The FAIR Principles: What Open Data Actually Requires

Posting a file somewhere is sharing, but whether anyone, or any computer, can find it, open it, understand it, and reuse it correctly is a separate matter. The FAIR principles name what usable actually requires: findable, accessible, interoperable, and reusable. Notably, accessible does not mean open, so data behind a clear application process can still be FAIR.

Why sharing data is not the same as making it usable

Open data has become a slogan, and like most slogans it hides a harder question. Posting a spreadsheet somewhere is sharing. Whether anyone, or any computer, can find it, open it, understand it, and reuse it correctly is a separate matter. A file with no labels, in a format nobody can read, behind a dead link, is technically shared and practically useless.

The FAIR principles were written to name what usable actually requires. FAIR stands for findable, accessible, interoperable, and reusable. Their framers had machines in mind as much as people, because the volume of modern research data is too large to curate by hand.

Findable

Data cannot be used if it cannot be located. Findable means the dataset carries a globally unique and persistent identifier, the kind that keeps pointing to the same object even as websites are redesigned. It also means rich metadata, a description of what the data are, who made them, and under what conditions, and that this metadata is indexed somewhere searchable.

The persistent identifier is the quiet workhorse here. A normal web link rots. An identifier such as a DOI is designed to resolve to the same resource for the long term, which is what makes a citation to data trustworthy far into the future.

Accessible, and why accessible is not open

Accessible has a specific and often misread meaning. It says that once you find the data, you can retrieve it through a standard, open protocol, and that the metadata stays available even if the data themselves are later restricted or removed. It does not mean the data must be free for anyone to download.

This is the point most people miss. Sensitive health data may sit behind an application and a data use agreement. That can still be FAIR, as long as the rules for access are clear, standardized, and the description of the dataset remains public. FAIR governs how data are described and reached, not whether every dataset must be wide open.

Interoperable

Interoperable means the data can be combined with other data without a human hand-translating everything. That requires shared vocabularies and formats, so that a field labeled a certain way in one dataset means the same thing in another, and so that software can join them.

In medicine this is where much of the friction lives. When one registry codes a diagnosis one way and another codes it differently, merging them is slow and error-prone. Interoperability asks researchers to use community standards and controlled vocabularies up front, so the data speak a common language later.

Reusable

Reusable is the goal the other three principles serve. It means the data are described in enough detail, with an explicit license and a clear record of where they came from and how they were processed, that someone else can use them appropriately and know the limits.

Provenance is central. A number without its context, how it was measured, on whom, with what instrument and what cleaning steps, can be reused wrongly precisely because it looks simple. A clear license matters too, because data whose terms of use are unstated leave a would-be user unsure whether they are even allowed to touch it.

FAIR in the clinic, and its limits

FAIR has moved from data science into clinical publishing. Journals following the major editorial recommendations now ask trials to include a data sharing statement, declaring whether individual participant data will be shared, what will be shared, and how someone can request it. Undecided is not accepted as an answer.

Two honest caveats belong here. FAIR is about machine-readable stewardship, not a promise that data are correct or that a study is sound. And FAIR is deliberately not the same as fully open; it is compatible with the real need to protect participant privacy. Read a data sharing statement as a statement of intent and access rules, not as proof that the underlying evidence is strong.