Frequently Asked Questions - CLARIN Concept Registry
There is much debate about this and other question and there is no good universal answer yet. However, we need to start using the concept registry to find out how the definitions can be improved, which concepts are missing and which granularity should be chosen for metadata, morphology and semantic annotation to just mention a few examples.
No - it is just a start to offer a reference, so that users creating new resources could use the registered concepts and schemas describing legacy data can refer to them. But we will found that not all tag sets which are in use for various purposes can easily be mapped on another one. It also will largely depend on the intended usage. For searching an imperfect mapping may result in less precision, but for a researcher this may not be a problem.
Not yet, as in most cases they are dependent on theories and practical intentions. We're looking in to supporting multiple sets of relations for different needs. From CLARIN we intend to offer at least one set of relations with a large coverage, which users may want to use or manipulate.
The CLARIN Concept Registry (CCR) is an OpenSKOS instance, which implements the W3C SKOS recommendation and data model. It can be accessed via https://www.clarin.eu/ccr/
Currently it is filled with many concepts from for example the EAGLES project, various metadata initiatives and hopefully other sub-disciplines and initiatives. The national CLARIN Concept Registry Coordinators take care that the registry is not too fragmented and meets a number of criteria.
The CLARIN Concept Registry is a step in the direction of interoperability at the level of linguistic encoding (tag sets, metadata elements, etc.). The basic idea is to register all widely used concepts/terminology in CLARIN so that everyone can refer to them. The CLARIN Concept Registry is based on the W3C SKOS recommendation, which is a generic model not restricted to linguistics.