I am a little bit confused about the overall DCAT-AP approach.
Sometimes it looks like a full operational linked data enviroment is being assumed, such as in the IDs discussion thread, but other times it looks like laxed linked data and/or semantic web requirements are being allowed or even promoted (I suppose that for wider compability, which is a good thing), such as at the controlled vocabularies discussion thread.
I know that this is not the place to discuss deployment issues, but a brief clarification about the expected operational enviroment for DCAT-AP implementations would be something nice-to-have as an introductory section for clarification on expectations.
More specifically such section should give response to questions such as:
- Is DCAT-AP intended to be used only in full linked data enviroments?
- Are DCAT-AP providers expected to follow also linked data and semantic web best practices?
- Are DCAT-AP receivers expected to be able to manage full RDF expressivity?
- Will be any chance for european data portals not implementing a full linked data operational enviroment of using DCAT-AP as a means for interoperability?
Comments
Perhaps the basic use case (see Section 3) could be a good starting point for this discusssion. It reads: "[DCAT-AP] intends to enable a cross-data portal search for data sets. This can be achieved by the exchange of descriptions of data sets among data portals.". So, in theory, this means that DCAT-AP only specifies in which format metadata is exchanged between data portals. It does not say anything about the operational environment per se. This also means that data portals should in any case not need to change their systems. For example, if data portals are based on CKAN this is fine.
The basic use case also clarifies the role of metadata brokers: “Metadata Brokers facilitate the exchange of description metadata between data portals by ensuring conformance to the DCAT Application Profile. They provide metadata harvesting, transformation, validation, harmonization, and publication services. The Open Data Support project funded by the European Commission will operate a Metadata Broker service for data portals in Europe.”
So far, there is no use case that requires data portals or metadata brokers to adhere to Linked Data design principles. What matters is that they can export (and import) description metadata in a format that (possibly through conversion) adheres to the specifications of the DCAT application profile.
The DCAT Application Profile is indeed not imposing a Linked Data solution on metadata publishers, consumers and brokers. The only requirement posed by the application profile is the need for them to be able to import and export RDF data, which if not already supported natively, requires minor effort.
In the Open Data Support project, we decided to go for a Linked Data solution, in order to enjoy the benefits offered by this technology. However, even in then, we do not require open data portals to change their modus operandi. The broker takes care of harvesting, transforming to RDF and linking the metadata of the datasets.
I think that the document should contain a paragraph in this issue in the scope section 1.2.
Will include in next version, additonal txt in section 1.2:
The Application Profile is intended to facilitate data exchange and therefore the classes and properties defined in this document are only relevant for the data to be exchanged; there are no requirements for communicating systems to implement specific technical environments. The only requirement is that the systems can export and import data in RDF in conformance with this Application Profile.
Included in final draft http://joinup.ec.europa.eu/release/dcat-application-profile-data-portals-europe-final-draft