FPF Reference

    Improvement-Oriented Quality Evaluation Question Framing

    E.22 · Stable · Part E - The FPF Constitution and Authoring Guides

    About this pattern

    This is a generated FPF pattern page projected from the published FPF source. It is canonical FPF content for this ID; it is not a FPF Reference product feature page.

    How to use this pattern

    Read the ID, status, type, and normativity first. Use the content for exact wording, the relations for adjacent concepts, and citations to keep active work grounded without pasting the whole specification.

    Status: Core.

    Use E.22 when someone is about to ask for a quality evaluation, quality review, returned-finding absorption, improvement proposal, or next-move hypothesis over an object version named by value, and the question needs to say what kind of evaluation is wanted before the evaluator starts.

    Relations

    E.22builds onEvaluation CharacteristicSpace Construction
    E.22builds onFPF Pattern-Quality Evaluation CharacteristicSpace
    E.22builds onDRR Decision-Adequacy Evaluation CharacteristicSpace
    E.22builds onFPF Pillar-Adequacy Evaluation CharacteristicSpace
    E.22coordinates withQuality Improvement Loop Method
    E.22coordinates withPattern Quality Gates: Review and Refresh Profiles
    E.22coordinates withLEX-BUNDLE: Unified Lexical Rules for FPF
    E.22coordinates withOntology-First Plain Technical Rewriting
    E.22coordinates withQ-Bundle: Authoring "-ilities" as Structured Quality Bundles
    E.22coordinates withCreativity‑CHR — Characterising Generative Novelty & Value
    E.22coordinates withExplore–Exploit Governor (E/E‑LOG)
    E.22coordinates withMulti‑Method Dispatcher & MethodFamily Registry
    E.22coordinates withParity / Benchmark Harness
    E.22coordinates withTelemetry-Driven Refresh & Decay Orchestrator
    E.22explicit referenceFPF Pattern-Quality Evaluation CharacteristicSpace
    E.22explicit referenceDRR Decision-Adequacy Evaluation CharacteristicSpace
    E.22explicit referenceFPF Pillar-Adequacy Evaluation CharacteristicSpace
    E.22explicit referenceQ-Bundle: Authoring "-ilities" as Structured Quality Bundles
    E.22explicit referencePattern Quality Gates: Review and Refresh Profiles
    E.22explicit referenceCreativity‑CHR — Characterising Generative Novelty & Value
    E.22explicit referenceNQD‑CAL — Open‑Ended Search Calculus
    E.22explicit referenceExplore–Exploit Governor (E/E‑LOG)
    E.22explicit referenceMulti‑Method Dispatcher & MethodFamily Registry
    E.22explicit referenceParity / Benchmark Harness
    E.22explicit referenceTelemetry-Driven Refresh & Decay Orchestrator
    E.22explicit referenceLEX-BUNDLE: Unified Lexical Rules for FPF
    E.22explicit referenceRelational Precision Restoration (RPR) — Kind-Explicit Qualified Relation Discipline
    E.22explicit referenceEpistemic Precision Restoration
    E.22explicit referenceLocal-First Unification Naming Protocol
    E.22explicit referenceMM-CHR — Measurement & Metrics Characterization
    E.22explicit referenceA.CHR-NORM — Canonical “Characteristic” & rename (Dimension/Axis → Characteristic)
    E.22explicit referenceA.CSLC-KERNEL — Minimal CSLC in Kernel (Characteristic/Scale/Level/Coordinate)
    E.22explicit referenceCharacteristicSpace & Dynamics Hook (A.CHR-SPACE)
    E.22explicit referenceDecision Theory (Decsn-CAL)
    E.22explicit referenceAgentic Tool-Use and Call Planning (C.Agent-Tools-CAL)
    E.22explicit referenceRole-Method-Work Alignment (Contextual Enactment)
    E.22explicit referenceU.Flow.ConstraintValidity — Eulerian
    E.22explicit referenceGateProfilization: OperationalGate(profile) (GateFit core)
    E.22explicit referenceEvidence Graph Referring (C-4)
    E.22explicit referenceTrust & Assurance Calculus (F–G–R with Congruence)

    Content

    Problem frame

    Use E.22 when someone is about to ask for a quality evaluation, quality review, returned-finding absorption, improvement proposal, or next-move hypothesis over an object version named by value, and the question needs to say what kind of evaluation is wanted before the evaluator starts.

    E.22 frames the question. It does not evaluate the object. The values, coordinates, statuses, and stop meanings come from the named object-under-improvement evaluation: for example E.21 for one pattern version, E.9.DA for one DRR, E.2.DA for an FPF-level object, C.25 for an engineering quality bundle, or another declared characteristic space, scale set, rubric, or review profile. E.19 is different: it supplies an admission or refresh review gate and findings profile. Use E.19 as the object-under-improvement evaluation only when the object being evaluated is an E.19 review-profile result itself. For one FPF pattern version, E.21 supplies the coordinate values and PatternQualityStatus; E.19 may later check that the E.21 result is valid, sufficient for the release seam, and not overread as project evidence, release, gate, assurance, or work.

    Not this pattern when the question is already scoped and one direct evaluation is enough. Run the object-under-improvement evaluation directly. Use E.23 when repeated improvement across passes is needed.

    First useful move: write a QualityEvaluationQuestionFrame naming the object version, the object-under-improvement evaluation, the purpose, the floor or improvement aim, protected trade-offs, expected evidence basis, and expected result form.

    What goes wrong if missed: "review this" can mean too many different things. A floor check may be mistaken for exceptional improvement, a review may suggest work without naming quality movement, absorption may count closed rows without re-evaluating the changed object, or a next-move suggestion may be overread as a decision, work plan, gate, evidence, assurance, or release.

    Primary EntityOfConcern in plain terms: the framed quality-evaluation question for one object version.

    Problem

    Quality evaluations fail when the evaluator has to infer the question. The same object can be checked for floor adequacy, improved toward exceptional expression, compared across trade-offs, mined for open questions, or evaluated after finding absorption. Those purposes produce different findings.

    The defect is not that reviewers need more ceremony. The defect is that an unframed question hides the object under improvement, the evaluation that supplies values, and the allowed shape of returned work.

    Forces

    ForceTension
    Cheap readiness vs ambitious improvementA floor evaluation should be short; exceptional improvement needs richer proposals.
    Explicit purpose vs reviewer discoveryThe request names the purpose, while the reviewer can still report important unasked questions.
    Evaluation vs next moveA useful evaluation may suggest a next move, but the suggestion remains a hypothesis until the pattern that governs the claim, relation, or boundary is applied.
    Multi-coordinate gain vs Goodhart riskRaising one visible value can damage usability, affordability, locality, source preservation, or corpus ecology.
    Proposal portfolio vs selected resultSeveral candidate improvements may be useful without becoming a selected set, pool policy, front insertion, parity, or refresh result.

    Solution

    E.22 gives one compact declaration for improvement-oriented quality evaluation questions. It keeps the question from replacing the evaluation and keeps the evaluation result from becoming a decision or work product beyond its authority.

    Local names and kind settlement

    Local nameKind and role
    QualityEvaluationQuestionFrameCompact declaration of the requested quality evaluation before it runs.
    ObjectVersionUnderQualityEvaluationExact object version being evaluated.
    ObjectUnderImprovementEvaluationRefExact evaluation pattern, characteristic space, scale set, rubric, review profile, or quality bundle that supplies values.
    QualityEvaluationPurposeSelectionRequested evaluation purpose or combined purposes.
    DeclaredQualityFloorMinimum acceptable coordinate or status floor when the frame declares a floor claim.
    DesiredImprovementAimRequested movement beyond the floor when improvement beyond the floor is requested.
    TradeoffProtectionSetQualities that must not silently worsen while visible values improve.
    ExpectedEvaluationEvidenceBasisEvidence loci the named evaluation must check or name for the requested purpose: object version, corpus/projection loci, source-currentness loci, comparator loci, worked cases, returned findings, or missing loci named by value.
    ExpectedQualityEvaluationResultFormThe result-row shape required by the named evaluation, including coordinate/value/short-rationale rows, mandatory attention-discharge profiles such as E.21 PrecisionRestorationProfile, and any evidence-locus or coordinate-specific payload fields. For E.21, this profile collapses word/head/use, phrase-apparatus, repetition/distribution, role-carrier, and pattern-application layers into affected-coordinate effects.
    QualityReviewFindingRowActionable row for a returned finding, expected movement, correction direction, and closure test.
    KindRestorationCheckRequired field for any finding or proposal whose correction direction changes wording, naming, or precision-restoration content: pre-repair kind/relation/slot-or-use-position/admissible use/scope, proposed post-repair kind/relation/slot-or-use-position/admissible use/scope, or not triggered/ordinary prose/already satisfied/blocker disposition with loci.
    CandidateImprovementProposalPortfolioBounded set of proposal rows returned by the evaluation when alternatives are useful.
    NextAdmissibleMoveHypothesisStop, repair, proposal, trade-off warning, outside-evaluation statement, new-frame statement, or governing pattern for a specific claim, relation, or boundary suggested by the evaluation. This is the proposed next improvement move, not a substitute for the evaluation result.

    These names frame and report quality evaluation. They do not select candidates, publish sets, plan work, certify evidence, approve release, or create new values.

    Quality evaluation purposes

    Purpose valueUse whenExpected result
    floorEvaluationThe question is whether the object reaches a declared floor.Values below floor, first repair, architecture hold, refresh, new-frame assignment, or admissible stop.
    exceptionalImprovementEvaluationThe floor is reached and the requester wants non-dominated improvement toward exceptional expression.Per-coordinate proposal or no-candidate disposition.
    paretoTradeoffEvaluationA candidate change may improve some values while worsening protected qualities.Trade-off account and non-dominated comparison.
    candidateImprovementProposalEvaluationThe requester needs candidate-change proposals before changing the object or generating variants.Proposal row or bounded proposal portfolio with expected evaluation movement.
    openQuestionDiscoveryEvaluationThe requester wants important unasked questions surfaced.Question classified as existing-coordinate issue, candidate future coordinate, or outside-evaluation issue.
    absorptionEvaluationReturned findings or suggestions have been applied or rejected.Quality-impact account over the changed object.

    Purposes can be combined, but the result keeps them distinguishable. A floor result does not answer exceptional improvement. Absorption count is not quality movement. A proposal is not a selected work item.

    Question frame

    QualityEvaluationQuestionFrame:
  Object version under quality evaluation: <object version named by value>
  Object-under-improvement evaluation: <exact evaluation>
  Evaluation purpose selection: <floor | exceptional | tradeoff | proposal | open-question | absorption | combined>
  Declared quality floor: <floor and scope, or evaluation default>
  Desired improvement aim: <floor-only | raise toward exceptional | compare variants | propose candidate changes | discover questions | absorption impact>
  Protected trade-offs: <usability | affordability | locality | corpus ecology | neighbour fit | source preservation | other property named by value>
  Expected evidence basis: <object, corpus, source, comparator, worked-case, returned-finding, projection, or missing loci named by value required by the named evaluation and purpose>
  Expected result form: <named evaluation's result-row shape | finding rows | proposal rows | trade-off table | open-question list | absorption-impact account | next-move hypotheses>
  Non-use boundary: <what this result must not decide, certify, publish, plan, execute, or prove>

    The shortest floor frame may name only object version, object-under-improvement evaluation, purpose floorEvaluation, and declared floor. The named evaluation still runs its required coordinate set and returns the result-row shape, evidence basis, rationales, coordinate-specific payloads, and mandatory attention-discharge profiles required by that evaluation. For one FPF pattern version under [E.21](/generated/patterns/E.21), compactness never permits omitted coordinates, missing ShortRationale, absent PrecisionRestorationProfile, inactive/triggered-coordinate shortcuts, scope narrowing, or a blocker-only substitute result.

    The frame does not authorize post-hoc scope replacement. If the requested floor is landing-input, corpus-facing, Stable, release, external-review, or another stated use, the evaluator measures that use. If a different use becomes interesting, open a new QualityEvaluationQuestionFrame; do not report the current request as passed under an easier scope.

    Finding and proposal rows

    An actionable finding has this shape:

    QualityReviewFindingRow:
  Review locus: <where the issue was found>
  Object locus: <where the object would change>
  Evaluation effect: <coordinate/status/floor/protected quality/outside evaluation>
  Current value or status: <if known>
  Expected movement: <repair floor | raise toward exceptional | prevent loss | classify outside>
  Correction direction: <what should change>
  Kind restoration check: <required when wording, naming, or precision restoration changes text; otherwise `not triggered`, `ordinary prose`, or `already satisfied` with loci>
  Closure test: <what evidence would close the row>

    A proposal row uses the same shape plus expected trade-offs and the governing pattern for any outside claim, relation, or boundary when needed. One edit may close several rows, but each row keeps its own disposition and closure evidence.

    For wording, naming, and precision-restoration proposals, the correction direction is not "replace X with Y". It must state the recovered object kind, relation, slot or use-position when live, admissible use, and scope before and after the change, or state not triggered, ordinary prose, already satisfied, or blocker with loci. A proposal that only removes the suspicious word, that leaves the text unchanged without by-value discharge, or that narrows one kind into another without an accepted decision, is a finding, not a closed repair.

    Absorption impact values

    Absorption impactMeaning
    coordinateImprovedA named coordinate or status has stronger content evidence after the change.
    floorOnlyClosureA below-floor defect was repaired enough for the floor but not exceptional expression.
    unchangedBecauseAlreadySatisfiedThe suggestion was already satisfied by value, with loci named by value and the evaluation property it already satisfies.
    tradeoffIntroducedA repair raised one property and damaged another.
    qualityLossDetectedThe applied or proposed change lowers a value or protected quality.
    outsideObjectUnderImprovementEvaluationThe suggestion belongs under another exact evaluation or pattern.
    notAdmissibleForDeclaredUseThe suggestion is rejected for the declared purpose and boundary.

    The absorption result is quality movement under the object-under-improvement evaluation, not a count of accepted rows.

    OEE/NQD and proposal portfolios

    When the object is a candidate, archive/front member, selected set, parity report, refresh report, or declared transduction result, E.22 can frame the quality question and return proposal rows. C.17, C.18, C.19, G.5, G.9, and G.11 keep authority over candidate characteristics, archive/front semantics, pool policy, selected-set publication, parity, and refresh.

    Worked slices

    Floor evaluation. A reviewer is asked whether one pattern is ready for ordinary use. The frame names E.21, purpose floorEvaluation, the declared floor, and the expected E.21 result form. The result is a complete E.21 coordinate table with ShortRationale and EvaluationEvidenceBasis, not a narrative "looks fine."

    Exceptional improvement. A pattern already passes the floor. The frame asks for non-dominated improvements toward 5 while protecting usability and related-pattern fit. The result returns proposal rows for missing worked cases and source-currentness, plus no-candidate dispositions for coordinates already strongly expressed.

    Absorption. External review returns many suggestions. The frame asks for absorptionEvaluation. The result says which changes improved coordinates, which were already satisfied, which introduced trade-offs, and which belong outside the evaluation.

    Proposal portfolio. A candidate improvement campaign needs alternatives before editing. The frame asks for candidateImprovementProposalEvaluation. The result returns bounded proposal rows; selection or generation stays with the pattern that governs that claim and is not decided by the evaluation frame.

    Bias annotation

    This pattern biases FPF toward asking the quality question by value. The bias is useful because unframed review requests often produce plausible but wrong answers.

    The bias is bounded. E.22 does not supply quality values, run repeated improvement, publish selected sets, decide work, or certify project claims.

    Conformance checklist

    CheckRequirement
    CC-E22-1Name the object version and object-under-improvement named by value evaluation.
    CC-E22-2State purpose, declared floor or improvement aim, protected trade-offs, and expected result form.
    CC-E22-3Keep the object-under-improvement evaluation as the source of values and required coordinates.
    CC-E22-4Represent actionable returned work as row-level findings or proposal rows with expected quality movement, closure tests, and KindRestorationCheck when the row proposes wording, naming, or precision-restoration repair. When a relation/signature/lens slot or use-position is live, the row cites the governing pattern named by the evaluation or restoration result; E.22 frames the improvement question and does not restate that ontology.
    CC-E22-5For absorption, report quality impact on the changed object, not only applied/not-applied disposition.
    CC-E22-6State a compact declarative non-use boundary when the result might be overread as decision, work, evidence, assurance, gate, release, certification, publication, parity, refresh, or selected-set authority. Keep the result on the evaluation question and name only the specific outside claim plus the pattern that governs it when one is needed; precision-restoration or phrase-apparatus issues belong to the named evaluation profile and F.19, not to a local boundary catalogue.
    CC-E22-7State what became worse when a proposed or applied improvement raises visible values.
    CC-E22-8Send repeated improvement to E.23 after one framed evaluation returns findings or proposals.
    CC-E22-9Name the expected evidence basis and result-row shape from the object-under-improvement evaluation; E.22 cannot authorize omitted coordinates, missing rationales, missing mandatory attention-discharge profiles, missing PrecisionRestorationProfile when E.21 is used, unchecked loci, inactive/triggered-coordinate shortcuts, scope narrowing, or a weaker result form.

    Common anti-patterns and repairs

    Anti-patternRepair
    "Review this" prompt. The evaluator infers purpose.Add a QualityEvaluationQuestionFrame.
    Floor pass sold as excellence. Readiness is mistaken for exceptional improvement.State exceptionalImprovementEvaluation if wanted.
    Frame replaces result. The question frame names a purpose but returns prose, a two-column value table, or proposal rows without the named evaluation's result form.Re-run the named evaluation and return its required coordinates, evidence basis, rationales, and payload fields.
    Scope laundering. The frame asks one use, but the result answers an easier, local-only, diagnostic, or evaluator-selected use.Re-run the named evaluation under the requested use; if another use is needed, open a new frame rather than saving the current result.
    Applied-count absorption. Closure count replaces quality movement.Re-evaluate the changed object and classify impact.
    Goodharted improvement. Visible values rise while protected qualities worsen.Add trade-off protection and reject dominated changes.
    Recommendation as decision. A next-move hypothesis is treated as chosen work.Open the exact decision, work, publication, parity, refresh, evidence, or assurance pattern if that claim is needed.
    Lexical repair request. A finding says only "replace this word" or "avoid that wording."Rewrite the row as a precision-restoration finding with pre/post kind, relation, admissible use, and scope; if no kind-preserving repair is recoverable, leave it blocking.

    Consequences

    ConsequenceBenefitCost
    Review requests become typed.Evaluators answer the intended quality question.Requesters must name the object and evaluation.
    Exceptional improvement becomes explicit.Reviews can propose non-dominated improvements rather than stopping at floor defects.Protected trade-offs must be named.
    Absorption becomes quality-aware.Follow-up says what improved or worsened.Row discharge alone is not enough.

    Rationale

    There is no neutral generic request when a quality result is wanted. The useful artifact is the framed question: object version, evaluation, purpose, expected evidence basis, expected result form, and boundary. This keeps review helpful without turning it into process control or project authority.

    SoTA-Echoing

    ClaimCurrent or retained source lineLocal adoption
    Quality evaluation should be multidimensional, diagnostic, and actionable.Current rubric and long-form evaluation work, including multidimensional LLM rubric evaluation and meta-evaluation lines, treats rubric validity and actionable feedback as current problems.Findings name evaluation effects, expected movement, correction direction, and closure tests.
    Feedback needs desired condition, current condition, and next action.Hattie/Timperley and Sadler lineage, retained through current feedback-evaluation work.The frame states floor or desired aim, current evaluation object, and expected result form.
    Evaluation questions must derive from purpose.GQM lineage and current task-specific rubric evaluation work.QualityEvaluationPurposeSelection precedes values.
    Multi-criteria improvement needs trade-offs and non-dominated alternatives.MCDA, Pareto, ATAM lineage plus current architecture trade-off evaluation work.paretoTradeoffEvaluation and TradeoffProtectionSet prevent one-score closure.
    Proxy optimization can degrade intended value.Goodhart taxonomy and current proxy/reward/rubric failure work.Findings ask what became worse and keep popularity, review count, and discharge count out of values.
    OEE/NQD needs proposal-shaped quality pressure before candidate change.Current quality-diversity and open-ended exploration lines.Proposal rows name expected quality movement before generation or selection neighbours consume them.

    Relations

    PatternRelation
    E.21Supplies pattern-quality values and required coordinates.
    E.9.DASupplies DRR decision-adequacy values and required coordinates.
    E.2.DASupplies FPF Pillar-adequacy values.
    E.19Supplies admission or refresh review profiles when that is the evaluation.
    E.23Governs repeated improvement after framed evaluations return findings or proposal rows.
    E.10, A.6.P, C.2.P, F.18Repair load-bearing wording and names introduced by frames or findings.
    C.16, A.17, A.18, A.19, C.25Govern characteristics, scales, measurements, characteristic spaces, and quality bundles.
    C.17, C.18, C.19, G.5, G.9, G.11Govern OEE/NQD candidate, archive/front, pool, selected-set, parity, and refresh claims.
    C.11, C.24, A.15, A.20, A.21, A.10, B.3Receive decision, call-planning, work, gate, release, evidence, and assurance claims when a quality result is reused beyond evaluation.

    E.22:End

    Last Updated: 2026-06-05 — this section last modified in upstream FPF commit eaafd3a4 (github.com/ailev/FPF)