Research principle
Unknown is a decision state, not an inconvenient blank
In high-assurance research, timeout, unsupported, and inconclusive outcomes should remain visible. Turning them into an implied pass makes the evidence easier to read and harder to trust.
The pressure to simplify
Technical review benefits from clear status. That can create pressure to collapse every result into pass or fail, even when the available evidence supports neither. In safety-critical work, that simplification changes the meaning of the record.
A timeout does not establish the property. An unsupported case does not establish the property. An incomplete run does not establish the property. Each should be recorded as unresolved and withheld from any deployment-eligible conclusion.
Why this builds trust
A conservative result may be less satisfying in a dashboard, but it is more useful to an engineer, reviewer, or grant evaluator. It identifies where evidence exists, where work remains, and where the system must refuse to overstate its confidence.
Fionn Labs uses this as a public research principle. The technical controls that implement it are part of the company's confidential work and are not described on this site.
The plain-language rule
When conclusive support is absent, the status remains unknown and the candidate is withheld. That is not a certification decision; it is evidence discipline for research conducted before any responsible authority or program makes one.