On Sat, Nov 28, 2015 at 5:23 AM, Andreas Kolbe <jayen466(a)gmail.com> wrote:
To the extent that Wikidata draws on Wikipedia, its CC0 license would
appear to be a gross violation of Wikipedia's share-alike license
It's essential to also consider whether the factual information derived
from Wikipedia (or any other copyrighted source) is subject to copyright.
For instance, a biography might contain facts like "born in year" and
in place" and "elected to XYZ position". I don't think facts like those
copyrightable in any jurisdiction. Perhaps there are copyrightable elements
from Wikipedia that are brought into Wikidata, but I don't know offhand
what they might be.
The generation of data always has a social context. Knowing where data come
from is a good thing.
Knowing where data comes from is a good thing, yes; but "copyright holder"
and "intellectual source" are not identical concepts. If the purpose is to
preserve the integrity of a line of reasoning, copyright law is probably
not a very good tool for that purpose.
A related question was recently asked on the web site Quora; here's my
answer for why CC0 is generally preferable for data sets. (I may update it
with some of the points brought up here.)