence of contractual elements mixed with license elements. The
former comprises *obligations* as a result of agreement, while the latter
represents *permissions*. I find the resulting mix very difficult to
ascertain or separate.

>> RMS dislikes the notion of information as property.
>Then it is curious to me why he considers his own works, as
>demonstrated by his actions, to be intellectual property. Regardless of
>whether he considers himself and the FSF as owners or as mere
>caretakers, he has imposed upon his works terms and conditions that
>only owners are allowed to make. He claims to be giving his software
>away when in fact he is sharing instead.

[my proposition] The extremist position RMS can be consistent with
intellectual property. In short, I understand RMS to be against restrictive
(closed) licensing, but NOT against copyright.

I think his key ambiguity is not the word 'freedom' (which the adage
'liberty not price' explains) but instead, his (IMHO) clumsy use of the
word 'proprietary'. If RMS clarifies this word, I believe his position can
be consistent with my proposition stated above.

I must stress the legal concept of what is 'proprietary' is very different
from the word 'proprietary' defined in the GPL. For RMS and the GPL, a
'proprietary' license means any license that gives you less rights than the
GPL. Thus, RMS' use of 'proprietary' relates  to extent(s) of *permission*.
I surmise that RMS would accuse Microsoft of being 'proprietary' because
Windows 2000 licenses are restrictive (as compared with GPL), and NOT
because Microsoft owns copyright to Windows 2000. Such is the way RMS's
construes the word 'proprietary'.

OTOH, in law, it is the copyright that is 'proprietary', since it is
copyright that you actually *own* as an author. The license represents your
*permission* to what becomes lawful use. HOWEVER, the problem is the
license potentially takes on a proprietary (property) nature when there is
a contract (whether of service or sale). Faced with a contractual
agreement, and the obligation (from the licensor) that a licensee may
expect therewith, it thus appears possible to 'own' a license as property.
Indeed, some in this forum have incisively indicated that the UCITA (which
I am not familiar with) makes every software license a contract.
As a result, we can see that the legal notion of a license has become
*stretched*, and our problems really issue from this area.

Some points I have been thinking

a. Revocation - A *bare* license can be revoked anytime (eg. you ask guest
in your house-party to leave, and their failure to do so would make them
trespassers under law.) However, what about an OSS license? If an OSS
License is a contractual license, what sort of damages would revocation
bring, supposing revocation amounts to contractual breach? Would these be
restitutive, or would they compensate expectation losses? How to quantify?

b. 'Rights to software' - Are my 'rights' to the software given by contract
(obligations following an agreement) or by license (permission to perform
acts otherwise restricted by copyright?) A further question is whether
these rights (i) are 'attached to the *program*' [cf. s.7 and s.8 of OSD]
or whether these rights are attached to the *person*. Legally, a license is
personal, whether it be a bare permission (from licensor to licensee) or a
contractual agreement (two parties within privity of contract). Thus, I
would challenge the wording of the OSD on the grounds that it seems to
claim that rights subsist 'in rem' , that is, in the program itself. If the
OSD were correct, the 'rights' or 'permission to use' follows the program
wherever it may go, without any need for contractual agreement or execution
of license involving persons. But I contend surely this is not possible,
because it GOES AGAINST the very nature of a license under law [that it is
*personal*]. Alternatively, I would claim that an OSS License is not a
license at all, but a sui generis (new form) of right. If databases could
be protected by the sui generis Database rights, why can't the Open Source
community push for their own category of right? In so doing, we would evade
of the trouble of licenses.

>Is he being dishonest by doing this? In my opinion, yes...
>> Personally, I like this kind of pragmatic approach. Change the system
>> from within.

Granted that it would be overly radical and even revolutionary in an active
sense to go against copyright itself as information property, I would
concur with the somewhat passive stance that RMS takes (it is passive in
that it complies with Intellectual Property as far as legal mechanisms are

>The "viral" clauses may do some small bit of changing the system, but
>as a whole, Free Software sits squarely amidst the concepts of
>property. Essentially, you can't do anything with anyone else's
>property without their prior permission. Free Software gives you those
>permissions. Does taking down one's "no trespassing" sign count as a
>blow against real estate property? Hardly!

This other dimension affects the formal/functional aspects of Open Source.
One can argue that license like the GPL effectively 'licenses it ALL away'
the protection given by copyright law. It seems that the only benefit of
holding Copyright after adopting a GPL scheme is for the purpose of having
the 'standing' necessary to sue. 

And what of the grounds for suing? I have stressed the phrase 'violation of
GPL' is troublesome because it is ambiguous. Would the grounds for a
'violation' suit be (i) copyright infringment, and/or (ii) contractual

A. Copyright Infringement 

Is it even possible to infringe with the GPL? Does distribution
non-compliant with GPL terms invariably lead to infringment of copyright? 

Infringment may result as a result of

	i. COPYING, People commonly say the GPL restricts 'distribution'
but NOT 'copying'. I think this is wrong. GPL s. 4 provides that 'any
attempt otherwise (as expressly provided under this license) to
void, and will automatically terminate your rights...'. Do I see an
automatic termination based on INTENTION here? It seems that one who
'attempt(s)' to copy WITH INTENT to violate the license will necessary lose
the rights and thereafter infringe copyright by the said act of copying.

	ii. DISTRIBUTING, I see a paradox here. For distribution to
infringe copyright, there must be 'issuing of copies to the public' [s.16
of Copyright Designs and Patents Act 1988, UK] *without* prior permission.
The GPL thus is paradoxical because it gives permission to distribute on
the condition that THE SAME is made a condition. We are thus left with the
question of whether such a condition precedent for distribution can
effectively remove the permission in event of non-compliance, *in time* for
there to be distribution *without* permission (thus amounting to

B. Breach Of A Contractual Obligation

'Violation' of license may also mean contractual breach. Im told UCITA
finds the existence of a contract in every case of software licensing.
'Violation' seems thus more straightforward following UCITA in that it
would be contractual breach.

The UK, there remains the hurdle of (i) 'consideration' and (ii) 'intent to
enter legal relations', before the existence of a contract may be found.
This remains, at least in UK, a largely grey area for the courts to decide
on the facts.


More information about the License-discuss mailing list