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

Josh Berkus josh at postgresql.org
Fri Jul 31 17:09:56 UTC 2015


On 07/30/2015 09:59 PM, Tim Makarios wrote:
> You might prefer NCCL if you don't want to worry about checking whether
> the source of derivative works carries "prominent notices stating that
> [the creators of the derivative works] modified it, and giving a
> relevant date" and that when they "Convey individual copies of the
> object code", they include "a copy of the written offer to provide the
> Corresponding Source", and satisfy all the other requirements of GPL v3
> (for example).  You might prefer NCCL if you don't want to be making an
> implicit threat to sue creators of derivative works if they fail to
> satisfy these requirements.
> 
> Or, relative to CC0, you might prefer NCCL if you want to be sure that
> creators of derivative works won't sue you for taking the same liberties
> with their derivative works as they took with your original work.
> 
> In a sense, NCCL is aiming to be right at one edge of the space of
> possible licences --- the permissive-most edge of copyleft licences.
> You could imagine that minor requirements could be added to or removed
> from GPL v3, for example, without fundamentally altering its nature as a
> copyleft free software licence.  But if NCCL does what I want it to do,
> then you can't really add or remove any restrictions without making it
> easier to create a situation in which people can be sued for their use
> of derivative works.

Speaking as a non-attorney, this sounds like a bunch of balderdash to
me; I certainly wouldn't buy into it as a developer.  I'd need Larry and
Eben both to endorse this as a legal concept before I would believe in it.

Further, I really don't think OSI should be encouraging licenses which
depend on the reciplient's knowledge of the lineage of the software
they're receiving in order to determine lincensing, which is what the
NCCL effectively does.  The NCCL, if used as the submitter intends, is
basically a landmine, because it claims to still have authority over the
code even if not included with the code.  Far from removing the ability
to sue, the NCCL encourages surprise lawsuits.

Just imagine this scenario:

* Tim releases WidgetX under the NCCL
* Mary releases WidgetX.1 under the NCCL, but does not include a copy of
the license, as permitted by the NCCL.
* George releases WidgetY, which is really a derivative of WidgetX.1,
under the MIT license, out of a mistaken impression that Mary's code is
public domain.
* I merge WidgetY with Mozilla code to release WidgetZZZ, under MPL,
with portions under MIT.
* Tim sues me for violating the NCCL license on the WidgetX code.

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.

So, -1 from me.

--Josh Berkus







More information about the License-review mailing list