- From: Fadi Maali <fadi.maali@deri.org>
- Date: Tue, 25 Jun 2013 09:56:00 +0100
- To: "Simons, E.J. (Ed)" <e.simons@ru.nl>
- Cc: "public-gld-comments@w3.org Comments" <public-gld-comments@w3.org>
Hello, Thank you so much for your feedback regarding DCAT. I will try to summarise and reply to the points you raised in the attached document: 1. How to express dataset relationships such as "derived from" I totally agree that it is sometimes important to describe relationships between datasets. DCTerms provide a number of properties to describe such relations. Based on our survey of existing catalogs, such relations are rarely described in catalogs. This is not to say that they are not important and not to say that they cannot be described. As in other RDF data, properties from different vocabularies can always be used together. DCTerms properties as well as others can be used along with the DCAT ones. 2. Time-related properties don't allow stating what agent caused the event This can be described by W3C Provenance Ontology. I added a link to it from the DCAT Spec. 3. Catalog language is superfluous and ambiguous This property is used particularly by catalogues that federate a number of other catalogs which can possibly be in different languages. 4. No license attached to each individual dataset metadata record A license and rights statement can be associated with the catalog. It is stated in the description that this applies to the catalog records and not to the datasets. Additionally, license of each distribution can be described. The only case that cannot be described is when different records of a catalog has different licenses attached (please notice that I am referring to records and not the datasets themselves), we believe that this is a corner case. Addressing it requires removing the license property from catalog and requiring it on each record which complicates the common case where catalog records have the same license. 5. Link from CatalogRecord to Catalog not necessary; link to Dataset sufficient While the link can be inferred, making it explicit simplifies querying. Please let me know if I miss any of the points you raised. Thanks again for the feedback! -------------------------------------------------- Fadi Maali PhD student @ DERI Irish Research Council Embark Scholarship holder http://www.deri.ie/users/fadi-maali On 11 Apr 2013, at 15:48, Christopher Gutteridge <cjg@ecs.soton.ac.uk> wrote: > a polite -1 for defining classes for Person and Organisation when foaf: already defines these very clearly. > > The suggestions section sounds like a very useful *extension* but not part of the core and will muddy what's currently a nice tight(ish) looking specification. > > What's important is that dcat is something achievable and reasonably easy to get right and interpret. There's no reason that the CERIF world can't invent a vast number of extensions to it, but these should not be part of the core specification. > > The idea of a license on the metadata record (separate to the catalogue or dataset or distribution) makes total sense. If & when DCAT takes off there will be aggregated catalogues with records with different licenses, but this should be no more tricksy than a dcterms:license applied to a dcat:CatalogRecord. > > On 08/04/13 11:48, Simons, E.J. (Ed) wrote: >> L.s., >> >> On behalf of the euroCRIS Board, I, as President, herewith send you the response of euroCRIS to your Last Call on the Working Draft of the Data Catalog Vocabulary (DCAT). >> >> The comments and suggestions made in the document are based upon both the expertise of the authors in as well as their experience with Current Research Information Systems (CRIS) and the (meta)datamodel underlying these systems, more specifically the CERIF datamodel (Current European Research Information Format). The latter, an official recommendation of the European Union to its member states, encompasses the broad range of research information aspects - funding, input, researchers, organisations, projects, equipment, publications, underlying data(sets), ... - and includes specific linking and semantic technologies that, in our view, could bring substantial added value for the optimal description, discovery and therefore dissemination and (re)use of research datasets. In this respect the CERIF data model and its semantic technology might be of significant interest to your initiative. >> >> By sending you this document we, as euroCRIS Board, sincerely have the intention (as well as the hope) to contribute to an optimal creation of the DCAT and we kindly put ourselves at your disposal for further information and cooperation. After all, we, as parties working in the domain of research information, in the end all pursue the same basic goal: the optimization (by means of IT) of both scientific production and knowledge dissemination on a global scale. So let's join forces whenever we can. >> >> I wish you the best of success with the DCAT project. >> >> Yours sincerely. >> >> Ed Simons, Ph.D., >> Radboud University, NL. >> President of euroCRIS. >> >> >> >> >> >> >> >> > > -- > Christopher Gutteridge -- http://users.ecs.soton.ac.uk/cjg > > University of Southampton Open Data Service: http://data.southampton.ac.uk/ > You should read the ECS Web Team blog: http://blogs.ecs.soton.ac.uk/webteam/ > > Would you recommend the software you use to another institution? > http://uni-software.ideascale.com/ > >
Received on Tuesday, 25 June 2013 08:56:28 UTC