§6 RELEASE.md row 15 — proposed status

Proposed replacement for row 15's State cell (current text: "Numbers proposed and internally coherent; no corpus"):

Measured over the project's own corpus; deployment corpus invited via review period. docs/standards/VER-size-corpus.md and scripts/measure_record_sizes.py measure all five section 7 quantities across 87 of the 92 conformance fixtures (excluding the unparseable ver101 and the four fixtures — VER103/104/105/106 — built to deliberately trip one quantity at minimum size) plus the three published examples: N=90. The realistic corpus does not approach any budget (0.16% / 0.02% / 0.05% / 4.7% / 0.78% of (a)–(e) at the max observed), and a synthesized worst-realistic record — assembled from every distinct real preserved-metadata segment this repository holds plus inline vectors at both dimensions this project ships — still lands at 0.61% / 0.10% / 0.10% of (a)/(b)/(c). This is the project's own corpus, not deployment evidence: §7's numbers remain proposals strengthened by these measurements, not ratified by them. Still BLOCKING pending row 17's external review period, the named channel for deployment data.

This document does not propose closing row 15, and the row should stay BLOCKING until row 17 (external review period) has run and independent deployment data — or a documented absence of any objection to these numbers — is on record.

Scope note (31 August 2026): the corpus has since grown (86 fixtures at the dated run; 84 measurable then, plus additions from the 30–31 August audit response). The additions are small hand-authored records well inside every distribution above; re-running the command reproduces the current numbers.