A Discovery Package was developed along with a Discovery View. The package references non-DDI objects. This includes objects from the SKOS and other RDF vocabularies. The scope of the Discovery Package and the Discovery view is Disco. That being said, after review by the Discovery team it was determined that a second Discovery view was needed that supports researchers and librarians who are interesting in discovering, reviewing and comparing research data plans. Temporarily this view has been named rDisco. “r” is for research.
Review and update (by Flavio)
I populated the Discovery View in Drupal based on the objects from this UML diagram:
The diagram is part of the document in this link: http://rdf-vocabulary.ddialliance.org/discovery.html#
All objects in the view should be now in the New Objects for Discovery Package. There are two outstanding issues:
- There are many objects in that package that are not currently in the view and it's not clear how many of them are necessary. This needs to be sorted out by the modelers.
- There are still objects in the Discovery Package that need to be cleaned up: they should either be moved to the New Objects for Discovery Package and added to the view, or relocated elsewhere. This needs to be sorted out by the modelers in conversation with the Discovery team.
In the view I added comments with the mapping (develop jointly with Wendy) between DDI 4 objects and DISCO entities from the above diagram.
Most objects still need to pass a final content review.
Meeting Minutes Discovery group 2014-05-21
We discussed the need to rewrite the functional view description which is currently:
“Description: Discovery at study, data file, and data item level” functional view”
Kate volunteered to draft something .
We added Geographiclevel, and TimeMethod to the list of possible objects. The original list is at at http://www1.unece.org/stat/platform/display/DDI4/Discovery+-+presentations+and+docs
The group will review the revised list for completeness and rate the objects as either “core”, “extended”, or “exclude”
Jay raised the issue of whether this list includes what is necessary for aggregate data, covered by SDMX.
A use case might be a researcher wanting to find the microdata used to produce a table presented in a publication as well as the procedure used to produce the table from the microdata.
We also discussed looking at Data Without Boundaries, the DDI RDF Discovery vocabulary, and the NIH DDI to see which information objects they include for discovery.
We also discussed the possibility that information objects not in DDI (e.g. from GSIM, or Dublin Core) could be used in a functional view.
We should identify which elements can make use of a controlled vocabulary.
A recurring question in our conversation was: How comprehensive should the Discovery Functional View be and will it be used as a view or is it a collection that will be used in multiple other places in the model? Limiting the set of objects will be difficult.
What kind of search should we support:
Researchers looking for data?
Researchers looking for questions, represented variables, etc to reuse in new studies?
Archives looking for reuse or citation of data?
Coverage is not abstract.