EPIC: Accessing cross-domain knowledge sharing approaches

[StroemPhi]

This task was defined in our WG charter as: "Assessment of existing approaches to achieve “cross-domain knowledge sharing”, e.g., minimum common metadata schemas"

As an epic task, this issue serves primarily as a hub to link to the more specific subtasks associated with this epic. Its comment section should thus be used to only discuss the scope of this epic, e.g. what do we mean by “cross-domain knowledge sharing” approaches, how should the assessment be organized, what should the outcome of this epic be specifically.

To start (propose) a discussion around a specific aspect/subtask of this epic, e.g. about a specific “cross-domain knowledge sharing” approach like the DataCite schema, CDIF or I-ADAPT, please use this issue template.

The work done in this epic forms the basis of the work done in epic #5.

[StroemPhi]

To me, it seems as if this epic needs to be reformulated a bit to make it clearer what it all encompasses. The way I currently understand it, is this. When we speak of a cross-domain knowledge-sharing approach we always speak of a schema that defines which knowledge must, should and can be shared. As a part such a schema will either be reusing terminology defined externally in some ontology or controlled vocabulary, or it will define the needed terminology itself. For example, the DataCite schema defines its terms internally and determines what must be known about a digital resource to allow denoting it with a persistent ID. DCAT-AP on the other hand, although having a quite similar scope as DataCite's schema, depends on having the terms defined externally in OWL ontologies.

So to me, it seems as if this epic is all about the finding and evaluation of common schemas (aka shapes or profiles), like we started in the discussions around research information and schemats like DataCite, DCAT-AP, CERIF and KDSF.

[StroemPhi]

From our call on the 13th of March:

• @joshmoore: Does the mapping of “schemas” also fit?
  • Jakob Voß: that opens a big can of worms. For achieving interoperability, then we should stick to RDF. How to get to RDF is out of our remit. In NFDI4Objects, only support 1-2 data formats (=specific schema; LIDO) and RDF.
    when we share between the consortia we should use only RDF
    Josh: additionally, re-using schemas may help us not need to map at all.