Not existing concepts [in DCAT-AP]:
12: to express real relationships
Raised in https://joinup.ec.europa.eu/asset/ogd2_0/issue/dcat-ap
DCAT-AP is based on RDF and RDF first of all is about relations between resources. It therefore is quite astonishing to see the requirement of "real relations" put forward as a reason against basing OGD on DCAT-AP.
To proceed it would help to know where the "real relationships"- requirement comes from. Which use cases are known?
Component
DocumentationCategory
improvement
Login or create an account to comment.
Comments
The codelist on relations give an example of what we want to express in addition to "is related":
The most frequent use case might be at dataset level:
"a dataset is used in an app" or "an app uses data from this dataset".
At ressource level the requirement comes from the statistical domain, where the predecessor (Vorgänger) and succesor (Nachfolger) is frequently used for timeline series but can hardly be mapped to the existing OGD 1.1.
Currently relationships are optional concepts but we see strong potential to foster reuse and findability when propagating those relations to the GovData federation.
DCAT and DCAT-AP are extensible.
This is an appropriate issue for the W3C DCAT workshop in Amsterdam. Preferably a proposal can be developed which is also usable in other parts of the world.
>> Not existing concepts [in DCAT-AP]: 12: to express real relationships
>> Raised in https://joinup.ec.europa.eu/asset/ogd2_0/issue/dcat-ap
>> DCAT-AP is based on RDF and RDF first of all is about relations between resources. It therefore is quite
>> astonishing to see the requirement of "real relations" put forward as a reason against basing OGD on DCAT->> AP. To proceed it would help to know where the "real relationships"- requirement comes from. Which use >> cases are known?
We now created a German extension of DCAT-AP and can use the dct:relation and the dct:hasVersion / isVersion properties.
Unfortunately the semantic “isPart” and “hasPart” is restricted to the dct:catalog level in DCAT-AP V1.1. We further did introduce the dct:type “collection” to say that datasets with this dct:type are part of a “collection”.
More relationships are not foreseen in the “dcat-ap.de V1.0 public”: http://www.dcat-ap.de/def/dcatde/1_0/spec/specification.pdf,
we will see whether a W3C dcat revision (https://joinup.ec.europa.eu/sites/default/files/event/attachment/20161025_dcat-ap_webinar_v0.04.pdf ) adds more semantically typed relationships to the generic “dct:relation”. A first step at DCAT-AP level towards additional relationships could be to introduce dct:hasPart and dct:isPart at dataset level.