[ckan-discuss] CKAN package metadata: JSON output different from RDF output

Monika Solanki monika.solanki at gmail.com
Sun Mar 27 21:04:18 BST 2011


On 27/03/11 20:22, Friedrich Lindenberg wrote:
> Hi Monika,
>
> On Sun, Mar 27, 2011 at 9:15 PM, Monika Solanki
> <monika.solanki at gmail.com>  wrote:
>> I just discovered that the package metadata available for a CKAN package is
>> different in the JSON and the RDF version. For example, for the package
>> biolit, the RDF does not include the pointer to its void description while
>> the JSON does. Can this be fixed ?
> there's a ticket for it: http://trac.ckan.org/ticket/910 and I think
> Will committed to do this fairly soon, but I think there was some
> ambiguity over how arbitrary extras (and from CKAN's point of view the
> LOD cloud VoID stuff mostly falls into this category) should be
> represented: invent our own vocabulary and use extra names as
> predicates (e.g.<x>  ckan:my_extra "value") or if we should create
> BNodes with RDFS:label and RDF:value. What do you think?
That is an interesting question and prompted me to look into the void 
vocabulary to check if they have defined any predicate that takes as 
value the URI of the void document for the dataset. I could not find 
any. I do not favour BNodes personally and would prefer the former option.

Thanks,

Monika
> - Friedrich


-- 
Dr Monika Solanki
F27 Department of Computer Science
University of Leicester
Leicester LE1 7RH
United Kingdom
Google Coordinates: 52.653791,-1.158414
Tel: +44 116 252 3828
WebID: http://www.cs.le.ac.uk/people/ms491/foaf.rdf#monika
Web: http://www.cs.le.ac.uk/people/ms491/
      http://bit.ly/djvtiH






More information about the ckan-discuss mailing list