Positive control
Let the search succeed.
At a strict 1.400 ms, the unchanged searcher must return exactly J1748-2446ad. At 1.500 ms it must return exactly three named records.
Not yet run in this session.
The result
matches / 4,393 records / 4,331 tested / 62 missing
A null can be a proof, if you name its walls.
This one has hard walls. The source archive and its database are pinned by SHA-256. The database header says catalogue 2.8.1 and 4,393 pulsars. The distributed psrcat program returns 4,331 finite periods and 62 missing periods. The predicate is only: finite, positive, and strictly less than 0.001 seconds.
The null is not a sensitivity-corrected survey of the Galaxy. ATNF compiles published spin-powered pulsars plus coherently pulsed anomalous X-ray pulsars and soft gamma-ray repeaters, while excluding accretion-powered systems. Faster objects may be difficult to detect or absent from the published record.
The operable sweep
Move the edge of nothing.
The horizontal control is logarithmic from 0.5 to 30 ms. Every move runs the same predicate over the selected catalogue domain. Use the endpoint toggle to expose strict versus inclusive semantics.
Objects returned
0
The result is the empty set.
Waiting for catalogue…
Each bar counts finite P0 values in a logarithmic interval.Threshold line: 1.000 ms
Planted witness
Put one inside the void.
Inject one labelled synthetic record at 0.800 ms into the real catalogue array. The same search must return exactly one. Remove it and the set must return to empty.
Unplanted. Real domain result: 0.
Empty through1.39595482 mswith strict < at the endpoint
First includedJ1748-2446adimmediately above that endpoint, or at it with ≤
Published spin716 HzHessels and colleagues, 2006
The harder question
Does zero in the catalogue constrain the sky?
Only after adding an assumption the catalogue cannot supply. Choose a comparison set P0 below a ceiling. If each were an exchangeable opportunity to observe a sub-millisecond object, zero successes gives an exact one-sided binomial upper bound. Then weaken relative detection efficiency, q, and watch that bound lose force.
Comparison n762
95% bound if q = 10.3924%
Adjusted bound / q0.3924%
pupper = 1 − 0.051/n; adjusted display = min(1, pupper / q)
Conditional, not astrophysical. Exchangeable detection is not established. Short spin period, orbital acceleration, dispersion, scattering, low flux, and eclipses can all change detectability. As q approaches zero, the adjusted bound approaches 100 percent and says nothing.
Raw-field audit
Two roads to the same period.
The catalogue can publish P0 directly or let the program derive it as the reciprocal of F0. Missing values are neither road. The derived browser artifact keeps that provenance for every record.
Direct raw P0…records
Reciprocal raw F0…records
Missing both…records
Checking printed P0 against 1000 / F0…
Reproducibility
The check, including the gaps.
Live inputs, computed counts, uncertainty, and free choices
Source identityLoading…
Artifact identityDerived locally from the official program output. No runtime fetch. The research verifier independently scans the shipped 4,393-row array.
CensusLoading…
MissingnessRecords without P0 remain null and are reported separately. They are not treated as zero and do not enter the finite-period denominator.
Predicate choicesThreshold, endpoint rule, subgroup, and witness state are live reader choices above. The certified headline fixes 1.000 ms, strict <, all records, unplanted.
Statistical choicesThe comparison ceiling and q are reader choices. Confidence is fixed at one-sided 95 percent (alpha 0.05). The binomial layer assumes exchangeable opportunities and is not a population estimate without a defensible selection model.
Catalogue boundaryATNF is an updated compilation of published objects, not a complete or uniformly selected Galactic survey. Its documented scope excludes accretion-powered pulsars.
Unexercised hereThe distributed psrcat.db identifies itself as a selected version-1 projection of the version-2 parameter store. This page did not enumerate other historical parameter versions. It also does not rerun telescope pipelines, model survey sensitivity, resolve the 62 missing periods, or establish that the catalogue contains every neutron star. Browser rendering and assistive-technology combinations are tested by the site supervisor, not by the data verifier.
Primary sources
What the null rests on.
- ATNF Pulsar Catalogue download
Catalogue package, program, database, GPL v3-or-later notice, attribution request, and current version. Retrieved 11 August 2026. The shipped derivative preserves a local licence and attribution notice.
- Manchester, Hobbs, Teoh, and Hobbs, 2005
The ATNF Pulsar Catalogue, Astronomical Journal 129, 1993. DOI 10.1086/428488. Defines catalogue scope, including the exclusion of accretion-powered systems.
- Hessels and colleagues, 2006
A Radio Pulsar Spinning at 716 Hz, Science 311, 1901–1904. DOI 10.1126/science.1123430. The paper reports the positive-control pulsar and discusses low flux and an eclipse fraction near 40 percent.
- Burderi and colleagues, 2001
Where Have All the Submillisecond Pulsars Gone? Gives a population-synthesis motivation for the empty interval and warns of computational selection effects in short-period binary searches.
- D'Amico, 1999
The Bologna submillisecond pulsar survey. Contrasts shorter theoretical limiting periods with earlier survey bias against ultrashort periodicities.