On Tue, Feb 6, 2018 at 9:36 PM, Samuel Klein
<meta.sj(a)gmail.com> wrote:
> This has been missing for some time -- there is no way for researchers to
> use commons to share and archive freely-licensed data files. It came up
> again at the dev summit as a persistent problem (despite the existence of
> the Data namespace); has there been any recent talk about enabling this?
>
> SJ
In theory this would be jolly nice. However as the copyright issues in
the data namespace have been difficult, we know that introducing these
formats will require further guidelines and policy changes to cover
csv and other types of file.[1] Let's not underestimate how much time
this takes, and how few free volunteer days there are available to do
this in an adequate way, support by a consensus for the changes.
As an example, though individual items in a data set may be thought
uncopyrightable, there are many examples where the assembled data is
considered creative work, and so can be a new intellectual property.
This is clearly true with subjective data, such as the "top films of
the 20th century" which I can see in my Kodi app, which is probably
not "freely-licensed", it may not even be legal for me to download
it...
The red flag in SJ's email is understanding "free-licensed data
files". For me, this must mean being able to:
- validate the source
- apart from exceptional cases there must be an identified rights
holder to release the data
- the data should have "educational value" and so must have test-able
reliable content
- volunteers must have easy tools to validate content, in the same
way as we can use Google Images to search for copyright violations
Links:
https://commons.wikimedia.org/wiki/Commons:Village_pump/Proposals#Proposal_…
Fae
--
faewik(a)gmail.com
https://commons.wikimedia.org/wiki/User:Fae