#Discipline‑CHR - Field Health & Structure

C.21 · Cluster C.IV - Composite & Macro-Scale · Stable · Part C - Kernel Extension Specifications

#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.

Purpose. Give FPF a typed, reviewable way to characterize the health, maturity, and structure of a scientific or engineering discipline, without collapsing into taste, anecdotes, dashboard views, audit labels, or single-number scores. The pattern defines a portable set of Characteristics and guards (legality, freshness, scope) that any Context can specialize.

This pattern supplies the CHR vocabulary of health for disciplines. C.20 composes the discipline; C.21 declares discipline-health characteristics and admissible readings; Part G may publish SoTA palettes or time-series views; Bridges keep cross-context meaning honest; penalties touch R only.

Status & placement. Part C (Kernel Extention Specifications) → Cluster C.I (Core CHRs/CALs). Depends on: MM-CHR (C.16), KD-CAL (C.2), USM/Scope (A.2.6), Trust & Assurance (B.3), E.10 (LEX‑BUNDLE). Coordinates with: C.20 Discipline‑CAL (what a U.Discipline is), G.2 (SoTA palette), G.12 (dashboard), G.0 (CG‑Spec registry).

FPF treats disciplines as first-class holons (see C.20): they aggregate epistemes, practices, standards, institutions, and observed Work. Teams routinely say “the field is fragmented,” “standards are converging,” or “replication is improving,” but these claims are rarely typed (scale/unit/polarity) or replayable (evidence lanes, freshness, scope). C.21 supplies the CHR vocabulary: named Characteristics with CSLC typing, so discipline-health claims can be compared admissibly (CG-Spec) and monitored through time (G.12) when a project needs that use. Each published value declares ReferencePlane ∈ {world|concept|episteme} and DisciplineId (U.Discipline@UTS); cross-plane use applies CL^plane in Assurance (penalty to R_eff only).

#Keywords

• discipline
• field health
• reproducibility
• standardisation
• alignment
• disruption.

#Relations

#Content

#Problem Frame

FPF treats disciplines as first-class holons (see C.20): they aggregate epistemes, practices, standards, institutions, and observed Work. Teams routinely say “the field is fragmented,” “standards are converging,” or “replication is improving,” but these claims are rarely typed (scale/unit/polarity) or replayable (evidence lanes, freshness, scope). C.21 supplies the CHR vocabulary: named Characteristics with CSLC typing, so discipline-health claims can be compared admissibly (CG-Spec) and monitored through time (G.12) when a project needs that use. Each published value declares ReferencePlane ∈ {world|concept|episteme} and DisciplineId (U.Discipline@UTS); cross-plane use applies CL^plane in Assurance (penalty to R_eff only).

#Problem

Narrative health claims cause three recurrent failure modes:

1. Illegality. Averaging ordinals, mixing units, or comparing incommensurate Contexts ⇒ nonsense roll-ups.
2. Staleness. Health “scores” rarely declare freshness windows or evidence lanes (TA/VA/LA).
3. Scope slippage. “The field” is left implicit; cross-Context reuse lacks Bridges & CL, leading to silent semantic loss. Any numeric comparison or aggregation cites a CG-Spec row (characteristics, ScaleComplianceProfile (SCP), Γ-fold, MinimalEvidence) before computation.

#Forces

#Solution — Discipline Health Characterisation (DHC)

#Ontology quick sheet (normative, clarifying)

What “DHC” is. DHC is a CHR vocabulary pack that defines Characteristics + Scales/Units/Polarity for discipline health; it is not a document or a run. Artifacts. • U.DHCPack (I-lane name; published as an episteme): the slot set (Characteristic/Scale declarations) for a Context. • U.DHCMethodSpec (S-lane): the computational specification(s) for deriving each DHC slot (e.g., replication‑window definition, CD‑index class), table‑backed; multiple per slot allowed, editioned separately. • U.DHCSeries (episteme w/ EditionSeries): a time‑indexed publication of computed DHC readings for a Discipline×Context, each value bound to …Ref.edition for every referenced method/metric/distance. Edition subjects. (i) DHCPack.edition — when the slot semantics (Characteristic/Scale) change. (ii) DHCMethodSpecRef.edition — when a computation method (formula/class/policy) changes. (iii) DHCSeries.edition — when the published series changes its content (not carriers). Publication. Releases are Work on Carriers; no edition change unless content changes per U.EditionSeries. Ref discipline. All bindings to packs/methods/distances use ...Ref.edition (dot on the Ref).

Define a portable minimal set of CHR slots. Each slot is CHR-typed (Characteristic, Scale/Unit/Polarity per A.17–A.18), Context-local, and guarded by USM (Claim scope G), freshness windows, and evidence lanes (TA/VA/LA). Contexts may extend the set; conforming extensions do not alter scale types illegally.

“Health” is a vector of CHR‑typed coordinates; no single scalar is implied. Lawful scalarization lives in Acceptance (G.4) under an explicit CG‑Spec ScaleComplianceProfile (SCP) and Γ‑fold rules, and is never embedded in CHR.

#Core Characteristics (kernel-portable names)

1. ReproducibilityRate (ratio ∈ [0,1]; polarity ↑; ReferencePlane=episteme; CG‑Spec‑bound) Fraction of tested claims/benchmarks that independent teams replicate under a declared ContextSlice and Γ_time window. Lane tags: LA (validation) with TA (typing) for protocols.

2. StandardisationLevel (ordinal; polarity ↑; ReferencePlane=episteme) {none, emerging, de facto, de jure}. No mean. Use medoid/mode; legal comparisons are ≤/=/> only. Tracks convergence on vocabularies, interfaces, or procedures.

3. AlignmentDensity (ratio; polarity ↑; ReferencePlane=episteme; CG‑Spec‑bound) Density of Substitution Bridges (same senseFamily, CL≥2) between major U.Traditions per 100 DHC‑SenseCells (G.2 F‑hooks) in the SoTA palette. Substitution rule: free substitution permitted at CL=3; at CL=2 substitute only with extra‑guard (count in reporting, but this is not «free substitution») Units: bridges_per_100_cells. Cross‑Context use requires Bridge+CL; penalties → R_eff only.

4. DisruptionBalance (interval; polarity = target band; ReferencePlane=episteme; CG‑Spec‑bound) Relative share of disruptive vs consolidating works within Γ_time using a registered CD‑index class (editioned; cite method id in UTS). Default plane: episteme. Publish the target band via Acceptance (G.4); not in CHR.

5. EvidenceGranularity (Context-declared: ordinal|ratio; polarity ↑; ReferencePlane=episteme) If ratio: units = claims_per_artifact or anchors_per_claim (declare). If ordinal: publish level names and ORD_COMPARE_ONLY. Fineness of evidential units and declared envelopes (experiment cards, benchmark tasks, audit granules). Encourages smaller, well-scoped claims over monoliths.

6. MetaDiversity (portfolio dispersion; polarity ↑ to band; ReferencePlane=episteme; CG‑Spec‑bound) Use entropy/HHI over MethodFamily/Tradition shares (method edition id in UTS); publish guard‑band as Acceptance binding; cross‑ordinal scalarisation is forbidden. Entropy/Herfindahl-type dispersion across U.Traditions, method families, or data regimes, bounded by a Context-declared guard-band (too low ⇒ monoculture; too high ⇒ incoherence).

Typing & legality. Each slot declares Scale/Unit/Polarity; illegal ops (e.g., mean on ordinals; unit mixing) are fail-fast per A.18/MM-CHR.

#Engineering-grade and semio-substitution extension slots

Contexts MAY add these DHC slots when the discipline-health question includes engineering-grade reasoning, architecturing, optimization, prediction, comparison, assurance input, decision input, first-principles justification, mathematical-lens use, or source-publication overread. These slots remain discipline-health characteristics. They do not become evidence relations, assurance relations, gate decisions, mathematical-lens use, measurement legality, release permission, or project authority.

7. EngineeringClaimJustificationRecoverability (ordinal; polarity ↑; ReferencePlane=episteme|world by declared claim; CG-Spec-bound when aggregated) Degree to which engineering-grade claims in the discipline or Context expose the exact justification that carries their force for the declared use. The justification may be a named construction, source, model, lens, evidence relation, characteristic relation, assurance relation, gate relation, method relation, or heuristic triage boundary, but it must cite the neighboring pattern governing the claiming FPF pattern when that force is live (A.10, B.3, A.15, A.20, A.21, C.16, C.29, or another governing pattern). Heuristic examples may carry recognition and triage only; prediction, comparison, optimization, falsification, assurance-input, decision-input, or architecture-readiness force requires the recoverable justification.

8. SemioSubstitutionPressure (ordinal or ratio; polarity ↓ to band; ReferencePlane=episteme; CG-Spec-bound when aggregated) Degree to which discipline texts, patterns, dashboards, views, publications, source chains, or review artifacts substitute wording, publication form, record appearance, source appearance, or explanation fluency for the operative engineering entity, relation, work, evidence, assurance, gate, decision, method, or mathematical-lens claim. Lower pressure is healthier when the discipline keeps EntityOfConcern, episteme, publication, source, carrier, and project-side claim kind or admissible-use boundary separable and names the pattern governing the recovered claim for any claim being made or admissible-use boundary.

Extension guard. Activating either extension slot requires a local EngineeringClaimJustification note or semio-substitution note that names the claim kind being made or admissible-use boundary, neighboring pattern governing the claiming FPF pattern, admissible use, non-admissible overread, and stop or reopen condition. The note is a DHC value explanation, not a new evidence source, assurance case, gate, release record, or work authority.

#Guard Macros (normative)

• ORD_COMPARE_ONLY(x) — for StandardisationLevel (ordinal).
• UNIT_CHECK(x) — forbid cross-unit aggregation (AlignmentDensity, ReproducibilityRate).
• POLARITY_CHECK(x) — enforce declared polarity (↑/↓/target-band) per MM‑CHR.
• FRESHNESS(x; window) — ensure values come from evidence within declared Γ_time; record valid_until; stale ⇒ {degrade|abstain} at Acceptance.
• PLANE_NOTE(x) — record ReferencePlane; compute CL^plane on crossings; penalties → R_eff only.
• LANE_TAGS(x; {TA|VA|LA}) — annotate contribution lanes.
• SCOPE_COVERS(x; TargetSlice) — enforce USM coverage of the computation.
• BRIDGE_CL(x; id, CL≥k) — on cross‑Context roll‑ups, require Bridge with CL; penalties route to R only. If planes differ, apply CL^plane and cite Φ_plane policy id. Hint: for AlignmentDensity reporting, set k=2 (CL≥2); CL=3 counts as free substitution.

#Legality Matrix (extract)

Note: For MetaDiversity/EvidenceGranularity (ordinal) use median/mode; forbid affine ops; unit mix always fails.

#Interfaces & Data Paths

• Inputs. U.Discipline from C.20 (composition), SoTA Palette/BridgeMatrix from G.2 (DHC‑SenseCells included), EvidenceProfiles from G.4/G.6.
• Outputs. Per‑Context DHC rows (these six slots), UTS Name Cards with twin labels (E.5/F.17–F.18), Registry/RSCR hooks on method edition changes; feeds G.12 (time‑series).
• Cross-Context reuse. Only via F.9 Bridges with CL and loss notes; Φ(CL) penalties applied to R (never F/G).

#Archetypal Grounding (three fields)

#Computer Vision (Benchmarks 2015→)

• ReproducibilityRate. Ratio of independently reproduced results on ImageNet-style tasks within rolling 24 mo (LA lane).
• StandardisationLevel. de facto for dataset specs and metrics in Vision_2024; emerging for robustness protocols.
• DisruptionBalance. Use an editioned CD‑index class (e.g., Wu‑style disruption family) with method id; publish target band via Acceptance; annotate ReferencePlane=episteme.
• AlignmentDensity. Bridges with CL≥2 across sub-traditions (supervised vs self-supervised).
• MetaDiversity. Entropy across method families (CNN/ViT/Hybrid) kept within guard-band to avoid monoculture.

#Biomedicine (Gene–Disease Associations)

• ReproducibilityRate. Fraction of associations replicated in independent cohorts within Γ_time(36 mo); LA lane with TA (typing of protocols).
• StandardisationLevel. de jure for certain reporting guidelines; emerging for pre-registration norms.
• EvidenceGranularity. Move from “paper-level” to claim-level units (Context raises the score).
• DisruptionBalance. Target band discourages sustained “novelty spikes” unbacked by replication.

#Software Performance Engineering (SPE)

• StandardisationLevel. emerging → de facto for SLO taxonomies and trace schemas across vendors.
• AlignmentDensity. CL-rated Bridges between tracing ecosystems.
• ReproducibilityRate. Share of publicly replicable perf claims in rolling windows.
• MetaDiversity. Balance across load models, failure modes, and toolchains.

#Decision‑Making (2015→)

• ReproducibilityRate — share of causal effect estimates replicated across independent datasets within Γ_time; LA lane. • StandardisationLevel — emerging for identification checklists; de facto for SCM notation in leading stacks (ordinal; no means). • AlignmentDensity — CL‑rated Bridges between SCM/DoWhy‑style and RL/BO traditions per 100 DHC‑SenseCells. • MetaDiversity — dispersion across method families (SCM/RL/BO/DT) within guard‑band; entropy/HHI (units declared in CG‑Spec).

#Evolutionary Architecture (software)

• ReproducibilityRate — fraction of architecture fitness results reproduced on independent workloads (rolling 18–24 mo; LA lane). • StandardisationLevel — de facto for ADR/ATAM patterns; emerging for continuous fitness protocols. • AlignmentDensity — Bridges across ATAM/SAAM/ADR style guides (CL≥2) normalised per 100 DHC‑SenseCells. • MetaDiversity — portfolio dispersion across patterns (microservices, event‑driven, layered) with guard‑bands; no ordinal arithmetic.

#Characteristic reading and publication path

1. Declare Context & TargetSlice. (USM) Name editions, Standards, env params, Γ_time.
2. Collect evidence. Bind sources via G.6 EvidenceGraph; tag lanes and freshness.
3. Compute DHC slots. Enforce Legality Matrix and Guard Macros.
4. Bridge (if needed). Map via F.9; attach CL and loss notes; apply R penalties.
5. Publish to UTS. Name Cards (Tech/Plain), twin labels; bind DHCMethodSpecRef.edition, DistanceDefRef.edition, and, where templates are used, DHCMethodRef.edition; register RSCR triggers (method change, ScoringMethod/NormalizationMethod edits).
6. Publication view. Feed G.12 with time-series and guard-bands (disruption, diversity) when a dashboard or trend publication is live.

#Bias-Annotation (E-cluster lenses)

• Didactic. Plain names + twin labels; one-screen tables for managers.
• Architectural. No ordinals averaged; all cross-Context movement goes through Bridges+CL; penalties never touch F/G.
• Pragmatic. Freshness-aware; unknowns tri-state; values are decision-input cues, not trophies.
• Epistemic. Evidence lanes explicit; reproducibility is LA, typing is TA; validation distinct from dashboard or report publication.

#Conformance Checklist

This checklist verifies a DHC reading after the practitioner has selected the live discipline-health question. It is not an audit form and not a dashboard specification.

#Consequences

Benefits. Scale-legal comparisons; freshness-aware governance; explicit cross-tradition alignment; dashboard views that do not lie by averaging ranks. Costs. Some ceremony (scales, windows, lanes, bridges), offset by template macros and UTS automation. Risks avoided. “Phlogiston disciplines” (charisma-driven fields) surface as unhealthy in DHC readings; No-Free-Lunch preserved by G.5 (selector returns sets, not universal scalars).

#SoTA-Echoing and rationale

The practical consequence is that C.21 reads discipline health through typed characteristics. It can feed dashboards or time-series publications, but the dashboard is only a publication view over DHC readings; it is not the discipline-health ontology and not project authority.

#Relations

• Builds on: A.17–A.18 (Characteristic/CSLC), A.2.6 (USM scopes), B.3 (assurance lanes), C.16 (MM-CHR templates).
• Coordinates with: C.20 (what a U.Discipline is), G.2 (SoTA palette and BridgeMatrix), G.12 (Dashboard operationalization), G.9 (parity harness for fair comparisons).
• Constrains: G.10 (pack ships DHC rows + method ids), G.11 (refresh windows/decay), G.5 (selector may reference DHC only via admissible predicates; no cross‑ordinal scalarisation). Coordinates: F.9 (Bridges for cross‑Tradition comparisons).

#Annex — Author’s quick template (copy-paste)

C.21.DHC(Context: <name/edition>; TargetSlice: <tuple>; Γ_time: <policy>)
  ReproducibilityRate:
    value: <0..1>   lane: LA   window: <…>   scope: <…>
  StandardisationLevel:
    value: {none|emerging|de_facto|de_jure}   compare_only: true
  AlignmentDensity:
    value: <ratio>   units: bridges_per_100_DHC_SenseCells   CL_min: 2   scope: <…>
  DisruptionBalance:
    value: <−1..1>   method: <CD-index class / edition>   target_band: [l,u]
  EvidenceGranularity:
    value: <ordinal|ratio per Context>   notes: <…>
  MetaDiversity:
    value: <entropy/HHI>   target_band: [l,u]
Guards: ORD_COMPARE_ONLY(StandardisationLevel), UNIT_CHECK(*), FRESHNESS(*), LANE_TAGS, SCOPE_COVERS, BRIDGE_CL(if x-Context)
Publish: UTS twin labels; RSCR triggers on method edition change.

#C.21:End

Last Updated: 2026-06-03 — this section last modified in upstream FPF commit a0c90e3b (github.com/ailev/FPF)