Break
with intent.

RuleBreakTargetEvidence

RuleBreakTargetEvidence {
  targetEvidenceRef,

  targetFamily,
  targetRef,
  targetOwnerRef,
  targetModuleRef,
  targetSnapshotRef?,

  subjectRefs[],
  sourceEvidenceRefs[],

  canvasRef,
  coordinateFrame,
  sourceUnit,

  confidence,
  status,

  mutationAllowed: false
}

RuleBreakIntentSource {
  intentSourceRef,
  intentSourceOwnerRef,
  intentSourceKind,

  evidenceRefs[],
  confidence,
  provenance,

  mutationAllowed: false
}

RuleBreakIntent

RuleBreakIntent {
  intentRef,

  targetEvidenceRef,
  subjectRefs[],

  intentMode,
  intentSourceRef,
  intentSourceOwnerRef,
  intentSourceKind,

  semanticIntentRefs[]?,

  confidence,
  status,
  provenance,

  mutationAllowed: false
}

RuleBreakCorrectionRequest {
  requestRef,
  targetOwnerRef,
  reasonCode,
  evidenceRefs[],
  suggestedAction,
  status,
  mutationAllowed: false
}

{
  "module": "Compo 12.00",
  "name": "grid_break_taxonomy_and_intent",
  "version": "1.0-canonical",
  "canonical_main_title": "Grid-break taxonomy and intent",
  "topic": "12_НАРУШЕНИЕ_СЕТКИ_И_КОНТРОЛИРУЕМОЕ_НАПРЯЖЕНИЕ",
  "scope": "RULE_BREAK_ROOT_TAXONOMY_EXPLICIT_INTENT_TARGET_EVIDENCE_IDENTITY_AND_TOPIC12_OWNERSHIP",
  "main_topic_contract": {
    "topic_title": "Нарушение сетки и контролируемое напряжение",
    "topic_goal": "formalize intentional breaking of grid and alignment: deviation magnitude, protected relations, exception budget and justification so controlled tension is distinguishable from error",
    "expected_output": "RuleBreakIntent, deviation metrics, exception budget, evidence",
    "planned_files": [
      "12.00 Grid-break taxonomy and intent",
      "12.01 Deviation magnitude",
      "12.02 Protected relations and exception budgets",
      "12.03 Rule-break candidates and justification",
      "12.04 Override scoring and tests"
    ]
  },
  "review_basis": {
    "previous_topic": "Topic 11 field landmarks and local landmark fit",
    "related_owners_reviewed": [
      "Compo8.00 ColumnGrid root/domain/topology ownership",
      "Compo8.03 ColumnGrid placement candidates",
      "Compo8.04 explicit-target local ColumnGrid evaluation",
      "Compo9.00–9.04 ModularGrid geometry/occupancy/collision/variant ownership",
      "Compo10.00 AlignmentRelation taxonomy and owner-qualified anchor evidence",
      "Compo10.01–10.04 concrete alignment relations, tolerances and local evaluation",
      "Compo11.00–11.04 FieldLandmark family/distance/applicability/candidate/local-fit ownership"
    ],
    "cross_cutting_instruction": "COMPO project instruction v2.0 IDEAS/LAB/DATA-aware"
  },
  "consumes": [
    "owner-qualified upstream grid/alignment rule or relation identity",
    "canonical subject refs affected by that upstream rule/relation",
    "explicit break-intent source evidence",
    "optional opaque semanticIntentRefs[] as deferred semantic evidence",
    "optional human feedback reason codes"
  ],
  "provides": [
    "RuleBreakTargetEvidence",
    "RuleBreakIntentSource",
    "RuleBreakIntent",
    "RuleBreakIntentResolution",
    "RuleBreakCorrectionRequest",
    "Topic12OwnershipMap"
  ],
  "owns_metrics": [],
  "descriptors_only": [
    "ruleBreak.target.targetFamily",
    "ruleBreak.target.targetRef",
    "ruleBreak.target.targetOwnerRef",
    "ruleBreak.intent.intentMode",
    "ruleBreak.intent.intentSourceKind",
    "ruleBreak.intent.subjectCount",
    "ruleBreak.intent.status",
    "ruleBreak.intent.confidence"
  ],
  "target_family_taxonomy": {
    "COLUMN_GRID": "owner-qualified Topic8 ColumnGrid rule/constraint evidence",
    "MODULAR_GRID": "owner-qualified Topic9 ModularGrid rule/constraint evidence",
    "PAIRWISE_ALIGNMENT": "owner-qualified Topic10 pairwise geometric alignment relation",
    "TEXT_BASELINE_ALIGNMENT": "owner-qualified Topic10 baseline relation",
    "SHARED_AXIS_ALIGNMENT": "owner-qualified Topic10 multi-object shared-axis relation"
  },
  "target_family_policy": {
    "closed_enum_v1": "Only the five target families above are canonical in 12.00 v1.0.",
    "field_landmark": "Topic11 FieldLandmark is not automatically a Topic12 rule-break target.",
    "other": "Unsupported target families require a reviewed extension; do not silently map them to OTHER."
  },
  "rule_break_target_evidence_schema": {
    "targetEvidenceRef": "stable identity derived from targetFamily + targetRef + targetOwnerRef + targetSnapshotRef + canonical subjects",
    "targetFamily": "COLUMN_GRID|MODULAR_GRID|PAIRWISE_ALIGNMENT|TEXT_BASELINE_ALIGNMENT|SHARED_AXIS_ALIGNMENT",
    "targetRef": "exact upstream rule/relation/constraint identity",
    "targetOwnerRef": "canonical upstream owner",
    "targetModuleRef": "exact upstream module/service ref",
    "targetSnapshotRef": "optional exact versioned evidence snapshot",
    "subjectRefs": "one or more distinct canonical subject refs; set-like lexical canonicalization",
    "sourceEvidenceRefs": "one or more owner-qualified evidence refs supporting target existence",
    "canvasRef": "required for geometric target evidence",
    "coordinateFrame": "required for geometric target evidence",
    "sourceUnit": "required for geometric target evidence",
    "confidence": "0..1 evidence confidence only",
    "status": "OK|N_A|INVALID",
    "mutationAllowed": false
  },
  "intent_source_schema": {
    "intentSourceRef": "required explicit source identity",
    "intentSourceOwnerRef": "required owner of source declaration/policy",
    "intentSourceKind": "HUMAN_DIRECTIVE|STYLE_PRESET|SEMANTIC_POLICY_REF|EXTERNAL_POLICY_REF",
    "evidenceRefs": "one or more refs supporting the declaration",
    "confidence": "0..1 evidence confidence only",
    "provenance": "DERIVED|STANDARD|EMPIRICAL|INITIAL_HEURISTIC|IMPLEMENTATION_HEURISTIC|STYLE_PRESET|DEFERRED",
    "forbidden_sources": [
      "AUTO_INFERRED_FROM_DEVIATION",
      "GENERATOR_SELF_AUTHORIZATION",
      "MODEL_GUESS_WITHOUT_EXPLICIT_REF"
    ],
    "mutationAllowed": false
  },
  "intent_mode_taxonomy": {
    "ALLOW_EXCEPTION": "A break may be accepted later if all downstream protection/budget/justification checks allow it; it is not preferred and need not be generated.",
    "EXPLORE_BREAK": "Downstream candidate generation may explicitly explore controlled-break candidates; no candidate is required to survive.",
    "REQUEST_BREAK": "An explicit design directive requests a break candidate/solution, but does not outrank protected relations, hard constraints or future conflict-priority policy."
  },
  "rule_break_intent_schema": {
    "intentRef": "stable identity from targetEvidenceRef + sorted subjectRefs + intentMode + intentSourceRef + intentSourceOwnerRef",
    "targetEvidenceRef": "exact RuleBreakTargetEvidence ref",
    "subjectRefs": "one or more distinct subject refs; must be a subset of target evidence subjects",
    "intentMode": "ALLOW_EXCEPTION|EXPLORE_BREAK|REQUEST_BREAK",
    "intentSourceRef": "exact explicit source ref",
    "intentSourceOwnerRef": "exact source owner",
    "intentSourceKind": "copied from validated RuleBreakIntentSource for auditability",
    "semanticIntentRefs": "optional opaque deferred semantic refs; no interpretation in 12.00",
    "confidence": "minimum target/source evidence confidence as serialization policy only; not tension strength",
    "status": "OK|INVALID",
    "provenance": "canonical provenance class",
    "mutationAllowed": false
  },
  "intent_resolution_schema": {
    "status": "OK|N_A|INVALID",
    "intentRef": "present only when valid explicit intent exists",
    "reasonCode": "EXPLICIT_INTENT|NO_EXPLICIT_INTENT|INVALID_TARGET|INVALID_INTENT_SOURCE|SUBJECT_SCOPE_MISMATCH|UNSUPPORTED_TARGET_FAMILY|...",
    "controlledTensionEligible": "true only for status OK; eligibility is not final acceptance or score",
    "important": "Observed deviation alone resolves to N_A / NO_EXPLICIT_INTENT, never to OK."
  },
  "identity_policy": {
    "targetEvidenceRef": "hash(targetFamily,targetRef,targetOwnerRef,targetModuleRef,targetSnapshotRef,sorted unique subjectRefs,canvasRef,coordinateFrame,sourceUnit)",
    "intentRef": "hash(targetEvidenceRef,sorted unique subjectRefs,intentMode,intentSourceRef,intentSourceOwnerRef)",
    "set_semantics": "subjectRefs and evidenceRefs are set-like and lexically canonicalized",
    "excluded_from_identity": [
      "display label",
      "display color",
      "free-form human comment",
      "human feedback text",
      "confidence"
    ]
  },
  "status_policy": {
    "OK": "explicit supported target + explicit valid intent source + compatible subject scope",
    "N_A": "no explicit break intent declaration exists for the target/scope; controlled-tension interpretation is unavailable",
    "INVALID": "malformed/unsupported target, malformed source, subject mismatch or forbidden self-authorizing inference",
    "N_A_not_zero": true,
    "observed_deviation_without_intent": "N_A, not controlled tension"
  },
  "topic12_ownership_map": {
    "12.00": "root rule-break taxonomy, explicit intent schema, target evidence contract, identity, accidental-vs-intentional firewall",
    "12.01": "deviation magnitude descriptors/metrics and their units/normalization",
    "12.02": "protected relations, protection policy and exception budget accounting",
    "12.03": "bounded rule-break candidate generation and explicit justification evidence",
    "12.04": "override/local controlled-break evaluation, scoring policy and consolidated Topic12 tests"
  },
  "integration_firewalls": {
    "Topic8": "owns ColumnGrid geometry/topology/placement candidates/local grid evaluation; 12.00 references target evidence only",
    "Topic9": "owns ModularGrid cells/regions/occupancy/collision/reflow; 12.00 does not reinterpret collisions as intentional tension",
    "Topic10": "owns alignment relation identity/measurement/tolerance/local fit; 12.00 references an existing relation as a possible break target",
    "Topic11": "FieldLandmark avoidance/non-use is not automatically a grid/alignment rule break and is outside the canonical target enum in 12.00 v1.0",
    "12.01": "owns deviation magnitude; 12.00 owns no deviation metric",
    "12.02": "owns protected relations and exception budgets; 12.00 cannot spend or reserve budget",
    "12.03": "owns rule-break candidates and justification evaluation; 12.00 creates no candidates and does not judge rationale sufficiency",
    "12.04": "owns override scoring/tests; 12.00 owns no numeric score",
    "future_semantics": "semanticIntentRefs remain opaque until their canonical semantic owner evaluates them",
    "future_priority": "module execution order and REQUEST_BREAK do not define global rule priority"
  },
  "metric_policy": {
    "direct_scoring": "FORBIDDEN in 12.00",
    "deviation_magnitude": "DEFERRED to 12.01",
    "exception_budget": "DEFERRED to 12.02",
    "justification_score": "DEFERRED / no universal scalar in 12.00",
    "override_score": "DEFERRED to 12.04",
    "intent_confidence": "evidence confidence only; not tension quality",
    "beautyScore": "FORBIDDEN"
  },
  "intent_firewalls": [
    "observed deviation != explicit break intent",
    "constraint violation != controlled tension",
    "large deviation != stronger intention",
    "intent confidence != tension strength",
    "ALLOW_EXCEPTION != preference",
    "EXPLORE_BREAK != acceptance",
    "REQUEST_BREAK != global priority",
    "REQUEST_BREAK != permission to violate hard constraints",
    "target existence != permission to break it",
    "protected relation != ordinary target; protection semantics are owned by 12.02",
    "candidate != intent; candidate generation is owned by 12.03",
    "justification text != proof; justification evidence/evaluation is owned by 12.03",
    "rule break != foreign mutation",
    "N_A != 0",
    "no beautyScore"
  ],
  "hard_constraints": [
    "12.00 owns only the root rule-break taxonomy, RuleBreakTargetEvidence, RuleBreakIntentSource, RuleBreakIntent, intent resolution semantics, identity and Topic12 ownership map.",
    "12.00 owns no numeric metric.",
    "12.00 does not measure deviation magnitude; Compo12.01 owns deviation descriptors/metrics.",
    "12.00 does not define protected-relation evaluation or exception budget accounting; Compo12.02 owns them.",
    "12.00 does not generate rule-break placement candidates and does not evaluate justification sufficiency; Compo12.03 owns them.",
    "12.00 does not score or accept/reject final overrides; Compo12.04 owns local controlled-break evaluation/scoring and consolidated tests.",
    "A rule break is eligible to be interpreted as controlled tension only when explicit intent evidence exists.",
    "Observed geometric deviation alone never creates RuleBreakIntent.",
    "An upstream grid/alignment violation cannot self-authorize its own exception.",
    "Generator output or model inference without an explicit owner-qualified intent source is insufficient to create valid RuleBreakIntent.",
    "Target evidence must reference an upstream canonical owner; 12.00 does not copy or re-own grid/alignment geometry.",
    "Target family is a closed enum in v1.0: COLUMN_GRID, MODULAR_GRID, PAIRWISE_ALIGNMENT, TEXT_BASELINE_ALIGNMENT, SHARED_AXIS_ALIGNMENT.",
    "Unsupported target family is INVALID until a reviewed canonical extension exists; do not map silently to OTHER.",
    "FieldLandmark is not automatically a Topic12 target and landmark avoidance is not automatically controlled tension.",
    "RuleBreakTargetEvidence subjectRefs are set-like, unique and canonicalized lexically.",
    "RuleBreakIntent subjectRefs must be a non-empty subset of RuleBreakTargetEvidence subjectRefs.",
    "RuleBreakIntent references exactly one target evidence object in v1.0; multiple targets require multiple intents and later conflict/budget handling.",
    "Intent mode is categorical: ALLOW_EXCEPTION, EXPLORE_BREAK or REQUEST_BREAK.",
    "ALLOW_EXCEPTION permits later acceptance but does not express positive preference.",
    "EXPLORE_BREAK permits downstream candidate exploration but does not imply acceptance.",
    "REQUEST_BREAK records an explicit design request but does not define global priority and cannot bypass protected/hard relations.",
    "Intent source must be explicit and owner-qualified.",
    "Allowed intent source kinds are HUMAN_DIRECTIVE, STYLE_PRESET, SEMANTIC_POLICY_REF and EXTERNAL_POLICY_REF.",
    "AUTO_INFERRED_FROM_DEVIATION is forbidden as an intent source.",
    "GENERATOR_SELF_AUTHORIZATION is forbidden as an intent source.",
    "MODEL_GUESS_WITHOUT_EXPLICIT_REF is forbidden as an intent source.",
    "semanticIntentRefs are opaque deferred evidence; 12.00 does not interpret semantic meaning.",
    "Intent confidence is evidence confidence only, not quality, tension strength or score.",
    "Missing explicit intent resolves controlled-tension eligibility to N_A, not score 0 and not INVALID by default.",
    "Malformed target/source/scope is INVALID, not N_A.",
    "N/A != 0.",
    "RuleBreakIntent identity must be deterministic and insensitive to subject/evidence input order.",
    "Display metadata and free-form comments do not change canonical intent identity.",
    "12.00 mutates no grid, object, text, alignment relation, cell, region, landmark or canvas state.",
    "Any correction is an owner-qualified request with mutationAllowed=false.",
    "Module execution order is not rule priority.",
    "REQUEST_BREAK is not a priority override and cannot resolve conflicts by itself.",
    "A future conflict-priority owner must resolve collisions between intent, protected relations and other hard constraints.",
    "IDEA != CANON.",
    "EXPERIMENT != CANON.",
    "Human feedback does not auto-mutate CANON.",
    "No universal tensionScore is defined in 12.00.",
    "No beautyScore."
  ],
  "correction_request_schema": {
    "requestRef": "stable request identity",
    "targetOwnerRef": "owner that must review/correct the target/source evidence",
    "reasonCode": "TARGET_EVIDENCE_WRONG|INTENT_SOURCE_WRONG|SUBJECT_SCOPE_WRONG|INTENT_MODE_WRONG|RULE_BREAK_WAS_ACCIDENTAL|OTHER_REVIEW_REQUIRED",
    "evidenceRefs": "canonical set-like refs",
    "suggestedAction": "non-binding text",
    "status": "OPEN|INVALID",
    "mutationAllowed": false
  },
  "IDEAS_REFS": [
    "IDEA-01 full human test after all 72 topics; Topic12 captures perception hooks now",
    "IDEA-02 semantics required before placement; semanticIntentRefs remain opaque/deferred",
    "IDEA-03 explicit rule priorities; REQUEST_BREAK must not become hidden priority",
    "IDEA-04 collect structured data for future composition model",
    "IDEA-05 preserve reasons for human corrections",
    "IDEA-06 structured machine-readable layer",
    "IDEA-07 graphics-for-post-text research remains downstream/separate"
  ],
  "DATA_CAPTURE": [
    "sceneRef",
    "targetFamily",
    "targetRef",
    "targetOwnerRef",
    "targetSnapshotRef",
    "subjectRefs[]",
    "intentMode",
    "intentSourceKind",
    "intentSourceRef",
    "intentSourceOwnerRef",
    "semanticIntentRefs[]",
    "intentResolutionStatus",
    "intentResolutionReasonCode",
    "humanReasonCodes[]",
    "future deviationEvidenceRef",
    "future budgetDecisionRef",
    "future justificationEvidenceRef",
    "future overrideEvaluationRef"
  ],
  "DEFERRED_SEMANTIC_INPUTS": [
    "semanticIntentRefs[]",
    "semantic role interpretation",
    "meaning/hierarchy rationale",
    "brand/editorial tone rationale"
  ],
  "HUMAN_FEEDBACK_HOOKS": [
    "RULE_BREAK_WAS_ACCIDENTAL",
    "RULE_BREAK_INTENT_WRONG",
    "TARGET_RELATION_WRONG",
    "INTENT_MODE_TOO_STRONG",
    "INTENT_MODE_TOO_WEAK",
    "SHOULD_PRESERVE_RELATION",
    "TENSION_READS_AS_ERROR",
    "TENSION_TOO_WEAK_TO_NOTICE"
  ],
  "CONFLICT_PRIORITY_ASSUMPTIONS": [
    "module execution order is not priority",
    "REQUEST_BREAK does not outrank hard/protected relations",
    "12.00 records intent but does not resolve cross-owner conflicts",
    "future reviewed priority/conflict owner remains authoritative"
  ],
  "LAB_HOOKS": [
    "A/B: same deviation with explicit intent cues vs accidental deviation",
    "human perception: when controlled break reads as tension vs error",
    "compare ALLOW_EXCEPTION vs EXPLORE_BREAK vs REQUEST_BREAK workflows",
    "measure disagreement on target family and subject scope",
    "collect reason codes for rejected rule-break interpretations"
  ],
  "candidate_generation": {
    "rule_break_candidates": "NONE in 12.00",
    "reason": "candidate generation belongs to 12.03"
  },
  "foreign_mutation_policy": {
    "foreign_mutation": "FORBIDDEN",
    "correction_request": "owner-qualified only",
    "target_state_change": "not performed by 12.00"
  },
  "stop_condition": "After 12.00 validation, STOP. Do not create 12.01 automatically."
}