[open-bibliography] Metadata for data dumps

[Rufus Pollock]

On 5 March 2010 15:37, Adrian Pohl <pohl at hbz-nrw.de> wrote:
> Hallo,
[...]
> How should I describe raw datasets? Does a vocabulary exist so that I can
> describe the relevant properties of a dataset (size, number of records,
> format, date of export, extent: whole database or part, query parameters if
> only a part is opened etc.)?

As others have already pointed out for *rdf* datasets the way to go is
probably VoID. If your dataset isn't rdf then it is a little more
complicated since VoID is very rdf/linked-open-dataset oriented. One
option is just to use dct:Dataset. At the moment we are using a
combination of these approaches for the RDF version of ckan data ...

Regarding the second question I think for something as major as
different licensing for different parts of the dataset you might want
to consider splitting the dataset up.

> I think it would be a good thing to describe an open raw dataset with an
> rdf-file so at least the dataset could be a Linked-Data-Resource as a
> whole... I've been searching for a vocabulary for quite a while but couldn't
> find one. Any suggestions or do we have to create a vocabulary ourselves?
>
> What kind of vocabulary uses CKAN? Are there any plans to describe the open
> datasets in rdf? Is the OKFN perhaps already working on a vocabulary,

See comment above. More info on the CKAN RDF setup can be found here:

  <http://wiki.okfn.org/ckan/doc/guide/rdf/>

Our aim is to reuse existing vocabs wherever possible ...

Rufus
-- 
Open Knowledge Foundation
Promoting Open Knowledge in a Digital Age
http://www.okfn.org/ - http://blog.okfn.org/