[ywz65]

This closes a terminal-output version of PR2199 Open next question2 in its declared local-call-only scalar projection. For selected calls time0 and time10, replace only their16 private centre values by zero and retain all11 chronological local calls through the second endpoint. The exact residual has source-only rank8 with target/helper dirt fixed, despite88 overlapping immediate scatter destinations. An explicit224-by8 repair map and a terminal minor of determinant1/64 certify sharpness. Allowing target values to vary raises the residual rank to16.

The package distinguishes the paired residual difference E(d,h)-E(0,h) from a complete dirty implementation: it supplies no free dirt offset, native512-coordinate lift, physical sharing invoice or changed exponent. Four bridge and four idle events are explicitly omitted. A concrete control shows that summing the original scatter increments predicts0 where the actual terminal residual is−1.

Validation: the registered Python3.12.7 suite passes exact source bindings/factorization/minors, five effective semantic controls, an independently implemented literal-event replay covering112 source columns and16 variable-target-minor columns (507,626 endpoint checks), a fixed-dirt/private-clear control, and Git-free/source-corruption/typed-receipt checks. A separate Codex agent reviewed the proof/code and independently reran both mathematical checkers under Python3.12.14 with no blocker.

Reproduce: python3.12 scripts/research_package.py --package research/centre-terminal-source-quotient.

Credit Adam Gall’s PR1918 projection and transcript lemma, thakrarsagar’s PR2199 census/open question, and the selected PR1094 construction lineage. Prepared by ywz65 with substantial OpenAI Codex assistance. No human review or maintainer admission is claimed. Exact overlap/provenance and remaining native/dirty-offset questions are recorded in the package.