[License-review] Request for approval of the Non-Coercive Copyleft Licence (NCCL) 1.0

Tue Aug 11 05:25:12 UTC 2015

On Tue, 2015-08-04 at 10:08 -0700, Josh Berkus wrote:
> *shrug*.  You're still missing the point entirely.  Explicitly allowing
> the code to be redistributed without a copy of the license is setting up
> a licensing entrapment, and there's no two ways about it.

Entrapment of who?  Of people who think they can, without permission,
apply a licence of their own choice to any code they find on the
internet without a licence notice at the top?

> Believe it or not, OSI does not have the mission of making sure that
> every single combinatorial leaf node of license possibilities is filled
> in.  Quite the opposite: we have the mission of *restricting* the number
> of new licenses we approve (and even deprecating existing licenses)
> because, all things being equal, having fewer licenses available is a
> benefit to the public.

The stated purpose of the Licence Review Process specifies vanity and
duplicative licences as the ones the process is meant to discourage [1].
Much of the criticism of the NCCL so far has not been about how
*similar* it is to existing licences, but about how *different* it is:
it doesn't prohibit the removal of licence notices, and it's a copyleft
licence that doesn't explicitly require the publication of the source
code of derivative works.

Indeed, one of the critics of the NCCL even said "you are proposing a
radical new thing" [2] ;-)

I stated in the very first email requesting approval of the NCCL [2a]
that "I would be willing ... to withdraw this request if someone can
point out an existing licence that achieves these goals".  This offer
stands.

> You seem to be under the mistaken impression that if you can game the
> OSD then we're somehow obligated to approve your proposal.  We are not.

I'm not trying to play games here.  I actually want to use a licence
like the NCCL.  I submitted it in good faith, believing that it
satisfies both the letter and the spirit of the OSD, and I'm trying in
good faith to engage with the suggestions and respond to the criticism
received on this list.

The reason I haven't already started using the licence is that the OSI
recommends submitting new licences before they're finalized, "to allow
for any necessary changes" [1].

>  *You* are obligated to prove that your license is open source,
> necessary, and well-written.
> 
> Therefore, you need to provide:
> 
> 1. What real need (preferably with project names) is this license
> designed to serve?

Well, I would use it for rootstock [3] and rttp-proofs [4], but the
licence wasn't designed for a particular kind of project; it was
designed for a particular kind of person:  the kind of person who
distinguishes between what people *ought* to do, and what they ought to
be *forced* to do; the kind of person who says yes, it *is* wrong to
unreasonably withhold accurate attribution; the kind of person who says
yes, ensuring that interested users are informed of the licence applying
to the software they're using is the right thing to do; the kind of
person who might even say yes, a married person should be faithful to
their husband or wife; but the kind of person who says no, none of that
makes it right to threaten coercive application of the law against
violators of these principles.  After all, "the law never made anything
perfect" [5].

That's why I want a licence whose *only* requirement is designed to
prevent (to the greatest extent possible) creators of derivative works
from threatening coercive application of the law against users of those
derivative works.

When a licence says, for example "Redistribution and use in source and
binary forms, with or without modification, are permitted provided that
the following conditions are met" [6], there's an implicit threat to sue
anyone who redistributes the software without meeting those conditions;
and when those conditions are about things far less important than
faithfulness to the one person who should be able to trust you more than
anyone else in the world, the kind of person I described above will not
be happy making such threats.

I won't refuse to contribute to projects under other open source
licences, but when the choice of licence is mine, I'd rather not make
unnecessary threats.

> 2. Why are none of the existing mainstream licenses adequate?  Please
> address the BSD, MPL, LGPL, and GPLv2 licenses.

The BSD implicitly threatens coercion against redistributors who fail to
retain or reproduce the copyright notice and the licence; also, it does
nothing to prevent creators of derivative works from coercing users of
those derivative works.

The MPL implicitly threatens coercion against (for example) distributors
of executables who fail to inform the recipients of the executables
about how to obtain the source code.

The LGPL implicitly threatens coercion against people who copy,
distribute, or modify the program but fail to meet a somewhat convoluted
set of conditions.

The GPLv2 implicitly threatens coercion against (for example) people who
modify the program but fail to add prominent notices stating the date
the modifications were made.

> 3. What legal support do you have to indicate that the innovations you
> are making (such as the ability to remove the license but still enforce
> original creator rights) have any effect in law?  Please cite
> references, or provide the opinion of a qualified attorney.

I'M NOT A LAWYER; WHAT FOLLOWS IS NOT LEGAL ADVICE; IT DOESN'T CREATE AN
ATTORNEY–CLIENT RELATIONSHIP.  I PROVIDE NO WARRANTIES REGARDING THE
INFORMATION BELOW, AND I ACCEPT NO LIABILITY FOR ANY DAMAGE THAT RESULTS
FROM THE USE OF IT.  AND I'M TYPING THIS WITH CAPS-LOCK ON SO THAT
NO-ONE CAN SUE ME FOR NOT BEING EMPHATIC ENOUGH.

The Berne Convention for the Protection of Literary and Artistic Works
[7] requires member countries to recognize copyright in a large variety
of works, regardless of formalities such as whether copyright notices
have been applied to the works.  The convention does, in Article 2(2),
allow member countries to restrict copyright to works that have been
fixed in some material form [8], but this is hardly relevant to the
question of whether the removal of licence and copyright notices
inhibits the ability of the copyright owner to enforce their monopoly.

The Berne Convention currently has 168 contracting parties [9],
representing jurisdictions covering the vast majority of the world's
population.  Member countries implement the convention in their own
legislation.  For example, New Zealand implements it in its Copyright
Act 1994 [10].

Section 18 of that act provides that a work qualifies for copyright if
its author (or any of its authors) is a citizen of, is a resident in, or
is a body incorporated under the laws of New Zealand or a prescribed
foreign country [11].  Sections 19 [12] and 20 [13] further broaden
those works that qualify to those that are first published in or
broadcast from New Zealand or a prescribed foreign country.  In
accordance with the Berne Convention, there is no requirement that the
work have any copyright notice or licence attached.

However, section 121 does provide that in cases of copyright
infringement, the entitlement to damages is affected if "the defendant
did not know, and had no reason to believe, that copyright existed in
the work" [14].  There are two possible reasons for the belief that
copyright does not exist in a work: one is the belief that the work did
not qualify for copyright; the other is the belief that the copyright
has expired.  (Another possibility may be a belief that the copyright
has been renounced by its owner, but this is not implied by the mere
absence of a copyright notice.)

If no copyright notice, licence, or claim of authorship is attached to a
computer program, the belief that the work did not qualify for copyright
would be very hard to justify, given the very broad reasons for
qualification for copyright in sections 18 to 20.  To understand how
broad those reasons are, it's necessary to see how the legislation
defines the phrase "prescribed foreign country".

Section 2 defines that phrase to include (via section 230 [15]) any
"convention country" (other than New Zealand), which is, in turn defined
as "an entity that is a party to an international agreement or
arrangement relating to copyright" [16].  Given the broad membership of
the Berne Convention alone, this makes it very likely indeed that a
computer program without a licence or copyright notice qualifies for
copyright unless there is specific evidence that it does not.

The remaining question — of whether the copyright has expired — is also
unlikely to be able to be used to justify a reasonable belief that
copyright does not exist in a computer program.  Section 22 allows that
the duration of copyright in computer programs is generally at least 50
years even if none of the authors are known; if any of the authors are
known, it can last for 50 years after the death of the last of them
[17].  Even if the software looks more than 50 years old, and is still
relevant, the authors still had approximately half a century in which to
make themselves known, if they wanted to extend their copyright until 50
years after their deaths.

So, not only does the absence of a licence or copyright notice not
affect the validity of the copyright; it is very unlikely even to affect
the ability of the copyright owner to obtain damages if an infringement
occurs.

If there are still lingering doubts about the enforceability of
copyright on works without licence or copyright notices, it is
worthwhile considering section 128, which states that if a program
issued to the public in electronic form carries a notice asserting that
a particular person owned the copyright when the copies were issued,
then that assertion is presumed to be accurate unless it is proved false
[18].

There is no requirement that *every* such issuance to the public carries
such a notice; still less is there a requirement that every licensee
authorized to distribute the program must be required to preserve such
notices.  In fact, subsection (4) explicitly states that the issuance to
the public of copies carrying such a notice need not even occur before
any alleged infringement; the presumption is still valid in proceedings
related to such an alleged infringement if the issuance to the public of
copies bearing the notice occurred only after the alleged infringement.

Perhaps you wonder whether the NCCL can bind people who receive
NCCL-licensed works that don't carry a copy of the licence, or any
licence or copyright notices.  Actually, the NCCL doesn't purport to
bind such people, if they haven't relied on it for their right to copy,
distribute, or make a derivative work of the Creative Work (see the
second sentence of the licence text [2a]).  What *does* bind them is
ordinary copyright law, for the reasons detailed above.

If they aren't aware that the NCCL gives them rights far beyond the
ordinary exceptions to copyright law, then that's unfortunate for them,
but their ignorance doesn't harm the copyright owner in any way.

One paragraph of the licence says the following [2a]:

> If you rely on this licence for your right to distribute a Derived Work,
> you must give back to the community by agreeing that the Derived Work is
> governed by this version of the NCCL or substantially similar terms,
> without adding further restrictions to the rights provided.  If you do
> not conspicuously announce which such licence governs the Derived Work,
> then you agree that this version of the NCCL will govern it, and that I
> and other people are allowed to rely on that fact.

This is intended to *protect* the recipient (say, Alice) of
NCCL-licensed software that doesn't come with a copy of the licence or
with any licence notices; it isn't intended to *bind* Alice.  It
protects her because if the person (say, Bob) who supplied her the
software (a derivative of Charlie's NCCL-licensed work) tries to sue her
for her use of the software, then she can argue in court that Bob has,
in fact, licensed his derivative work under the NCCL, even though he
hasn't explicitly said so; if this argument succeeds, then Charlie never
needs to sue Bob for copyright infringement.

"But", you might object, "how can it protect Alice if she doesn't know
about the licence?".  Well, it can't, of course.  But how could a clause
requiring Bob to include licence notices with his derivative work
protect Alice?  If he's willing to breach the NCCL by trying to sue
Alice for her use of his derivative work, why wouldn't he be willing to
breach a modified NCCL's licence-notice-preservation clause?

Tim
<><

[1] http://opensource.org/approval
[2] http://tinyurl.com/qdmxped
[2a] http://tinyurl.com/ov5zd6g
[3] http://repo.or.cz/rootstock.git
[4] http://repo.or.cz/rttp-proofs.git
[5] http://tinyurl.com/ozh3zq7
[6] http://opensource.org/licenses/BSD-2-Clause
[7] http://www.wipo.int/treaties/en/text.jsp?file_id=283698
[8] http://www.wipo.int/treaties/en/text.jsp?file_id=283698#P85_10661
[9] http://www.wipo.int/treaties/en/ShowResults.jsp?lang=en&treaty_id=15
[10] http://tinyurl.com/p8eoxnj
[11] http://tinyurl.com/p59k9kp
[12] http://tinyurl.com/q4wp8a8
[13] http://tinyurl.com/p7dfc45
[14] http://tinyurl.com/nojh3ed
[15] http://tinyurl.com/pltofr4
[16] http://tinyurl.com/pqcd7o6
[17] http://tinyurl.com/pp2xmep
[18] http://tinyurl.com/qh6occg