Hi James,

Le 30/11/2017 à 23:54, James Heald a écrit :
Mathieu,

You don't seem to grasp the essential legal point, though several people in this thread have already tried to tell you.

Copyright protects expression and creative originality.  It does not protect merely a collation of facts.
Well, let's recall https://meta.wikimedia.org/wiki/Wikilegal/Database_Rights#Copyright_protection_in_the_US:
A database is protected by copyright when the selection or arrangement is original and creative.[2] The level of creativity required is low, so it doesn’t have to be very creative — as long as the author had some discretion and made some choices in what to include or how to organize it, the database is likely to be protected.
So, depending on how creative your collation arrangement is, copyright might apply (in the US). In Europe, as you point bellow, sui generis rights might be enforceable.


The CC-SA licence is based on copyright.  Anything that is not protected by copyright is not protected by the CC-SA licence.

To the extent that an article can be reduced to a mere collation of facts, it is not protected by copyright.  What is protected is any originality or creativity in how those facts are organised and presented -- the expression, the sequence of thought, the selections of words, all the authorial choices in the text.
The problem not addressed in this reasoning is that all this "creative choices" can themselves be exposed as factual statements. This could be exposed in extensive development of several concurrent theses regarding the problem of knowledge and creativity from both a gnoseologic and epistemic perspectives. But admittedly, here this would be useless offtopic logorrhoea. So in short, through history people developed, inter alias, theories which states that everything is creative, nothing is creative, only some things are creative.

So the problem here is not that I can't grasp the legal point about the creativity argument, but that I'm not in position of enforcing what is considered creative nor predict whatever some undetermined legal entity might prefer to declare to be creative or not.

For that, you have to get the answer from some legal entity which through their mystic power inaccessible to mere mortal like me will be able to operate the magical performative statement that will seal the destiny of a work into the realm of creativity or relegate it to vulgar combinatorial material for the rest of eternity (in the scope of its jurisdiction, until some other legal decision states otherwise).


*That* is the difference between a copyright-protected Wiki article on the one hand, and a Wikidata collection of facts on the other.


In the European Union collections of facts can be protected by database rights.

That is the path Open Streetmap chose, when they designed the ODbL, to prevent their work being eaten up and assimilated by closed commercial rivals.

It is not the choice Wikidata made.  And it is not the choice any of the Wiki projects made before Wikidata -- CC-SA disclaims database rights.
Actually, as far as I know, CC-by-sa-3.0-undeed states nothing about suis generis rights, and so don't disclaim it but let it applied in all its extensiveness.

And that is the license that cover all other Wikimedia wiki projects (with a dual GFDL 1.3), except Commons where users chose whatever free licenses they want, and Wikidata which permit exclusively CC0.

And a large part of the inquiry on this topic is to determine who decided to use exclusively CC0, through which process and with which goals/perspectives. Some answers stated "long discussions on the topic", but I wasn't given any link so far with something like a vote on the topic, and until something like that is provided, it can't be checked that indeed the community made this decision. So a statement like "the choice Wikidata made" is inconvenient, as what denotation is supposed to be done of "Wikidata" in this context is all but trivial.

Yes, CC0 causes us some difficulties.

It means what we can import from OpenStreetmap is very restricted -- mass import falls foul of OSM's database rights; and also coordinates and boundaries are somewhat susceptible to judgment, so there is probably a copyright element to.

It also makes it difficult to import from official sources (eg the UK Open Government Licence) that use database rights to require attribution -- that is not an obligation we are prepared to pass on to out re-users, which means we generally have to forego such sources.
I think that with the solution already previously proposed to integrate a license attribute, it would be extremely easy for end user to filter items and statements that come with license they don't want to respect, while still enabling other to benefit from their presence in Wikidata.
The important point though is that this boat has sailed.  Wikidata is CC0, and it is not going to change now.
Why not? It was envisioned and explicitly stated from the very beginning of Wikidata that it might switch license at some point. And actually, I'm not even supporting such a move, but simply to also open Wikidata to sources with other free licenses.
Yes, somebody could fork the data from Wikidata into their own ODbL project if they wanted to.   CC0 allows that.  (The reverse direction is what is difficult).  You might have preferred ODBL on viral GPL-style community-building (or community-isolating) grounds.  But that is not going to happen.
I would prefer to avoid fork if possible, this is scattering of resources. And personally I'm not favourable to enforce a single license. ODbL alone would keep entire the legal uncertainty issue of massive import from incompatible licenses such as Wikipedia. However if there was a community driven decision on this topic going in this direction, I would follow it.

Also a fork would not be accessible from other Wikimedia projects, which is as far as I'm concerned the main interest of Wikidata (indeed I have no interest in how it might be used by some random organisation where I don't have any possibility to contribute). Now if the foundation would be interested to run and integrate such a fork, that might begin to become interesting, but otherwise a fork would be useless for the Wikimedia environment.

But there is no restriction, not from copyright law, nor from the CC-BY-SA licence, to stop Wikidata -- or anyone else -- extracting and systematically storing standard uncontroversial facts, so long as nothing of original expression is taken.

Please confirm that you understand this.
I understand that "uncontroversial facts" and "original expression" is too subject to interpretation for enabling any accurate prediction on what will be qualified as covered under it by some undetermined legal entity, all the more as I am not a lawyer. If you have some jurisprudence references that pertain to this topic, surely that would be far more interesting than what I might understand and opinion.

I'm afraid this will not be the reply you would have preferred, but I greet your educational effort toward me and deeply thank you for taking time to write such an extensive and detailed answer.


Cheers,
mathieu