Topics of interest

The issues below is not an exhaustive list, instead it should be considered as a first attempt at bringing up issues for discussion that should be complemented.

Metadata profiles within data publishing

Today we see a variety of adaptations of DCAT-AP in the nordic countries. Based on previous discussions there are similar needs and perspectives that should be investigated. Clearly there is a lot of experience that can be collected and fed back into the standardization process at the level of W3C. There are also adaptations of PROF and SKOS that should be investigated.

Best practices for a national web of data infrastructure

What entities should be considered the fundamental building blocks of a web of data infrastructure? Clearly datasets (DCAT) are vital, but what about concepts (SKOS), specifications (PROF), information models (RDFS/OWL/?), public services (CPSV/?), APIs (DCAT again), data comments (DQV) etc.

Furthermore, what is the recommended way to interconnect these parts in a way that enables their metadata to be harvested into portals? Will Linked Data principles be enough or should there be simplified harvesting solutions, and what about those that feel that the base technology of RDF is too complicated. Should alternative non-RDF formats be established that can still be mapped to RDF, e.g. upgrading JSON into JSON-LD.

Best practices for domain specific data specifications

There is a wide range of domains where there is an interest for harmonizing data specifications. For instance the tourism sector, e-health, cultural heritage, geodata to name a few.

How to utilize linked data principles to improve data quality and interoperability

At the core of Linked data principles lies the concept of a thing identified via URIs that can be described using statements that together form a distributed graph. That this graph can be accessed in chunks using the HTTP protocol is also fundamental and transforms the information into a web of data. Clearly there are other ways to exchange information, but the principles of linked data are more than a transport mechanism. The principles provide a perspective on data that is both practical and useful when bringing information together without losing semantics, especially across organizational borders.

A useful perspective on expressing data is that even if the information is not expressed in one of the recognized RDF formats, the data can still be defined with reference to the established semantics of RDF vocabularies. An interesting possibility would be to establish principles for how to do information models with reference to established RDF vocabularies. Weather this is done in a formal language like RDFS, OWL, in SHACL, more loosely in application profiles or in flowing text in a specification remains to be seen. 

A very concrete issue that should be tackled is how to mint and utilize identifiers across systems, should URIs be preserved and when is it ok to introduce other identifiers.