[Gta04-owner] Updates for gta04 DTS file.

[xdrudis at tinet.cat]

On Tue, Jan 06, 2015 at 09:38:19PM +0100, Dr. H. Nikolaus Schaller wrote:
> > 
> > "Signed-off-by" isn't about authorship.  The "Author" tag or "From:" line
> > gives authorship.
> > “Signed-off-by" is a statement about copyright.  
> 
> Ok, this might be a wording issue introduced with slightly different copyright laws.
> In Germany the author automatically owns the copyright and can’t transfer it. He can
> only licence something (which is the same as transferring in practise). This background
> has an influcence why my view is that author and copyright/signed off is the identical
> and can be exchanged. But there are other legislations where it can and must be separated.
>

Check with a lawyer but it is my understanding that Neil is right.
Signed off is not about who is the author, it is about who takes
responsability that the incorporation of a contribution is legal (at
least according to copyright/author rights law). The contribution may
be completely original by the "signer off" or may be taken from some
other author/s who has/have given permission to incorporate it in the work
the "signer off" is proposing to incorporate the contribution in. 
Or the contribution may be a contribution of material from the "signer off" 
and material from others, as long as all authors have given permission. 

There are many differences between licensing something and trasferring
copyright. Mostly about sublicensing.

When you transfer copyright you keep only moral rights, and the only
people who can grant further licenses (for explotation rights) are the
people you transferred copyright to.  You can transfer in exchange of
money, and/or rights, for example you can negotiate to get a right to keep
licensing, etc. but that would be in addition to copyright transfer.
That's why there are fiduciary copyright assignment contracts that 
set up (by contract) the rights of the transferring party and receiving
party as coditions attached to the copyright transfer. 

When you license you keep copyright ownership and can use it to give
different licenses in the future. If you take a work of someone else
and they didn't transfer the coyright to you you need them to license
this use of their work. If you then send that work to be integrated
in a larger derived work you need to check that the license you got
from all authors grants the right to integrate it in the larger work
and distribute it under the conditions the larger work will be distributed.
That's why copyleft licenses exist, to avoid negotiating copyright
at each collaboration because it is already licensed beforehand. 

The "signed off" means that the sender of the contribution has checked
that he has right to incorporate the contribution in the larger work
and distribute it and sublicense the larger work under the conditions
the larger work was previously distributed, so that the maintainers
don't need to check they have the copryrights needed if they trust the
"signer off". I haven't seen explicitly written but I assume it is
implicit that the "signer off" does not need to audit the legal status
of the larger work before the contribution, just that if the larger
work was legal to distribute as it was being distributed before the
contribution, the "signer off" states that it will still be after
incorporating the contribution.

Some projects require copyright assignment. Sometimes Ubuntu, or Qt
do, if I remember right, because they want to be able to license the
works under propietary licenses. Often the GNU project or others
require copyright assignment too (giving back some rights to the
author and some more to the public), because they want to make sure
they can relicense under newer free licenses or sue someone who breaks
the licenses. Only those in posession of the copyright can sue for
copyright infringement (at least so it used to be).
  
Some projects work in a decentralized copyright fashion (linux,
coreboot, etc.), every contributor keeps its copyright, every
contributor trusts the others have checked they have the right to
contribute when they "sign off", and no single entity keeps the
copyright for all of the work. This make it harder to relicense the
work, but makes it easier to receive contributions.

In continental Europe, and I believe this includes Germany because I
thought it was a little the German model other legislations copied,
the author has all copyright rights inmediately by the mere creation
of the work (but software created by an employee is copyrighted by the
employer unless otherwise agreed). Some of these rights, the moral
rights, are inalienable and the author can't sell or transfer in any
way. Some other, the explotation rights, the more economic rights such
as making the work available, copying it, modifying, sublicensing,
etc. can be transfered. In Germany a copyright assignment does not
transfer 100% of the rights and in UK or USA maybe it does, but in any
way the only difference is moral rights that are almost never used and
licenses try to work around them (moral rights include things as the
author at any moment can prevent use of the work if it is done
endorsing ideas the author does not share, but the author has to pay
the expenses/losses of stopping the publication, the author can decide
whether the work must be attributed to the author or not, can complain
of alterations of the work that may damage the authors image, etc.).

"signed off by me" does not mean "I wrote this and you can use it" it
means "I know this can be incorporated in your project, because I or
whoever it takes are giving you permission". For example if I work for
a company and don't have any special agreement, the signed off of a
contribution that includes my paid work in that company should be by
whoever my employer designates as empowered to grant copyright (might
be myself or not).  There could be a lawyer in the company checking
where I got the code from, if its license allows our use and whether
the company wants to allow others the use of whatever I wrote new, and
then either allow me to send it signed by me (more usual) or signed by
the company lawyer. Or even more usual, company lawyers tell employees
what they can do and resolve doubts and the company does not have a
lawyer whatching over the shoulder of each employee.
 
> > 
> > If both you and I wrote identical patches, then there in clearly no
> > significant creative input and nothing that can be copyrighted.
> 
> That is of course also right and makes the discussion quite irrelevant.
> 

It may well be so, yes, but I've never known what the threshold for
creativeness is. The Oracle/Google battle for Android seems to imply
what everybody assumed can be at odds with judges sometimes. And it
may depend on jurisdiction, or even on particular judges ?

But yes, my intuition is like yours. If two people independently write
the exact same code it possibly can't be copyrighted. 

> > In fact, we didn't write identical patches at all, though some of the code is
> > the same.  We split the code up into patches quite differently in most cases.
> 

Well, in case one could prove that a) the repeated code is
copyrightable and b) someone wrote it first and put it somewhere the
other could have seen it, then I don't think it would matter how you
split it and mix it with original code, in case a part of a patch was
copied (or could not be proved otherwise) then authorship should
acknowledge it (as Neil proposes later just in case, to avoid any
issues). But any patch may be attributed to multiple authors if they
agree to be coauthors.

> Ah, ok. I think I have learned something. Until now I always thought that software
> is copyrighted. I.e. the resulting file. But apparently, the focus is on copyrighting
> or licensing of patches. Which raises an issue like the one we discuss if two
> developers independently come up with different patches that result in the same
> code.
> 

Not sure whether the focus is in copyright the whole source or
patches, but of course the copyright of the whole combined work is not
controlled by a single contributor (in linux, anyway, where there are
no copyright assignments), so you can only look at what license you
grant patch by patch. A patch might have multiple authors, though,
just like the patched file. Since everyone grants the same copyleft
license, it is just a problem of authorship attribution. 

I think what one usually does is dodge the question of creativeness by
making sure that if the contributor holds some copyright it is
licensed under the project license. If the contributor does not grant
some license that is needed it is a problem for the project and all
users. If a contributor grants a license that is not needed (or is not
possible because there is no right to license) it isn't a problem for
anyone, it's just void.  So when licensing you assume you hold more
copyright than you possibly do (like in "if this belongs to anyone at
all it surely belongs to me, so I'll license it"), and before sueing
you try to make sure you really can prove you have that copyright (or
if you are like Oracle you risk it to push jurisprudence).

> > 
> > Still, if you think you wrote it first and want your code to go upstream with
> > your authorship and your signed-off-by, please do that.  I'd be very happy to
> > see the code upstream no matter how it got there.
> 
> Ok. I understand.
> 
> Then we should do it this way.
> 

Yes, thank you both for your efforts and trying to get things
upstream.  I don't have enough experience to tell who is right about
development models (and I'm user not developer here, so it'd possibly
be irrelevant if I could tell). I understand what Niklaus tells about
the closer community being people who can test things (because they
have GTA04s) and he trying all these to work together by publishing
work in progress as soon as possible, but if that's too much work for
too little help for Neil, someone who is just giving away his work to
us as well, then I don't see how to get coordination work done without
detracting from development work.

But I'm learning from your dialogue. So thank you both for code and 
for discussion.

Remember IANAL, get legal counsel if you need it. Don't trust me as
legal advice.