[hipl-dev] Re: Copyright Attribution in HIPL
- From: Diego Biurrun <diego@xxxxxxxxxx>
- To: hipl-dev@xxxxxxxxxxxxx
- Date: Thu, 09 Jun 2011 20:19:16 +0200
Executive summary: I see no advantage for centralized copyright
management for HIPL. It would just add a lot of red tape for
no gain.
On Mon, May 09, 2011 at 07:14:29PM +0200, Stefan Götz wrote:
>
> >> 1. copyright stays with RWTH and Aalto: For every contribution to
> >> HIPL, it is implicitly assumed that the author is willing to transfer
> >> their copyright (in German 'Verwertungsrechte') to RWTH and Aalto and
> >> to distribute their contributions under the current license of HIPL.
> >
> > You wish... Copyright is not “Verwertungsrechte”. And you cannot implicitly
> > transfer it. – You can assume they distribute their share according to the
> > license used by the project (a new file should include a header with the
> > standard boilerplate), but they still stay with their copyright.
>
> Ok, I was terse on the subject for brevity but some clarification is in order.
>
> Not being a lawyer, just having talked to our university copyright
> management department, the way I understand it, the U.S. legal term
> 'copyright' breaks down into the German legal terms 'Urheberrecht' and
> 'Verwertungsrecht'.
>
> As you pointed out, Urheberrecht / authorship always stays with the
> author and cannot even voluntarily be transferred to a third party.
>
> Copyright and Verwertungsrecht can be transferred to a third party,
> either explicitly by signing an agreement to that effect, or
> implicitly in certain jurisdictions, e.g., when an employee creates a
> works for their employer in Germany.
>
> Correct me if I'm wrong about any of this because it is the basis for
> this discussion.
This is correct AFAIK.
> > Someone here said Libreoffice does not require Copyright transfer, either. –
>
> Ok, this opens the question how we handle the above legal situation.
> Rene and I see two options:
>
> 1) Individual copyright: Include everyone who contribututes to a file
> in the file's copyright statement (I guess that's the LibreOffice
> model).
>
> 2) Central copyright: The copyright stays with a legal entity (such as
> a project foundation or company, e.g., as with OpenOffice, or with the
> RWTH and Aalto universities)
>
> Pro 1): it's the legal default and requires no contract between
> contributors and copyright holder(s) as they implicitly become
> copyright holders themselves. Con 1): a) the set of copyright holders
> diversifies and asserting the rights of copyright holders, such as
> relicensing, becomes much more difficult b) we need to distinguish
> between individual contributors who gain copyright (in terms of
> Verwertungsrecht) on a file and contributions provided under contract
> for RWTH and Aalto in which case the copyright (in terms of
> Verwertungsrecht) stays with RWTH or Aalto. c) We need to define what
> kind of contribution makes someone a copyright holder (Diego claimed
> that 'trivial' contributions do not imply copyright on the whole file,
> so we need to define 'trivial')
>
> Pro 2): it keeps the legal situation simple in the sense that RWTH and
> Aalto stay the sole copyright holders (in the sense of
> Verwertungsrecht) and can decide on relicensing. Con 2): a) it is not
> the legal default and, if challenged, requires HIPL to prove that the
> author deliberately or implicitely transferred their copright to the
> documented copyright holder so b) we need a mechanism for copyright
> transfer.
>
> So yes, our suggestion is to say: if you contribute to HIPL, you
> implicitly transfer your copyright to Aalto and RWTH. If you do not
> agree to this, do not contribute.
>
> Does that seem legally stable enough to at least tell people what our
> policy is instead of having no policy at all?
I fail to see the point.
HIPL is already permissively licensed, so what sort of relicensing do
you wish to embark on in the future? Relicensing that requires the
permission from copyright holders usually just goes in one direction:
restrictive ---> permissive.
All types of uses are already allowed by the permissive license of HIPL,
so no extra control is needed by RWTH or Aalto. Both institutions, as
well as any external entities, can use HIPL for whatever they like.
> > I like the idea, and agree with him in that it is not necessary in with
> > “gift” style licensing allowing everything anyway.
>
> Gift style licenses do not allow 'everything' in the sense that the
> copyright holder and licensor has a number of important rights (such
> as re-licensing) which a licensee does not have.
It is comonly understood that permissive licenses allow relicensing
to more restrictive licenses.
> > If you/others notice a patch sneaking in with a different license (e.g.
> > GPLv3 only), reject it and ask the contributor to add a compatible license
> > as well. (Compatible means here: according to HIPL policy.)
>
> Absolutely.
Yes.
> >> 2. authorship (attribution of the German 'Urheberrecht'): every
> >> contributor is recommended to add themselves to the list of authors in
> >> every file they modify as well as the central list on the HIPL
> >> website. It is a contributor's own responsibility to include this
> >> author information in their patches directly (i.e., please don't come
> >> pleading later on, even though we'll try to be nice about it ;-)
> >
> > “try being nice about it”? Author stays author forever. Even if you’re not
> > nice about it. To not get lost, every author should add his name and email
> > in the file header comment (e.g. in a Copyright statement) in some standard
> > way. A simple script can go over all files and collect all contributors this
> > way. The script defines that standard and avoids the need for an extra
> > manually maintained list.
>
> Absolutely. 'Being nice' was just meant in the following sense: if a
> contributor without commit rights submits a patch and forgets to add
> his name to the list of authors, "we'll be nice about" adding his name
> later through a separate commit. Of course, legally the authorship of
> his contribution cannot be challenged and the HIPL developers have no
> such intention. Practically however, the information about who
> contributed to HIPL might get lost over time, if contributors
> themselves do not add their names to the list of authors. We'll be
> nice about reminding contributors of asserting their authorship
> attribution but we do not take responsibility for meticulously keeping
> track of this information unless contributors do.
We already have Bazaar to track authorship information. I fail to see
what reinventing that particular wheel would gain us.
> >> Does this sound good? It seemed like a practical approach to Rene and
> >> me because it is fairly easy to handle, puts responsibility back on
> >> every author, and keeps the copyright holder question simple.
> >
> > Your policy does not work, because of that “implicit” copyright transfer.
>
> Do you see a possibillity to make copyright transfer work? An implicit
> mechanism would be nicest but if it means that we need every
> contributor to provide some electronically or physically signed
> agreement, that might still work somehow, too, but it's a hassle.
There is no way to transfer copyright implicitly AFAIK.
Diego
Other related posts:
