The Cross-Domain Interoperability Framework (CDIF) is a set of guidelines and practice for using domain-agnostic standards to support the interoperability and reusability of FAIR data, especially across domain and institutional boundaries. It is being developed in response to the need for agreements on the use of standards in FAIR implementations, to support all areas of research but in particular those grand challenge research questions that are necessarily interdisciplinary and cross-domain in nature. The core of the work is the identification of a set of common functions needed in FAIR implementations. Guidance is provided as to which standards can be used for the information required and how they can be implemented to better support interactions between FAIR systems.
CDIF does not aim to be a comprehensive description of FAIR functionality, nor does it compete with domain standards. It attempts to provide a core set of agreements, which integrate the generic aspects of FAIR (basic discovery metadata, structural aspects of data needed for processing and integration, basic access information, provenance, etc.) with links to domain-specific semantic components such as controlled vocabularies and ontologies.
The functional areas which CDIF will initially address include:
1.0: Discovery of “static” data sets (draft available at https://doi.org/10.5281/zenodo.10252564);
1.1: Discovery of queryable services which provide data;
2.0: Description of data intended for integration;
3.0: Description of controlled vocabularies and mappings between them;
4.0: Approaches to provenance and process description;
5.0: Guidance on publishing “universal” temporal and spatial information;
6.0: Providing metadata needed for use as ML training input;
7.0: Packaging sets of data and metadata for exchange.
The initial CDIF profile for the discovery of static data sets is now available for comment at https://doi.org/10.5281/zenodo.10252564. This gives a sense of both the content and level of detail which can be expected in other functional areas. Feedback is still welcome to: Simon Hodson simon@codata.org, Stephen Richard smrtucson@gmail.com, Arofan Gregory ilg21@yahoo.com.
CDIF aims to be practical, rather than conceptual, advocating widely used, existing standards and technology, and building on proven approaches within existing large-scale FAIR networks. In some cases, CDIF will provide detailed recommendations for specific standards (Schema.org, DCAT, ODRL, DDI-CDI, SKOS/XKOS, SSSOM, etc.) but in other cases there is no well-established practice to recommend. Here, CDIF will describe possible approaches without making specific recommendations. In line with a set of established design principles, the CDIF recommendations will change in response to emerging standards and practice in these areas and others. The initial draft of CDIF is an output of the WorldFAIR project, to be delivered by the end of May 2024.
CDIF has been developed based on input from the 11 case studies within WorldFAIR, cutting across a broad range of domains and disciplines. This input includes the case studies’ FAIR Implementation Profiles, a series of dedicated meetings, and the participation of all the case studies in two successive workshops hosted at the Leibniz Foundation’s Dagstuhl Centre.
Thirty invited experts have participated in drafting the CDIF guidelines, including members of many related FAIR initiatives and standards bodies, and has therefore, drawn on significant expertise both within and outside the WorldFAIR project. Just as CDIF is designed to complement and add detail to the EOSC Interoperability Framework, as well as work coming out of other EOSC-related projects, it also draws on a number of global developments represented in WorldFAIR. CDIF is designed to build on the work of such groups, complementing their efforts and making them practical for implementers in Europe and globally.
It is the intention that work on CDIF will continue, initially under the auspices of WorldFAIR and of a series of follow-up projects and initiatives (coordinated as WorldFAIR+). Such guidelines are only meaningful in relation to developments in the realm of standards and technology, and must be maintained accordingly.