# OUDEA and the FAIR Guiding Principles

OUDEA is an architecture and specification; FAIR is a set of guiding principles. OUDEA does not claim that an implementation or data object is automatically FAIR merely because it uses OUDEA.

The mappings below describe how OUDEA architectural concepts can support FAIR-oriented implementation.

| Problem | OUDEA model | FAIR / related principle |
|---|---|---|
| Distributed data is difficult to find. | Search operates over public index projections of DEIDs and standardized discovery metadata. Implementations may support text, domain, Class, supply-type and location filters. | FAIR F4 |
| Dataset links break when infrastructure changes. | A stable DEID identifies the entity while Find can resolve changing infrastructure. | FAIR F1 |
| Every source has a different access process. | OUDEA separates discovery, resolution and access into the consistent Search → Find → Exchange model. | FAIR A1 |
| Consumers cannot understand data before accessing it. | DEID metadata can expose Class, optional Schema, source/provenance descriptors, description, location and available supply metadata before Exchange. | FAIR F2 / R1 |
| Consumers may use the wrong data revision. | Supply items can carry version metadata; OUDEA v0.1 exchanges the latest authoritative version by default. Historical version-targeting is reserved for a later version. | FAIR R1 |
| Dependencies between data entities are unclear. | Qualified DEID References explicitly connect related or composed entities using roles such as `part_of` and `benchmark`. | FAIR I3 |
| Data cannot always be publicly accessible. | OUDEA can operate with authenticated or controlled access; the Matrix binding can provide authentication and authorization without requiring every payload to be public. | FAIR A1.2; FAIR does not mean open |
| Central platforms take control away from publishers. | Domain and supply authority remain with publishers while federation enables participation across infrastructure. | Related: Common European data spaces |
| Broken or incomplete sources are hard to identify. | Implementations may report availability, completeness and source health separately from claims about intrinsic data correctness. | FAIR does not itself guarantee data quality |
| Cross-organisation exchange requires bespoke integrations. | DEID identity, discovery, references, resolution and Exchange provide a shared public interoperability model. | Related: OECD data access and sharing |

Sources:

- FAIR Principles: https://www.go-fair.org/fair-principles/
- Original FAIR Guiding Principles publication: https://www.nature.com/articles/sdata201618
- FAIR is not equivalent to open and does not itself prescribe intrinsic data quality: https://www.go-fair.org/resources/faq/what-fair-is-not/
- Common European data spaces: https://digital-strategy.ec.europa.eu/en/policies/data-spaces
- OECD G20 Compendium on Data Access and Sharing: https://www.oecd.org/en/publications/g20-compendium-on-data-access-and-sharing-across-the-public-sector-and-with-the-private-sector-for-public-interest_df1031a4-en.html