[License-discuss] GPL and proprietary WebAPIs
btilly at gmail.com
Sun Dec 25 18:22:12 UTC 2011
The real question is not what the GPLv3 does or does not allow, it is
what copyright does or does not allow. If a work is derived under
copyright law from a GPLed piece of work, then it must be GPLed. If a
work is *not* derived under copyright law from a GPLed piece of work,
the GPL is going to have trouble restricting it. If you write any
other copyright license, you'll run into the same issue.
That said, this is all very vague. There isn't a lot of case law
about what is and isn't a derived work in software. There is a lot of
folklore and opinions. However until an actual judge makes an actual
decision, nobody really knows what will happen. And different
countries have different copyright laws and precedents. What flies in
the USA may not fly in Germany, and vice versa.
Note, I am not a lawyer, and this is not legal advice. I highly
recommend not taking any concrete action that potentially skirts the
GPL until after you engage a real lawyer. Even then I'd recommend
caution. The legal basis of the GPL has held up in several countries,
so the only real question is whether it applies to you.
On Fri, Dec 23, 2011 at 9:20 AM, Clark C. Evans <cce at clarkevans.com> wrote:
> My question is rather straight-forward. Does the GPLv3
> permit the distribution of derived works that require
> an independent and non-free work for its operation .
> I was under the impression that the Corresponding Source
> ("all the source code needed to... run the object code")
> and 5c ("the whole of the work, and all its parts,
> regardless of how they are packaged") would effectively
> prevent this sort of distribution. However, this seems
> not to be the case.
> If so, and this seems to be the consensus I'm hearing,
> then I think the GPLv3 is ineffective; more of a nuance
> than effectively protecting the free commons. Since,
> if I wish to distribute an extension of GPL'd work,
> all I have to do is factor out the critical parts of
> the my work and make them available as an independent
> and proprietary web service.
> On Fri, Dec 23, 2011, at 03:52 AM, Rick Moen wrote:
>> I doubt very much that the recent queries here qualify as
>> that variety of public service.
> You are being unnecessarily argumentative. I'm
> trying to find an appropriate licensing strategy
> for our company, and I'm expressly trying to prevent
> and understand the sort of shims that seem to be
> standard industry practice. If our work can't be
> protected from these "creative circumventions" by
> the GPL, then we probably won't use this license.
> It's my position that if you wish to create a derived
> work that incorporates proprietary functionality,
> you should also provide an equivalent implementation
> under a compatible license. The style of linking
> and the question if the combined work is also derived
> from the proprietary work are largely irrelevant in
> my estimation. Yet, these two considerations keep
> emerging as if they are limitations of the GPL.
> Part of this problem is legal, but the other part
> is what the community accepts as being acceptable
> and that depends upon the public opinion of legal
> and technical professionals on this list.
>  For purposes of this question, you can consider the
> dependency to be declared as part of the derived work,
> but resolved at runtime via sockets or WebAPI; also
> assume that the derived work is *not* a modification
> or transformation of the independent, non-free work.
> License-discuss mailing list
> License-discuss at opensource.org
More information about the License-discuss