[License-discuss] Copyright on APIs

[Richard Fontana]

On Tue, Jul 2, 2019 at 12:53 PM Pamela Chestek <pamela at chesteklegal.com> wrote:
>
> On 7/2/2019 11:25 AM, VanL wrote:

> Van, I agree with everything you say. But that doesn't answer the same question as "is it open source"? Add to that the interesting possibility that currently-existing licenses will now reach beyond what everyone thought their scope was. Do we expand the meaning of "open source" to match? Or do we just accept it as an unfortunate outcome but not embrace it?

This reminds me a little of the policy issue raised by the largely
forgotten section 8 of GPLv2:

"If the distribution and/or use of the Program is restricted in
certain countries either by patents or by copyrighted interfaces, the
original copyright holder who places the Program under this License
may add an explicit geographical distribution limitation excluding
those countries, so that distribution is permitted only in or among
countries not thus excluded. In such case, this License incorporates
the limitation as if written in the body of this License."

Incidentally, this section embeds the FSF's early policy opposition to
copyrightability of software interfaces as well as software patents.
But leaving that aside, I would argue that this section, if activated
by "the original copyright holder", transforms GPLv2 into a
non-open-source, and indeed non-free-software, license. (Surely if a
copyright holder takes the BSD license and tacks on a geographical
distribution limitation we would say that the resulting amended
license would fail to satisfy the OSD and fail to provide software
freedom; how could we justify treating the GPL differently?) What
makes this largely uninteresting is that hardly any GPL licensor has
made use of this section, perhaps in part because doing so would
obviously clash with basic software freedom norms. (When I looked into
the question around ~2006 I found evidence of only one GPL licensor
who had done so, a developer based in Germany who had apparently made
a deal with a US patent owner to do so.) The non-use of this provision
is why it was not carried over into GPLv3.

So the mere fact that GPLv2 specifically authorizes a software
freedom-inconsistent application of the license -- at least if the
underlying legal system provides some justification for doing so --
does not mean that we have to adjust our understanding of software
freedom, or the OSD, to take account of this possibility.

Richard