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

[Josh Berkus]

On 07/31/2015 05:29 PM, Tim Makarios wrote:
> As far as I can see, the only added risk with the NCCL is that you can't
> sue people who deliberately go around removing references to the
> licence, perhaps in an effort to deceive people about the licence that
> applies to their derivative works.  But how likely is this to happen in
> practice?  Any more likely than that Goldman Sachs would do the same to
> something licensed under an existing free software licence, and not be
> sued for it?

The NCCL is the first proposed license I've ever seen which *encourages*
distributors to distribute the code without the license attached.  While
other licenses exist which are liberal about how the license text is to
be distributed, you are proposing a radical new thing.  And you have
yet, as far as I've seen, to state any real need for it.

Disributing a copy of the license with the software isn't just about
protecting the rights of the original author; it's about protecting the
users, but informing them of their rights under the license.  The GPL
allocates quite a bit of text to this concept, and for good reason.

> The GPL (or any other free software licence) allows people to charge a
> fee in exchange for copies of the software, but people very rarely do,
> because others are always free to distribute the same software free of
> charge.  Likewise, the NCCL allows people to distribute copies of the
> software without the licence attached, but I suspect that people very
> rarely would, because if there's demand for explicitly and accurately
> licensed copies of the software (which there will be among people who
> aren't copyright-naive, if the software is popular), then others are
> always free to supply such copies.

Useful code not infrequently survives the original author, and not
infrequently becomes more popular in a fork or derivative than it did in
the original release.

>> Based on Tim's explanation of the NCCL, I would neither use it for my
>> own code, nor would I be willing to use any code licensed under it.
> 
> Just to be clear, do you mean you wouldn't even use software licensed
> under the NCCL?  Or that you wouldn't contribute code to NCCL-licensed
> projects?  Or just that you wouldn't incorporate NCCL-covered code into
> other projects?

I wouldn't contribute to it, nor use its code in any project I have.
And I would actively reject any contribution which I knew to contain
NCCL-licensed code.  *Assuming I knew*, which is the whole problem with
your license.  For example, let me revise my original scenario:

1. Tim releases widgetX under the NCCL.

2. Mary releases WidgetX.1.  As is her right, she removes all references
to the NCCL from the code.

3. George picks up WidgetX.1 and incorporates it into FrameworkQ. The
rest of FrameworkQ is MIT-licensed.  No license violations have occurred
at this point.

4. Sheila pulls some code from FrameworkQ and sends me a pull request to
incorporate it into my Framework100 project.  The code she pulls happens
to be mostly WidgetX.1.

5. I accept her pull request and release it under the MPL, AGPL, or
other more-restrictive license . At this point, I'm in voliation of the
NCCL, even though I would have no reasonable way to know this.

Again, encouraging derivative creators to strip the license from
distributions is one of the dumber ideas I've heard on this forum, and I
really don't see why we should discuss it any further than we already have.

--Josh Berkus