[OSM-legal-talk] "A Creative Commons iCommons license"

[Richard Fairhurst]

80n wrote:
> Indeed it is exactly this case I had in mind, where the license gives 
> the contributor fewer rights.  It creates a class of derivative works, 
> called Produced Works, that are not share alike.

No. This is really, really important.

The concept of a "Produced Work" is not ODbL magically exempting more works
from copyleft. Produced Works is simply ODbL's effort to _define_ something
that exists in all copyleft licences. It isn't a new class of restriction.

* CC-BY-SA uses two terms to describe how work "uses" the original:
Derivative Work and Collective Work.
* ODbL uses three terms: Derivative Database, Produced Work and Collective
Database.

Very roughly (I'm generalising here), in both cases, Derivatives refer to a
situation where the entire result is copyleft, Collectives refer to
something where only part of it is. Produced Works are a subclass of the
latter, not a new class at all. The data component is still copyleft, and a
stronger copyleft than CC-BY-SA gives, but other independently sourced
components may not be.

We have this at the moment. Think of CloudMade's* routing application. Its
raison d'etre is OSM data, and anything derived from that is theoretically
copyleft (not, ahem, that CC-BY-SA requires it to be contributed back). But
no-one is suggesting that CloudMade have to release their code under
CC-BY-SA. The GPL is similar. If you compile a program with GCC, the binary
contains all the optimisations that are probably the most "creative" aspect
of GCC, so in moral terms it could be considered a derivative. But no-one's
suggesting your code therefore has to be GPLed.

ODbL just uses a new term to help firm up the boundaries. The fact we're
still arguing five years on (cf flosm.de) about what's derivative and
collective when CC-BY-SA is applied to data shows how better terminology is
desperately needed.


FWIW, I do think that the ODbL Produced Work provisions _may_ need
rewording. There seems to be a myth around here that a Produced Work can be
public domain. Clearly it can't - not in the traditional sense of PD -
because of 4.7 (the Reverse Engineering provision that dictates that the
data is still copyleft). If there is any restriction on a work, it isn't PD.
This is perhaps not apparent and could do with hardening up.

> The attribution is to the owner of the database, not the author of the
> work.  There is no requirement in ODbL to provide attribution to the 
> authors of the database's content.  Indeed the ODbL asserts that it 
> provides no protection over any of the content, just on the database 
> as a collective whole.  It makes the provision for the database content 
> to be protected by some other mechanism, such as copyright, but we 
> see that the proposed FIL license doesn't provide that protection.

Again, this hinges on whether FIL _is_ what's proposed. I don't believe that
was Jordan's intention. 

The wiki is contradictory: in one place it says
  "Sign up page now states you agree to license your changes under both
CCBYSA and also ODbL."

but in another
  "when you upload your individual contributions, you agree to licence them
under the Factual Info Licence."

You're on the OSMF board, you can tell us. :) For what it's worth, I
distinctly remember Jordan telling me in Reading that he expected individual
users to license their contributions under ODbL; and though in my heart of
hearts I'm a PD person and prefer things like the FIL, I too think that ODbL
is pragmatically what OSM should be adopting here.

> > [...]
> > Database rights only exist for collections.  A single person's
> > contribution may not, on its own, be a database.
> [...]
> Let me clarify.  The database right applies to a collection of facts.  
> An individual contribution may not qualify as a database if it is not 
> a significant collection of facts, not because it is just one person.  
> Most individual contributions will be insufficient *on their own* 
> to constitute a database.

In which case they're uncopyrightable and don't qualify for any protection
under CC-BY-SA either, so it's moot. I don't quite see your point.

>> I believe Jordan's original intent (but he can say this much better than
>> me, and contradict me if necessary) was that users' contributions could
>> individually be licensed under ODbL.
> If that were the case then the FIL license would not be necessary.

Yes, absolutely. Indeed for most people FIL is irrelevant. It's a specific
clarification to cover an issue in Australian law:
  
http://lists.openstreetmap.org/pipermail/legal-talk/2008-February/000607.html

In general terms, FIL covers the individual, atomic, uncopyrightable facts.
ODbL covers your aggregation of these into your contributions to OSM.

> Attribution to individuals is really really important to many
> contributors.
> They give their time and effort, attribution is the *only* reward for
> these
> people.  They want to be able to say "I did that".

Fine. So let's give them attribution. We don't at the moment, by the way. ;)

cheers
Richard

* or anyone else's, for that matter
-- 
View this message in context: http://www.nabble.com/%22A-Creative-Commons-iCommons-license%22-tp22260709p22261842.html
Sent from the OpenStreetMap - Legal Talk mailing list archive at Nabble.com.