Assessment 211: permanent ownership and query-paid codes

12 October 2026. The written native construction 69 handles unboundedly many learned contexts, including arbitrarily deep chains and late shorter contexts, under a checked immutable-code funding condition. This is substantive serial-reuse progress, not an independent-store join or root claim admission.

Strongest statement and proof mechanism

The greedy producer cannot create a descendant of an older context. Therefore the earliest context prefixing a target is its deepest matching context, and every produced target retains that owner forever. Selected subsets and retained source versions preserve this disjointness. The growing same-query OR diagnostic from assessment 210 applies to arbitrary overlapping kernels, but its conflicting true contributions cannot occur in this producer.

One current source index stores prefix-free creator codes followed by residual words. The original alias/residual class-span check and the new code bound imply a fixed composite-key class span, so the existing section 15 pending-window index supplies current selected answers. There is no uncharged per-context source-root lookup.

After the deepest context mark, actual residual bits service the code-plus-residual stream at a fixed faster rate. Backlog is exactly $\max(0,L-cn)$ after $n$ bits for code length $L$. If it remains positive at termination, the producer’s $L\le c|y|$ check certifies absence. Otherwise the inherited exact reader is caught up. Resets at deeper marks are sound by permanent ownership.

The ordinary producer is essential: its rolling window feeds a prospective context insertion capsule during the full target, and its creator code is built during the alias. A new composite word has length $3|v|+3+b$, whose class is prepared during that alias. Existing composite clocks are paid by residual service. Fixed tuples of FIFO pieces avoid delimiter concatenation. Full literal target comparisons enforce row degree without pointer identity.

Independent review and finite evidence

An independent adversarial reviewer derived permanent ownership and the stable-code approach separately, checked the exact backlog recurrence and the prospective composite-length clock, and supplied an audit with no primary-model imports. The maintained readable audit source uses actual right-record histories and compares literal-target unions with streamed encoded queries across retained selected versions.

The run passes 4,082 accepted and 3,136 rejected histories, 23,394 retained versions, 1,192,707 query comparisons, 252,321 deepest-mark switches, a 13-context descending chain, and 9,900 separate backlog checks. Cases include branching, late ancestors, duplicates, missing sources/aliases, exact-context targets, short targets and extensions. Code width is aligned with the written $\lceil\log_2(d+1)\rceil$ convention. These host-language checks establish finite semantic and abstract-scheduler evidence, not native capsule simulation.

The final independent adversarial review approved the complete written producer and query schedule relative to the stated inherited interfaces, with no blocking gap found. Caught-up service is tested by code-cursor exhaustion and residual-FIFO emptiness; the backlog formula is analysis, not native arithmetic.

The written construction explicitly inherits section 15’s dictionaries, strict queues, unary clocks and simultaneous-record compiler. Its own contribution is the producer geometry, stable global coding and input-paid delayed decoding. No external novelty claim is made. Native recognition gives a PEG for reversal of the complete checked ordinary-input language only.

Failed route, scope and next decisive test

A versioned contributor directory is unnecessary for this producer; arbitrary independently populated kernels still require a union mechanism. A later shorter alias cannot freely replace an old creator code: previously selected keys would change interpretation. The new theorem deliberately rejects residuals too short to pay that old code. The next test is removing this guard through paid stable recoding or another current-readout operation, beginning with one long creator followed by many short aliases and an immediate short negative query.

Bounded right degree and the residual length-class span remain checked restrictions. There is no context-count bound or common early route. Entry 129’s growing-alphabet matching and positive-path bounds remain imported results with their interface hypotheses. Replay slices, affine-fragment joins, persistent causal algebra and unary coherent cubical algebra retain their separate controls. Clipped-cone activation and independently prepared store joins remain open.

Continuity and publication

Custody is the canonical GitLab proof and audit; the existing Site receives generated readings and exact source downloads. The checkpoint, agenda, program entry and cadence move together. Root CLAIMS.yaml is unchanged. This is one mathematical construction, not one result for each supporting lemma or source projection.

Scheduled 115 to 116; mathematical 131 to 132, window three to four, reassessment 128 retained. Native construction 68 to 69. Next 117 is research; dedicated review and editorial duties both remain at 120. Build, scoped rendered/link checks, repository validation and deployment parity are recorded in integration state. Scientific admission is separate from those publication receipts.