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

[Carlo Piana]

On 06/08/2015 07:09, Tim Makarios wrote:
> On Tue, 2015-08-04 at 10:18 +0200, Carlo wrote:
>> On 04/08/2015 03:05, Tim Makarios wrote:
>>> I'm not the first person to want a copyleft version of something like
>>> CC0 or WTFPL [2], and I probably won't be the last.  I don't claim that
>>> everyone will suddenly want to switch to NCCL, but I do think there will
>>> be some people who will at least sometimes prefer it either for reasons
>>> of principle or merely for the sake of simplicity.
>>>
>> This is a red herring to me.
>>
>> You cannot have the CC0 equivalent of a copyleft license, just as you
>> cannot have the cake and eat it.
>>
>> A CC0 is a total waiver of any possible right in something. It's simple
>> because it's unconditional. A copyleft license is a *conditional*
>> license, and the conditions are strong that a simple liberal license.
>> Since the conditions must be written down as rules in natural language,
>> and natural language is inefficient in writing rules -- yet it's the
>> best option so far -- you need long and complicated text to make the
>> rules and exceptions.
>>
>> It is utterly naive thinking that some of the best legal minds in the
>> world have come to the GPL v.3 just because they could not write
>> something simpler than that, so somebody without legal legal training
>> can do better. I'm sorry, but I keeping hearing this nonsense.
>>
>> Writing conditions is difficult, writing them so that they are not
>> easily exploited is even more difficult and have them working in the
>> many different conceivable scenarios, present and future ones, is nearly
>> impossible.
>>
>> So you can dislike copyleft, and many do, and many advocate against it,
>> because they prefer liberty over control, so they accept the odds that
>> code will be misused, or simply don't care, for the benefit of having
>> simple and straightforward licenses and larger commons because of lack
>> of licensing incompatibilities. Others think different. But it's always
>> a trade off. You can't have both copyleft *and* simplicity, you can only
>> have a better noise/signal ratio.
>>
>> Sorry for the rant, but I think we should set the record straight here.
> Are you talking about the simplicity of the language of the licence?
> The NCCL inherits much of that from the SimPL, which, at around 300
> words, was, in fact, written by a professor of law at the University of
> Washington, is supposed to be equivalent to the GPL v2 [1], and has been
> approved by the OSI.  The text of CC0 is significantly longer, reaching
> nearly 1000 words [2].

Both.

In fact, my point is totally that complication of language derives from
complications in underlying rules, plus the necessary inefficiency of
natural language to lay down legal rules, plus unnecessary inefficiency
that I label "signal/noise ratio". My point is that you can only improve
this latter factor, or trade off simplicity with conditions.

The fact that SimPL has been approved says nothing about it being a
workable copyleft license by any stretch of imagination. Such as the
fact it has been written by a law professor, which I admit I have never
heard of (this is hardly the point, though).

>
> But actually, when I wrote the paragraph you quoted above, I didn't have
> the simplicity of language in mind; what I had in mind was the
> simplicity of the requirements that the NCCL imposes on the licensee.
> Or rather, "requirement", singular, since my intention is that the
> *only* requirement is the copyleft condition.

I appreciate the difference, and neither is achievable, or achieved.

The only requirement your license impose is that the derivative is under
the same license as the original. Apart from other deficiencies that
have been otherwise discussed, it does not even create the basic
conditions for such requirement to be effective in real life.

>
> Yes, by this standard, it isn't as simple as CC0, but it is simpler than
> any other copyleft licence I'm aware of, or can imagine.

Point is, it's not a workable copyleft license, because of the
deficiencies I spotted.

I can't really say it's OSD-compliant as such, OSD certainly does not
require copyleft (maybe copyleft can make being OSD-compliant more
difficult). But I refute the underlying concepts of your proposition and
I would be pleased if a rejection decision is achieved. Others have
added more reasons, no point in repeating them.

Sorry to be blunt, we are not taliking about the persons, we are talking
about legal positions. But your case is one of the most underwhelming I
have ever read on this list.

>
>> PS: I have expressed an opinion in the past that any copyleft license
>> which do not ensure access to modified source code is almost totally
>> moot, at best it's a shareware license. The same applies here.
> The GPL doesn't require that software it covers is distributed free of
> charge.  But we see that in practice, its effect *is* almost always to
> ensure costless access to such software.  Why?  Because the GPL creates
> a free market in the distribution of the software, and in a perfectly
> competitive market, the price tends towards the marginal cost, which,
> thanks to the internet, is effectively zero.  I believe that this
> economic principle is widely accepted by micro-economists belonging to
> various political traditions; even the author of The dotCommunist
> Manifesto has been known to refer to it [3].

How is this any relevant?

Incidentally, the GPL does not require to be distributed free of charge,
it does not even require software be distributed at all (perhaps such a
license would not be OSD compliant) . It prohibits to all who are
distributing GPL software to require monetary compensation for receiving
the complete corresponding source code, because that right is a
condition for the copyright grant.
 
In my humble experience of advising clients on GPL compliance, this is
something that many would avoid if it was not required by the license.
By avoiding this simple requirement, you are de facto removing any
meaningful copyleft from the license, what remains is confusion and
wishful thinking.

This is what I was speaking about.

>
> Likewise, I believe that despite not requiring that the source code of
> derivative works is published, the NCCL will in practice ensure that
> source code of published derivative works is almost always published,
> simply by ensuring a free market in the source code.  Sure, it's
> *possible* that sometimes the author of a derivative work might withhold
> the source code, forcing others to reverse engineer it before the source
> code is freely distributed; but it's also *possible* that under the GPL,
> authors of derivative works might sometimes insist on payment in
> exchange for copies of their derivative works.  But for those authors,
> such a strategy is roughly equivalent to a reverse bounty strategy [4],
> because as soon as several people have copies of the software, they'll
> bid each other's distribution price down to the marginal cost, which is
> probably zero.

This is the same reasoning of all non-copyleft licenses. Which
consequently have no pretence of being copyleft, as they see copyleft as
an unnecessary encumbrance. You are making the case for rejecting your
submission.

A strategy is different from a legal requirement, is different from
establishing a *condition* to the copyright grant.

You keep confusing the legal requirement of exchanging the complete
corresponding source code with asking to be paid for distributing
software at all. Honestly, it's difficult to have a conversation on
these bases.

>
> In a reverse bounty strategy, the bounty that can be obtained is bounded
> by the value of the software to be provided, and also by the cost of
> having someone else implement equivalent software.  But in a situation
> where the binary is already available under a copyleft licence, the
> bounty that can be obtained in exchange for the source code is bounded
> not only by the value of the source code and by the cost of having
> someone else write equivalent source code, but also by the cost of
> reverse engineering the binary.
>
> If reverse bounties are relatively rare for open source software (which
> I believe they are, compared to how much open source software is simply
> released free of charge), this analysis suggests that they will be even
> rarer for the source code of derivative works of open source software
> whose copyleft licences don't require the source code to be made
> available.
>
> What does the developer of a derivative work have to gain by withholding
> the source code?  Perhaps a one-off bounty (if they're lucky, and people
> don't simply exercise their right to reverse engineer the binary), after
> which the source code will be widely available free of charge anyway.
> What do they have to gain by releasing the source code?  More time with
> more people having the opportunity to find and fix bugs, and maybe even
> contribute new features.

Go and ask GPL-violations, or FSF. People KEEP distributing GPL software
and refusing to comply with the legal requirements of the GPL. Oddly
enough, the world is full of incentives to follow that strategy, and
sometimes you need to go to court, maybe simply to set an example. Very
good and clever legal minds are at work to circumvent the GPL, I don't
even have to cite examples, they are so widely known.

I agree that fear of retribution is not the *only* motive for people to
comply with a license, I concede this is not even the most important.
But we are talking *licenses*, as legal and *enforceable instruments*.
If they fail with enforcing the basic requirements for the software to
remain in the commons, all that remains is hoping for the best. Which is
not copyleft.

>
> But maybe my analysis is wrong.  Do you see a flaw in it?  Or have you
> seen software released under copyleft licences that don't require the
> source code of derivative works to be published?  Maybe someone has
> tried releasing software under CC-BY-SA?  If so, how did it turn out?

Surely people have attempted, but the absence of significant examples
would rather confirm, not debunk, my point.


Sincerely

Carlo