Three Operating Domains Get Portable Evaluation Field Guides

Release at a glance

Grid FYI now has three portable, domain-specific field guides for operating decisions that look simple at the final number but become consequential across sources, rules, time, review, and external authority. Each guide is a stable ten-page circulation view of a maintained web argument. Each has a paired use case that turns the architecture into a small, exact, synthetic evaluation fixture.

The three domains are deliberately different. A payment waterfall allocates accepted funds in an ordered calculation. A health-system capacity plan tests aggregate service capability across time, support, and movement gates. A deliverable commitment resolves usable inventory, planned production, transport capacity, and accepted demand. They share a discipline without pretending that financial release, clinical judgment, facility activation, quality release, freight tender, and customer commitment are interchangeable acts.

These are evaluation field guides, not implementation specifications or customer cases. Their value is that a team can carry an inspectable model, exact oracle, authority map, failure injections, acceptance questions, and public-source boundary into a working session without treating the PDF itself as proof that a product or operating process works.

Financial services: the arithmetic balances, but authority still matters

The finance guide begins with $1,000,000 of synthetic available funds. The represented priorities allocate $40,000 of a submitted $45,000 administrative expense because the rule caps it at $40,000; $300,000 of senior interest; a $100,000 reserve top-up; $400,000 of scheduled senior principal; $80,000 of subordinate interest; and an $80,000 residual. The exact baseline is:

$40,000 + $300,000 + $100,000 + $400,000 + $80,000 + $80,000 = $1,000,000

The $5,000 expense above the cap remains an unresolved exception. It does not disappear, enter another priority, or become approved because the sources-equal-uses invariant balances.

A corrected delinquency fact moves from 3.800% to 4.200% against an exact 4.000% trigger. At 3.800%, the $80,000 remains residual. At 4.200%, the trigger redirects that amount to senior principal, which becomes $480,000, while residual becomes zero and total uses remain $1,000,000. Boundary cases at 3.999%, 4.000%, and 4.001% expose the comparator directly. A simulated payment file that omits the $100,000 reserve row totals only $900,000 and must fail rather than inherit approval from the correct calculation.

The guide keeps term interpretation, source attestation, calculation preparation, independent review, exception disposition, treasury preparation, authorized release, bank response, settlement evidence, accounting, and reporting separate. Grid does not interpret an agreement, waive the $5,000 exception, authorize or transmit money, provide custody, determine accounting treatment, perform an audit, or establish that a payment is due.

Healthcare: a count becomes capacity only inside a supported window

The healthcare guide uses abstract aggregate service units, not people, patients, bed assignments, diagnoses, acuity classes, or care recommendations. Demand is 20 units for the full 18:00-22:00 horizon. Facility A contributes 8 supported units. Facility B declares 7, but its acceptance ends at 20:00, so it contributes zero to the full horizon. Facility C contributes 6. Reserve facility D declares 8 but initially contributes zero because activation, support confirmation, and two transport cycles remain unresolved.

The initial arithmetic is therefore:

8 from A + 0 from B + 6 from C + 0 from D = 14 usable

20 required - 14 usable = 6-unit shortfall

When D's support is confirmed and two cycles of four units complete before 21:00, 2 x 4 = 8 reserve units become represented in the plan. The total becomes 8 + 6 + 8 = 22, a two-unit margin. If one cycle is lost, D contributes only 4 and the total becomes 8 + 6 + 4 = 18, a two-unit shortfall. B's visible 7 never enters the full-horizon total because its corrected interval still expires early.

That arithmetic can expose a feasible aggregate option. It cannot diagnose, triage, prioritize, assign an individual, select a destination, activate a facility, compel acceptance, dispatch transport, direct care, or establish command. Facility, receiving, clinical, incident, transport, support, privacy, and security authorities retain their distinct responsibilities. A transport acknowledgment is evidence about a represented event, not proof of appropriate clinical action, completed care, or a health outcome.

Manufacturing: visibility becomes a commitment only after every gate

The manufacturing fixture evaluates an accepted commitment of 68 finished units by 17:00. Forty finished units are recorded, six are held, and four are already allocated, leaving 40 - 6 - 4 = 30 usable now. Component A supports 80 starts. Component B supplies 100 pieces at two per finished unit, supporting floor(100 / 2) = 50 starts. Five line hours at ten units per hour also support 50. The minimum represented production limit is therefore 50. After a five-unit release reserve, 45 planned units remain conditionally releasable.

Pre-transport availability is 30 + 45 = 75. Transport capacity before the cutoff is 72, so min(75, 72) = 72 units are deliverable under the fixture, a four-unit margin over the accepted commitment of 68.

The revision ladder then moves the binding constraint. Component B is corrected from 100 to 84, supporting 42 starts and 42 - 5 = 37 planned releasable units. Availability becomes 30 + 37 = 67, a one-unit shortfall. The quality owner later releases two held finished units, making usable finished goods 32 and total availability 32 + 37 = 69, a one-unit margin. Transport capacity then falls to 64, producing min(69, 64) = 64, a four-unit shortfall.

No calculation releases the two held units, changes a production schedule, approves a substitution, creates a purchase order, tenders freight, chooses which demand loses four units, accepts or amends a customer commitment, or proves pickup, receipt, or delivery. Inventory, quality, production, logistics, procurement, commercial, program, and integration owners retain those authorities.

The web is canonical; the PDF is for circulation

The Insights remain the canonical maintained reading surfaces. They carry the fuller argument, current editorial metadata, source context, internal links, and future correction path. The paired use cases remain the canonical operating walkthroughs. The PDFs are stable circulation artifacts for annotation, workshops, review, and offline discussion. A future substantive correction should update the maintained page and receive an attributable successor rather than silently changing the meaning of this release record.

The PDFs provide selectable text, designed tables and diagrams, evidence boundaries, public-source links, metadata, and links back to their canonical Insights. This release makes no claim that the PDFs are tagged PDFs, conform to PDF/UA, satisfy Section 508, satisfy WCAG, or replace accessibility testing of the complete intended experience. The web pages remain the preferred reading surface when a reader needs the maintained structure and navigation.

How to verify the release

1. Open each of the three canonical Insights, confirm its evidence level, review metadata, worked fixture, authority boundary, and download link, then follow its paired use-case route.
2. Download all three PDFs and confirm that each is a distinct ten-page, letter-size artifact with the expected title and Grid FYI authorship, nonblank selectable text, and no obvious table or text overflow.
3. Follow the canonical link from each PDF and test its eight public-source links. A functioning link proves navigation only; it does not validate the source's applicability or the guide's proposition.
4. Regenerate the guides from the deterministic domain-field-guide builder and run the collateral verifier. The checked-in and regenerated files should match in page geometry, normalized text, link targets, and expected metadata.
5. Build the static site and confirm that the three files reach the production output byte for byte, appear in the Resources library, and appear once in each applicable domain resource lane despite being linked from more than one maintained page.
6. Reperform the three fixture calculations independently. Confirm the exact threshold transition in finance, interval and cycle behavior in healthcare, and component, quality, and transport transitions in manufacturing.

Evidence and claim boundary

This release establishes that Grid FYI publishes three reproducibly generated editorial PDFs and connects each to maintained research and a substantive use case. That is a product-demonstration claim about collateral availability and inspectability. It is not evidence that the Grid runtime implements every depicted state, connector, permission, formula, or workflow.

The cited public sources establish selected context within their own scopes; they do not endorse Grid. Passing the synthetic fixtures can support only a bounded claim about the tested arithmetic, revisions, roles, failure cases, and environment. It cannot establish financial advice, transaction authority, custody, valuation, model validation, compliance, clinical decision quality, patient safety, facility readiness, command authority, physical inventory truth, quality conformance, delivery, security authorization, production suitability, service levels, savings, resilience, adoption, or improved outcomes.

The common standard across all three guides is therefore not automation. It is answerability: can a responsible reviewer identify the accepted sources, exact calculation, changing constraint, unresolved gate, role boundary, authorized act, external response, and evidence required before a larger claim is made?