The economics of unfinished work

Decision model · constructed example

Second Close

A first close records the operating position at a reporting cutoff. A second close revisits it after evidence, settlement, or permission changes. The earlier output can remain accepted while the next allocation changes.

In Records Framework 002’s worked example, ten outputs have been accepted and the original worker disappears. One unresolved effect remains. The successor has a limited review capacity and separately established permission to produce, investigate, or reconcile.

1. At the first close: the effect is uncertain. Under the model’s assumptions, investigating is the best next step.
2. After adverse evidence: the accepted outputs stay the same. The next step becomes reconciliation rather than new production.
3. After favorable evidence: the next step would instead be production. Evidence alone does not release an operational reservation; the modeled resolution is a separate event.

Investigation changes the evidence. It does not settle the effect. The original worker’s disappearance does neither. If production permission is withdrawn, that action leaves the feasible set even when it would otherwise have value.

This is a finite decision model using stipulated costs, permissions, signal reliability, and capacity. It applies established decision theory; it is not measured enterprise economics, a new economic law, or a claim of financing benefit. Full historical production cost remains unknown.

Read the worked example · Inspect its generated account · Read the model and assumptions

What the operating work establishes

Records Framework 001

Conventional, neutral, and Bulla-supported routes produced the same account. A Bulla-specific information advantage or total preparation-cost saving was not established. The work demonstrates a reporting convention over retained sources; independent reader usefulness and economic benefit remain unmeasured.

Its six-job cohort used synthetic inputs and induced interruption and error controls. Successor and underwriting memoranda were project-authored exercises, not independent recipient tests. Missing costs remain missing.

Read the original disposition and measurement boundaries →

Records Framework 002

A constructed finite decision model makes the choices inspectable. Its 26 constructed cases test conditional results under stated assumptions. They do not calibrate real prices, human effort, demand, or enterprise-wide costs.

Inspect the model results →

Records Framework 003

The main comparison did not run, establishing neither advantage nor tie nor loss. Twelve development roots completed; no main root opened. All 18 allocated main checkpoints remained unreached and all 72 prospective mechanical cells were unrun.

The selected report preserves the canonical study’s terminal disposition and import boundary. Development engineering does not substitute for a main comparison, independent successor use, or a total-burden measurement.

Read the selected report and unpassed gates →

Reliance Field 001: a pinned observation snapshot

At the retained September 24, 2026 snapshot, status was OBSERVING, with zero of six assignments enrolled and one observation run. Launch was September 24, 2026 at 06:50:29 UTC. This page is not a live status feed. Setup and validation are not field results; the empty cohort supports no comparative conclusion.

Read the dated snapshot →

The burden has to earn its place

A new verification step is added work unless a specified conventional obligation actually disappears. Acquisition, verification, interpretation, exception handling, fallback, and maintenance all belong in the comparison. File size, elapsed runtime, and model payoffs are not substitutes for that account.

Ordinary durable state or exported records may be sufficient. A more elaborate arrangement must show what useful work it replaces or enables for the recipient, and at whose cost.

Economics can reject machinery or deployment under the program’s constitutional commitments. That does not authorize dropping legitimate obligations, erasing receiptless claims, or treating an unverified allegation as permission for a penalty.