FAIR does not simply mean open: what makes research data reusable

A file needs lasting context

FAIR stands for findable, accessible, interoperable and reusable. It refers to data and accompanying descriptions that should be discoverable, accessible, compatible with other systems and reusable. A persistent unique identifier helps identify a dataset even if its technical storage location changes. Meaningful metadata explain what was measured, by whom, using which method and under what conditions. Without this information, even a carefully produced series of numbers remains difficult to interpret. A value of twenty could mean a temperature, a count or an already normalised signal. The principles explicitly address machine processing too. This does not mean that people become redundant. It means that standardised descriptions should reduce unnecessary ambiguity and help technical systems with searching and matching.

Accessible can also mean controlled

A common misconception equates FAIR with a free download. In fact, standardised access procedures may also require a login or authorisation. This makes sense, for example, for data requiring protection. What remains important is that the conditions are apparent and technically traceable. Wherever possible, metadata should remain available even when the actual data are no longer accessible. Interoperability, meanwhile, requires more than a widely used file format. Two tables do not automatically fit together if they use the same column names but different units or definitions of terms. Shared vocabularies and explicit relationships help make such differences visible. The scientific task is to ensure that meaning does not reside solely in the memory of the original team. Only documented context enables other groups to assess whether combining data is scientifically sound, rather than merely carrying it out successfully on a technical level.

Reuse requires limits and rights

Anyone wishing to reuse data must understand their provenance, processing and legal conditions. This includes clear statements on use, as well as information about any corrections or exclusions already made. FAIR therefore does not imply that a dataset is suitable for every new question. A sample may be properly documented yet unsuitable for a particular generalisation. Nor do the original measurement uncertainties disappear when data are reused. The principles are an established guide to data management, not a substitute for subject-specific quality assessment. Their practical implementation is often gradual and constrained by available resources. The greatest benefit comes when researchers plan documentation from the outset. Data then do not have to be painstakingly deciphered after a project ends, but remain connected throughout their creation to the methods and decisions that determine what they can tell us.

Keep exploring

• Diving measurement buoys: how Argo observes the ocean from within
• More than temperature: how measurement buoys capture biogeochemical processes
• Food from chemical energy: how deep-sea communities function