Elizabeth, thanks for a very informative answer!
You wouldn´t happen to have some links to sites where a "normal" person
could read about stuff like this so that those of us who tend to
accumulate data like this can actually make sense of these issues?
Karin (who is frequently confused regarding things like this)
On 25/05/16 18:43, E.W. wrote:
Hi Matthias,
A big part of my job is attempting to answer this question for
researchers at my (US) university and as part of on a team developing a
data repository.
I've done some analysis on the licenses that are used by datasets within
DataCite records. As of December 2015 when I scraped the data in, 59%
of the records had a rights statement and 95% of those were in the
Creative commons family. Yanking more out of my slides, when looking at
the CC uses: 62% CC-BY-NC; 36% CC-BY; 1% CC0; <1% other. These numbers
are heavily biased towards specific repositories using stock licenses
for all their records and having a high volume of records, so these
should not be interpreted as data representing the self-deposit data world.
CC has a nice wizard to select a license from, but CC0 or CCBY are
usually the ones we (the data repository team I work in) try to
recommend to people for open data. I can provided unapologetically
biased opinions about which to use, but I shall refrain unless prodded.
There may be a domain repository that specializes in this kind of data
and they likely have some recommendations. As far as adding it goes,
most repositories just have a declaration on the splash page for the
dataset, within the metadata, and sometimes a copy of the license as
part of the file set.
But to focus more on the third item, please do consider formally
depositing this into a data repository of some sort (versus just having
a public github repo). Zenodo has hooks to github and issues out
DataCite metadata when it generates the DOI. Figshare does this as
well, but Zenodo has better editing capabilities for the metadata. I'm
happy to brain dump about this more offline for the curious of if you're
confused as to how to use the elements (this is an open offer to anyone
on there wrangling with datacite metadata).
As far as other considerations about the question of making things
public, it depends on the source and content of the data.
1) Is work on the content creation and edit of these data files done?
You don't want to potentially be changing content under people's feet if
they are working with the data. There are ways to version the data and
I can expand on this if it is an issue.
2) Are there any data sensitivities? For example: Is this human subject
data? Could this potentially have a harmful impact on any subjects?
Looks like these are just models, so likely not, but always consider this.
3) Are there any contractual or licensing sensitivities for making this
open? For example, are these data files derived from a source with
restrictions on such derivatives? Any other contracts or IP issues with
tools used or the University in regards to licensing? University IP
concerns are highly variable by local laws and policies, but something
to consider if they would want to have a stake in this.
Just some things to chew on.
Elizabeth
(Data Curation Specialist, Research Data Service, University of Illinois)
On Wed, May 25, 2016 at 11:05 AM, Matthias Nilsson <matti...@chalmers.se
<mailto:matti...@chalmers.se>> wrote:
Hi,
I got a question at work today that I felt unable to answer, so I
thought I'd pass the question on to more knowledgeable people.
At my institution we have a set of metabolic models, which are
basically descriptions of reactions and metabolites and so on, stored
in an SBML[0] file. Internally, we have started to move them to
private Git repositories, but would now like to make them public.
As far as we can tell, there are no requirements from the institution
or the university on which type of license to choose, apart from that
the data should be "open".
So what I'd like to know is this:
1. What licenses are recommended for data? I've looked at Creative
Commons and Open Data Commons, but I suspect that there may be more.
2. How do we actually license things? Is it enough to add a file
called LICENSE to the repository and point to it in the README?
3. Is there anything else that we should consider when making the
transition from private to public?
Best regards,
Matthias
[0] A format based on XML.
_______________________________________________
Discuss mailing list
Discuss@lists.software-carpentry.org
<mailto:Discuss@lists.software-carpentry.org>
http://lists.software-carpentry.org/listinfo/discuss
_______________________________________________
Discuss mailing list
Discuss@lists.software-carpentry.org
http://lists.software-carpentry.org/listinfo/discuss
_______________________________________________
Discuss mailing list
Discuss@lists.software-carpentry.org
http://lists.software-carpentry.org/listinfo/discuss
