--- id: CON-PFM-ELG-003 name: Eligibility Decision feature: FEAT-PFM-ELG-02 product_area: platform module: eligibility revision: 1 status: draft product_intent: confirmed implementation_evidence: observed source_sha: services-platform: ad3f6a6582843008ed0f5a17b131878d15829af1 react-console: baf6dae382d6f41dce976c7f848667ccc37e076f definition: >- An eligibility decision combines a qualification state, the use’s policy outcome and an explanation for supplied subjects on a selected day. owned_by: MOD-PFM-ELG synonyms: - answer - evaluation result not_to_be_confused_with: - a permission grant - a population list - a test result facets: - name: Qualification and outcome meaning: >- The three states remain separate from denial, admission with review or deferral under policy. used_for: Distinguishing a known answer from a failure fallback. - name: Explanation meaning: >- Reason codes and permitted explanation levels describe the result with readable labels. used_for: 'Giving members, support and authors appropriate detail.' - name: Subject and date meaning: >- The question names a workspace subject and the day for version and date resolution. used_for: Evaluating the intended record in the intended temporal context. - name: Consistency choice meaning: >- Display, authorization and transactional modes have different freshness and recording needs. used_for: Choosing an answer suitable for presentation or consequential action. - name: Retained history meaning: An append-only record preserves important answers in monthly storage. used_for: Investigating decisions within the retention period. - name: Display cache meaning: >- Short-lived answers reduce repeated evaluation and can be invalidated by release changes. used_for: Serving repeated display questions efficiently. - name: Population work meaning: >- Durable jobs produce counts, subject pages, comparisons and exports for one version. used_for: Understanding the records a definition covers. - name: Draft impact meaning: >- A simulation compares a draft with the published expression over relevant subjects. used_for: 'Reviewing likely gains, losses and dependent uses before publication.' relationships: - to: CON-PFM-ELG-002 statement: A use supplies the uncertainty policy for its requested decision. provenance: observed source: 'services-platform:modules/eligibility/services/evaluate/policy.ts' - to: CON-PFM-ELG-001 statement: Stored decisions identify the expression and version used to answer. provenance: observed source: 'services-platform:modules/eligibility/db/eligibility-decisions.table.ts' lifecycle: [] acted_on_by: - CAP-PFM-ELG-003 - CAP-PFM-ELG-008 - CAP-PFM-ELG-009 - CAP-PFM-ELG-010 - CAP-PFM-ELG-011 - CAP-PFM-ELG-012 visible_to: - ACT-PFM-013 - ACT-PFM-024 - ACT-PFM-003 - ACT-PFM-025 evidence: - >- services-platform: docs/eligibility.md, docs/responsibilities.md, docs/attention.md; docs/milestones/PLAT-M11 to PLAT-M18 records and the PLAT-M11, PLAT-M15 and PLAT-M16 plans; docs/proposals/eligibility/17-decision-register.md, docs/proposals/responsibilities-and-attention/16-decisions.md; modules/eligibility/services/evaluate/**, services/population/**, services/simulation/**, db/eligibility-decisions.table.ts, eligibility-population-runs.table.ts, eligibility-simulations.table.ts. - >- react-console: components/eligibility/test/**, components/eligibility/governance/simulation-summary.tsx. supersedes: [] --- # Eligibility Decision ## What It Is An eligibility decision is the result of asking whether a person or organization qualifies in a particular context. Eligible and ineligible express a determined answer. Indeterminate means the engine lacks a dependable answer, perhaps because a subject is absent, a fact is missing or a provider fails. The use’s policy turns that uncertainty into the outcome the consumer can act on. Keeping state and outcome separate prevents a fallback denial from being mistaken for a successful negative evaluation. Each result identifies the expression and date used and carries reasons. Explanation detail varies with the caller. Many results serve display requests and are never retained; consequential decisions and a sample of display answers become historical records. The stored result allows support to examine what happened rather than rerun today’s expression against today’s data. ## What It Is Not A decision does not itself grant permission or write a business entitlement. Those actions belong to the consumer. A population run is a collection of evaluations with aggregate results, not a different eligibility state. An author’s trial produces an answer for inspection without establishing a historical decision record. ## How It Is Used Applications request answers for their registered uses, individually or in batches. Support readers inspect retained answers to understand a refusal and the expression behind it. Authors use population work and simulations to assess reach before a change. Operators observe decision measurements for rising uncertainty. Display caching supports frequent reads, but callers must choose the appropriate consistency when the answer authorizes work or participates in a transaction. ## Its Facets **Qualification and outcome.** The three states remain separate from denial, admission with review or deferral under policy. Distinguishing a known answer from a failure fallback. **Explanation.** Reason codes and permitted explanation levels describe the result with readable labels. Giving members, support and authors appropriate detail. **Subject and date.** The question names a workspace subject and the day for version and date resolution. Evaluating the intended record in the intended temporal context. **Consistency choice.** Display, authorization and transactional modes have different freshness and recording needs. Choosing an answer suitable for presentation or consequential action. **Retained history.** An append-only record preserves important answers in monthly storage. Investigating decisions within the retention period. **Display cache.** Short-lived answers reduce repeated evaluation and can be invalidated by release changes. Serving repeated display questions efficiently. **Population work.** Durable jobs produce counts, subject pages, comparisons and exports for one version. Understanding the records a definition covers. **Draft impact.** A simulation compares a draft with the published expression over relevant subjects. Reviewing likely gains, losses and dependent uses before publication. ## How It Relates The rule supplies the qualification expression, and the use supplies its context and policy. Security consumes answers for its own access decisions and derived-role work. Population jobs apply the same evaluation machinery across records and contribute simulation gains to recheck requests. Source records remain owned by the applications that register their conditions. ## What It Could Become The product should settle the permission to enumerate qualifying people before approving population planning. Run discovery, cancellation and cleanup also need attention. Display-record persistence can fail after an answer returns, so audit needs should account for that limit. A published consumer-process evaluator and future operational views could make results easier to use without overstating what exists now.