[License-discuss] Strong and weak copyleft

Maximilian maximilian at actoflaw.co.uk
Wed Apr 8 13:17:59 UTC 2015


Interesting - I always thought that the distinction between strong and
weak copyleft was in respect of how the code is linked.

Are there any/many examples however of "weak copyleft" given that
definition? I would have thought that "weak copyleft" under that
definition would be largely ineffective, as it would be easy enough to
hide any modifications to third-party copyleft code in other proprietary
source files with function calls.

Therefore, ought the definition of "weak copyleft" extend to include any
code additionally referred to in the third party source file? Also,
ought then the distinction not be on how individual source files are
licenced, but instead the subdivisions (e.g. functions) within the
source files?

This has probably already been considered and dealt with before - I'm
just being verbose and thinking out loud in my lunch break.

-- Maximilian


On 07/04/2015 19:40, Simon Phipps wrote:
> It looks like you may consider LGPL to be a weak copyleft license; my
> apologies if you don't!  But if you do...
>
> I do not believe the LGPL to be a "weak copyleft" license. "Strong
> copyleft" implies that the scope of the required reciprocity is the source
> needed to create the distributed binary, while "weak copyleft" implies that
> scope to be the altered source file alone. The LGPL requirements, like
> those of the GPL, are scoped at the distributed binary, but there is a
> restriction to what constitutes the distributed binary.
>
> Thus I refer to LGPL as "scope-limited strong copyleft" and discourage
> clients from regarding it as weak copyleft. Treating LGPL as weak copyleft
> is a dangerous thing to do as, in the absence of conditions to make the
> limitation of scope apply, LGPL has all the same consequences as GPL.
>
> S.
>
>
> On Tue, Apr 7, 2015 at 7:29 PM, Ben Tilly <btilly at gmail.com> wrote:
>
>> I believe that the legal key is distribution of the licensed code, not
>> linking to it.
>>
>> The LGPL defines a "Combined Work" and has requirements on what is
>> required when you distribute a combined work together.  The intent is
>> clearly that if you distribute the combined work together and DO NOT
>> meet those conditions, then you had no permission to distribute the
>> LGPLed code.  And this has force because while the proprietary half of
>> a combined work is not a derived work, you still need permission to
>> distribute some else's copyrighted code and that permission was
>> contingent on what you did with your application.
>>
>> The GPL defines a covered work to be, "either the unmodified Program
>> or a work based on the Program."  Later in the license a distinction
>> is drawn between that and "mere aggregation".  The intent is that
>> distributing your program + the covered GPLed code it depends on
>> creates a work and you need GPL permission to have distributed the
>> covered GPLed code.  (Whether a judge will agree with this
>> interpretation is another question, but I'm pretty sure that the
>> license drafters intended a judge to understand it this way.)
>>
>> With that said, the LGPL gives a lot of license flexibility for your
>> part of the combined work but says you must allow reverse engineering.
>> Which by default is allowed in many places, but is something that many
>> proprietary licenses take away.  By contrast the GPL offers no real
>> alternative but to license the code you own under the GPL.  Therefore
>> LGPLed code keeps itself copylefted but does not encourage developers
>> to GPL their own code.  While GPLed code pushes people who want to use
>> that code to have to GPL the code that they wrote.
>>
>> On Tue, Apr 7, 2015 at 10:23 AM, Lawrence Rosen <lrosen at rosenlaw.com>
>> wrote:
>>> Patrice-Emmanuel Schmitz referred me to this thought-provoking link:
>>>
>>> https://joinup.ec.europa.eu/community/eupl/news/meaning-%E2%80%9Ccopyleft%E2%80%9D-eupl
>>>
>>> Can anyone here precisely identify the language in the GPL licenses that
>>> makes it "strong" rather than "weak" copyleft? And can anyone here identify anything in copyright law or cases that allow this distinction in the
>>> meaning of "derivative work"?
>>>
>>>
>>>
>>> /Larry
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> License-discuss mailing list
>>> License-discuss at opensource.org
>>> http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
>>>
>> _______________________________________________
>> License-discuss mailing list
>> License-discuss at opensource.org
>> http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss
>>
> _______________________________________________
> License-discuss mailing list
> License-discuss at opensource.org
> http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss




More information about the License-discuss mailing list