DiSSCo Prepare Deliverable D6.4 - Implementation of the DiSSCo Data Management Plan (DMP) and ENVRI FAIR compliance of DiSSCo data services

Abstract

Data Management Plans (DMPs) are documents guiding actions and liabilities of research data management during the whole data lifecycle, including project outputs and involving the maintenance of correct attribution and provenance data. Following the approach of FAIR Digital Objects (FDOs) adopted by DiSSCo, we developed a provenance data model which preserves the operations acting upon DiSSCo’s core data model, the Digital Extended Specimens (DES). A DES is a specific FDO version which represents a particular geo- and biodiversity specimen in a natural science collection and binds - i.e. provides encapsulated links or bitstreams of - relevant derived information content about that physical specimen. Based upon the key objective of the FAIR principles referred to as “machine-actionability” - the capability of software agents or “machines” to handle data autonomously and appropriately with little or without human intervention - we adapted the concept of a “machine-actionable data management plan” (maDMP). The maDMP for DiSSCo is designed as a continuously self-updating structure which is an FDO itself and offers significantly more options over the traditional approach of research data management (RDM) driven by static documents and reports. The maDMP enables flexible integration of all relevant information in the data life cycle as well as the creation of downscaled reports containing structured data. Consequently, it facilitates data discovery and reuse, and supports automated evaluation and monitoring. Within this report, we present a prototype as part of DiSSCo’s specification for open Digital Specimen (openDS) which builds upon the openDS classes for provenance recording. This is supplemented by additional concepts from the W3C PROV Ontology (PROV-O) and type-specific operations providing timestamped records using JSON Patch (RFC 6902). To foster integration and reuse of research data managed in DiSSCo, we present an initial approach for the semantic alignment of the provenance and maDMP classes with corresponding terms in the W3C Data Catalog Vocabulary. On this basis, we discuss the relationship of approaches providing an overarching ontology as mapping target - one common ontology maps to many provider vocabularies - vs. the FDO approach using a common set of globally obtainable (Kernel) attributes - many-to-many mappings of different provider/registry standards using an agreed upon framework of specified types, profiles and Kernel Attributes - and compare both concepts with regard to achieving interoperability in hyper infrastructures like the European Open Science Cloud.