Re: [gentoo-project] Re: [gentoo-dev-announce] Poll: Would you sign a Contributer License Agreement?

[Rich Freeman]

On Thu, May 31, 2018 at 9:55 PM R0b0t1 <r030t1@×××××.com> wrote:
>
> On Thu, May 31, 2018 at 6:52 PM, Raymond Jennings <shentino@×××××.com> wrote:
> > I refuse to sign anything.
> >
> > If the GPL works as intended, then anyone using my work would
> > themselves be making a derived work of their own for which they
> > themselves could enforce the GPL.
> >

They could pursue the violation for just the copyright on the changes
they made, not to the work they derived it from.

For example, if somebody reproduced this email illegally (let's assume
it was distributed under a non-free license), I could sue them for
copying these two paragraphs, or for the email as a whole, but not if
they removed my additions and only sent the quoted text above.
Likewise you could sue them for copying the quoted text above, or for
the email as a whole, but not if they removed the quoted text and only
copied my additions.  This email as a whole is a derived work, but if
I didn't quote you it probably wouldn't be.

>
> It is for this reason I never understood the point of contributor
> agreements for open source projects. Even the FSF's justification, so
> that they may pursue GPL violations without requiring you present,
> seems to fall apart as various Linux kernel contributors have gone
> after companies on their own for profit without the consent of Mr.
> Torvalds, the other contributors, or the Linux Foundation.
>

You're comparing apples and oranges here.  The Linux kernel doesn't
require people to sign CLAs.  That means that the Linux Foundation
generally CAN'T pursue copyright violators, but the individual
contributors CAN.  If the contributors signed traditional CLAs, then
the Linux Foundation COULD pursue copyright violators, but the
individual contributors COULD NOT.  A traditional CLA transfers the
rights of the creator to the assignee, at least under US law.  I'm
speaking generally of course since what any particular CLA does is
governed by the wording of that particular CLA, and of course the
governing law.

The FSF wants people to sign CLAs so that THEY can pursue copyright
violators.  The contributor already has this right, probably.

There are a few other nuances (again, talking about traditional CLAs
that assign copyright):

*  Some question whether somebody holding copyright over only a small
part of software could on their own pursue a violator, or how
effective this would be.  These people argue that a CLA consolidates
the copyrights so that instead of a bazillion people owning copyrights
to 3 lines each, you end up with 1 entity owning copyright to the
whole thing, which eliminates this issue.

*  A CLA also can allow for relicensing beyond the limits of the
standard "or a later version" language (for projects that even use
this language).  If a CLA is signed then an organization could choose
to switch from GPL to CDDL, or from BSD to Apache.  Of course, many
traditional CLAs would also let them switch from GPL to "all rights
reserved," with the caveat that whatever was prevsiously released
under the GPL could still be redistributed and modified by recipients
under the GPL.

Note that the above pertains mostly to traditional CLAs.  For the FSFe
FLA approach the main benefits are:

*  Gentoo could re-license under a different free license, within the
limits of the agreement, assuming we owned enough of the code or
otherwise dealt with that issue.  This could be useful if some new
license comes out later not related to the GPL/etc.  If you care about
such things you'll want to read it for yourself, but the agreement is
structured to prevent shenanigans like proprietary re-licensing.

*  One of the goals originally in the FLA was to make it exclusive in
a way that would ensure that Gentoo would know that you hadn't given
somebody else permission to use your contribution under a non-free
license (which means we know anybody doing this is in violation).
However, looking at the wording of section 2.3 of the new version I'm
not sure that this feature even applies anymore, as there don't seem
to be any copyleft restrictions on the grant back and its ability to
relicense.  That seems like a bug the FSFe might want to look into.
In any case, without an FLA with the correct wording, there is no way
to know if somebody is in violation without contacting the original
contributor, because the original contributor could have given them
permission to use it under a non-free license.  This makes detecting
violations more difficult.

(I realize the above might be confusing, and can elaborate further if desired.)

-- 
Rich