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

[Rick Moen]

Since the licence review process is still pending for NCCL 1.0, I'll
add my assessment.


Conflicting aims
----------------

As some other critics have pointed out, author Tim Makarios's two primary
aims of the copyright owner renouncing 'the right to compell' and of all
recipients being passed along copyright terms are incompatible.  Because
the licence states...

   The NCCL applies to a work of creativity (this is called the "Creative
   Work") and comes with any rights that I have in it.

...that therefore includes the right to redistribute and create
derivative works under any terms whatsoever.  The recipient can then
elect to ignore the remainder of what NCCL 1.0 says, possessing all the
rights the copyright holder did.  Because the recipient can exercise
_any rights_, no action can be copyright infringement against the
stakeholder.  Ergo, as Josh Berkus points out, the sole person with
standing has no means to enforce his/her the copyright expectation.
Totally non-coercive or copyleft.  Pick any one.

/me nods at Thorsten Glaser's point that NCCL, even if valid and
enforceable, which it's not, would alienate the permissive-licensing
BSDniks every bit as much as NCCL's punting on access to modified source
code would make it useless to copyleft partisans (as Carlo Piana says).
Which is tangential to OSI review but worth noting in passing.



Defined terms
-------------

As a minor point, paragraph two defines the term 'Derived Work' as
meaning the copyright-law concept of derivative work -- apparently just
so the capitalised phrase 'Derived Work' can be festooned around the
rest of paragraphs two and three.  Why not just say derivative work?

For that matter, the licence uses the terms 'I' and 'you' quite a bit.
The 'you' can likely be safely inferred from context, but 'I' is more
critical, and nothing in this text identifies who's granting the right.
This matter is often taken care of via a template header such as the
Copyright line in the MIT License, http://opensource.org/licenses/MIT --
which is a model of clarity in that regard.  Moreover, come to think of
it, NCCL 1.0 lacks any statement of the year of copyright title on which
the licence grant is based -- a matter I'll elaborate on soon.

Kevin Fleming also alluded to minor problems inherent in the use of 'I', 
to which Tim Makarios made a telling rejoinder:  He says he took it from
SimPL 2.0, and (paraphrasing) figured it was thus blessed because OSI
approved SimPL 2.0.  Tim seems to have copied a large number of other
bits from SimPL, such as the warranty waiver (which I get to, below) and
the pointless definition of 'Derived Work' as a synonym for derivative
work.

Surprise!  SimPL is not a good licence.  Which makes the point that 
mixmastering existing licences is an extremely poor way to write a new
one, and IMO most certainly should not increase the likelihood of OSI
approval.  If I could make one earnest plea to future submitters, it
would be to please cease putting existing licence texts in a blender 
and imagining that to be a good way to make new ones.  It leads to
nothing good.


Copyright notice
----------------

I don't know if Tim was omitting the traditional Copyright line
deliberately or not.  (Quite possibly he merely copied that misfeature,
too, from SimPL 2.0.)  There is a faction of coders (and I don't wish to
be unfair to Tim in lumping him in with them) who appear to so dislike
the global copyright regime that they continually attempt to magick it
away by making no reference to its structures (e.g., they dislike
Copyright lines like the MIT License template's) and flock to whatever
is the ultra-short licence du jour that sounds least lawyerly.  I
sympathise with their dislikes, but the tactic doesn't work in a Berne
Convention world:  Copyright notices, licence texts distributed with
covered works, and very clear, non-morphable licence terms exist for a
reason.

This is why I commend CC0 to people wanting ultra-permissive licences 
(as opposed to copyleft, which Tim wants):  It attempts to do a public
domain dedication, and in the event that is locally impossible, backs
that up with a maximal grant that was well drafted.  But, of course, the
people who want the shortest possible legal instrument (or who are
kidding themselves into thinking of this as something other than a legal
instrument) don't like CC0 as too long, too complex.


Warranty disclaimer
-------------------

NCCL 1.0 has:

  You get NO WARRANTIES. None of any kind;

Certainly clear and bracingly concise (and I've written things like that
in ultra-short licences for non-critical works -- but does it work?
http://www.mslater.com/post/104847-why-are-warranty-disclaimers-in-all
In short, in USA law jurisdictions, the Uniform Commercial Code requires
that a warranty disclaimer be 'conspicuous' and gives the example of
disclaimer sentences being in ALL CAPS as qualifying.

Further, lawyers tend to take the further precaution of explicitly
listing all of the types of express and implied warranties that are
being disclaimed, to ensure that no court could apply.  Thus the hideous
and verbose paragraph at the bottom of
http://opensource.org/licenses/BSD-2-Clause .

Yes, it's hideous.  Yes, it's lawyerly.  Yes, it's verbose.  And all for
a reason.  Even if you think you as a licence drafter don't deserve the
protection of legal best practices, you ought not to thrust that risk
onto downstream re-users and redistributors of covered works.



I respectfully suggest the License Review Chair and/or OSI Board decline
this licence.


(Entirely-unserious query:  If OSI ever appoints a non-American to be
License Review Chair, would he or she be the Licence Review Chair?  ;->  ) 
http://linuxmafia.com/~rick/faq/?page=misc#english

(_Extremely_ unserious observation:  I suppose the correct way to
decline NCCL is would depend on whether it's a feminine, masculine, or
neuter noun in Latin.)

-- 
Cheers,                               "I don't need to test my programs. 
Rick Moen                             I have an error-correcting modem."
rick at linuxmafia.com                                         -- Om I. Baud
McQ! (4x80)                           https://thc.org/root/phun/unmaintain.html