Which recordings can support a scientific result?

Publishing an original recording makes it inspectable. It does not automatically accept it as scientific evidence. NASA archive analysis and stationary consumer measurements have separate selection, instrument and uncertainty requirements.

NASA archive

The saved eligibility rules and interval inventory define the existing airborne selection. Whole-stretch model comparisons, failures and processing assumptions remain unchanged. Repeated or overlapping instrument streams do not create independent flights. See the archive guide.

Stationary recordings

The participant protocol replaces passenger-flight collection. A proposed scientific corpus would require identifiable physical units, verified slow-rate response, complete timed reversals, supported level/alignment, retained interruptions and a validated uncertainty model. Selection must use these properties before inspecting the reversal rate or agreement with latitude.

The current signed-rate report is exploratory and always abstains from a scientific decision. There is no calibrated stationary admission policy or population estimator yet. Do not apply earlier aircraft quality flags, thresholds or unit certificates to it. Location may be unknown for a useful instrument control; latitude is needed only for the later geographic comparison.

Keep originals immutable, duplicate units identifiable, failed attempts visible and repeated instrument/day/site groups intact. Eight blocks from one device are one recording, not eight independent confirmations.

Historical server rules

The curation guide documents the earlier passenger-flight registry and approval commands. It remains available for old archives. New stationary uploads are stored and may be processed diagnostically; registry approval does not promote them to the historical primary flight corpus.