[open-linguistics] Datahub multiple licenses

[Sebastian Nordhoff]

On Fri, 31 Aug 2012 16:06:38 +0200, Jimmy O'Regan <joregan at gmail.com>  
wrote:

> On 31 August 2012 12:02, Sebastian Nordhoff
> <sebastian_nordhoff at eva.mpg.de> wrote:
>> Dear list members,
>> I/we have added a number of resources to thedatahub.org
>>
>> * Glottolog (http://thedatahub.org/dataset/glottolog-langdoc)
>> * WALS (http://thedatahub.org/dataset/wals)
>> * IDS (http://thedatahub.org/dataset/ids)
>> * WOLD (http://thedatahub.org/dataset/wold)
>> * ASJP (http://thedatahub.org/dataset/asjp)
>>
>> Feel free to review an interlink.
>>
>> For Glottolog/Langdoc, I now have the problem that languoid data are  
>> CC-BY
>> whereas bibliographic data are CC-BY-NC.
>
> Bibliographic data? Are you including abstracts and/or summaries?
> i.e., is there anything in this dataset that can _be_ copyrighted?
> Looking at  
> http://thedatahub.org/dataset/glottolog-langdoc/resource/a1fab3c6-ad56-423f-84f5-57bd3fbea8fc
> I see nothing that can be (except, of course, when the collection is
> taken as a whole).

Hi Jimmy,
you are right that the individual records are probably not copyrightable.  
The situation is, however, more involved in the sense that there are  
cultural issues upstream (relating to Australian aboriginal culture) which  
need to be taken care of. The thing which probably assures best that this  
is taken care of is CC-BY-NC. I am aware of the fact that this is a hack,  
and that I could probably not enforce this, but it is nevertheless  
socially important.

No such sensitivities exist with regard to languoids, so there I can use  
plain CC-BY
Best
Sebastian

>
> (BTW, there's a quoting error in the 'dcterms:rightsHolder' line)
>
>> I am looking for a good way to
>> state this, and I would like to avoid having two datasets in the end  
>> (one
>> with languoids and CC-BY, the other one with refs and CC-BY-NC).
>> Any thoughts?
>
> Do you mean that you want to avoid maintaining two datasets, or that
> you want to avoid there being two datasets? The former is
> understandable, but the latter may perhaps be inevitable - it could
> even be worse, if multiple people take different measures to extract
> the public domain portion of your bibliographic data.