diff --git a/docs/PAPER1_NOTE.md b/docs/PAPER1_NOTE.md new file mode 100644 index 000000000..74fbff4d6 --- /dev/null +++ b/docs/PAPER1_NOTE.md @@ -0,0 +1,69 @@ +# Short-lifetime H0 persistence: statement and measurement guide + +[Read the manuscript](research-translation/20260930/MANUSCRIPT.md) · [Follow the proof chain](research-translation/20260930/PERSISTENCE_BRIDGE.md) · [Reproduce the experiment](research-translation/20260930/EXPERIMENT.md) · [Inspect exact sources](research-translation/20260930/SOURCES.json) + +This is a short reading companion to the existing planar manuscript. It explains the quantity being counted and how to compare it with data. It does not introduce another theorem or replace the manuscript's source-bound proof chain. Scientific effect: NONE. + +## One field and one observable + +The model is the centered, variance-one periodized Bargmann-Fock field on the two-dimensional flat torus of side 24. Its covariance is + +```text +K_24(z) = sum[n in Z^2] exp(-|z+24n|^2/2) + / sum[n in Z^2] exp(-|24n|^2/2). +``` + +Decrease the level through the ordinary superlevel filtration. A finite H0 bar born at a maximum M and killed at its merging saddle S has lifetime f(M)-f(S)>0. Exclude the essential class born at the global maximum; do not discard the longest finite bar. For Borel sets A contained in (0,infinity), define + +```text +mu(A) = 24^(-2) E[# finite ordinary superlevel H0 bars with lifetime in A], +nu(ell) = dmu/dell. +``` + +The area divisor appears exactly once. This expected intensity includes all birth heights and spatial separations. It is not normalized by the random number of bars in a realization. The source chain supplies the density; it is not assumed merely from a plotted histogram. + +## What the existing manuscript states + +At its recorded planar consumption scope, manuscript sections 2-6 assemble D1's repaired global elder-selector and lifetime results, D2's bounded remainder, and the SIDE24 coefficient expression: + +```text +nu(ell) = c_(2,24) ell^(-1/3) + O(1), ell -> 0+, +mu((0,t]) = (3/2)c_(2,24) t^(2/3) + O(t), t -> 0+. +``` + +The torus side and dimension are fixed before the lifetime limit. The remainder constant and valid lifetime cutoff are existential, not evaluated error bars. The SIDE24 enclosure is an arithmetic result for the coefficient expression; its interpretation as a persistence coefficient consumes the repaired parent chain. See the [manuscript's exact statement](research-translation/20260930/MANUSCRIPT.md#2-the-assembled-planar-statement) and [source manifest](research-translation/20260930/SOURCES.json), whose mathematical source cut is Math- `fa2e9909d8cca40d38b7dc2eadda6766bd0788e1`. + +The parent must be read with its congruence erratum, Section 9 replacement v1.1, wording W1 and embedding restriction `r < L/(4 sqrt(2))`. The [statement crosswalk](research-translation/20260930/CROSSWALK.md) and [completion audit](research-translation/20260930/COMPLETION_AUDIT.md) explain the consumed interfaces and remaining assembly limits. This guide records their existing scope; it does not perform a new acceptance review of those proofs. + +A compact birth/gap population is a different measure. Restrict its count and use its compact coefficient c_(B,K) together. Do not attach unrestricted c_(2,24) to that population. The enclosure of c_(3,24) likewise does not provide a dimension-three persistence proof for this planar exposition. + +## Counts, density and bin width + +Under the leading density law nu(ell)=c ell^(-1/3)(1+o(1)), integration gives + +```text +mu((0,t]) = (3c/2) t^(2/3) (1+o(1)), +mu((ell,2ell]) = (3c/2)(2^(2/3)-1) ell^(2/3) (1+o(1)). +``` + +Raw counts in proportional-width bins therefore also have exponent +2/3. For m independent field realizations and a bin [a,b), 0