Skip to content

History / 09 Architecture Decisions

Revisions

  • ADR-061 (Proposed): operational κ-label composition surface over the Atlas's categorical operations Prism-runtime-level decision building on ADR-058 + ADR-059; no foundation or SDK change. Operationalizes ADR-059's five categorical operations on the Atlas image inside E₈ — product (G₂), quotient (F₄), filtration (E₆), augmentation (E₇), direct embedding (E₈) — as five ConstrainedTypeShapes in prism's standard type library (G2ProductShape<N> binary; the other four unary). A composition of N component κ-labels under an operation is a typed-input value the existing ψ-pipeline consumes through the existing σ-projection; the realization's canonicalize verb emits a canonical form whose κ-label is the composed κ-label — itself a κ-label, recursively composable (closed under the Atlas). Arity ≤ 3 / depth ≤ 8 per the Categorical X regime split (ADR-025) and ADR-059's H-/O-level bounds; wide compositions decompose via ConstraintRef::Recurse (ADR-057). σ-axis substitutability (ADR-047) and cross-substrate universality (ADR-058 + ADR-059) lift to compositions. Byte-concatenation / tree-hash composition rejected (collapses ADR-059's algebraic codomain to byte-pigeonhole). Substrate footprint: foundation unchanged (TypedCommitment per ADR-048 and ObservablePredicate per ADR-049 closed sets retained; no wire-format change), SDK unchanged, prism gains the five standard-type-library shapes; audit-trail row marked "No." Propagation: ADR-061 section + audit row + heading (ADR-038 through ADR-061) in ch 9 (lands on the Continued page, now ~322 KiB); composition shapes added to the ch 5 standard-type-library entry; ch 8 ADR-058/059 cluster extended with the composition-surface reading; ch 4 strategic row cross-ref; a "κ-label composition shapes" glossary entry. Applied domain-neutral per the wiki's discipline: the downstream content- addressing realization crate, its Lean module paths, and its realization roster are generalized to "a content-addressing realization (Layer-3 per ADR-031)"; format names (PNG/PDF/GGUF/ONNX/JSON) kept as illustrative domains only. Audit resolved, while applying: the source ADR's principle numbering gap ((1)(2)(4)(5)(6) → renumbered (1)–(6)); a dangling prism::std_types::RouteShape reference (generalized to "route-declaration shapes," no such type in the wiki); and grounded "Categorical X" in the existing ADR-025 regime split (T=3 H-level, O=8 O-level) rather than as a free-floating theorem name. All eight validators (V1–V8) pass; page-size guard OK (glossary at 494 KiB — warning band, under the 512 KiB limit). Status: Proposed (as authored). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 23, 2026
  • Split Architecture Decisions page at ADR-038 (GitHub 512 KiB render limit) The Architecture Decisions chapter grew to ~695 KiB, exceeding GitHub's ~512 KiB per-page rich-markdown rendering limit: the live wiki truncated mid-ADR-46 (ADR-047 began at byte ~512,380, right at the boundary), so ADR-047 through ADR-060 did not render. Fix: stage-to-wiki-root.sh now splits the staged page at the first `## ADR-038 ` heading — the start of the "substrate-amendment audit trail (ADR-038 through ADR-NNN)" cohort. Page 1 (09-Architecture-Decisions.md, ~396 KiB) keeps the chapter intro, the audit-trail table, and ADR-001..037; page 2 (09-Architecture-Decisions-Continued.md, ~300 KiB) carries ADR-038 onward. Both pages are well under 512 KiB, so every ADR renders. New ADRs append to page 2; ADR-001..037 is stable, so the split point and page-1 size do not drift. The split is deterministic (first ADR-038 heading line), so Stage 5 idempotence holds; it runs after image-reference rewriting so page 2 inherits the already-rewritten tail. Cross-navigation links added in both directions; the Sidebar gains a nested "ADR-038 onward" entry under chapter 9. The source stays one .adoc (09_architecture_decisions.adoc), so V1 arc42 structural alignment is unaffected. All eight validators (V1-V8) pass; idempotence check clean. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 22, 2026
  • Post-ADR-060 consistency sweep: residual byte-width-cap references Closes the current-fact references to removed constants an audit surfaced after landing ADR-060: - 12 glossary `Grounded::output_bytes` entry asserted the value-payload slot is bounded by the removed `ROUTE_OUTPUT_BUFFER_BYTES`; rewritten to the source-polymorphic TermValue carrier (Inline/Borrowed/Stream, no fixed slot). - 09 substrate-amendment audit-trail heading "(ADR-038 through ADR-055)" was stale (table runs through ADR-060); corrected. - 09 ADR-013 substrate-amendment narrative (line 186) and ADR-030 capacity-bound reconciliation asserted the writer-style resolver signature / AXIS_OUTPUT_BYTES cap as current fact; annotated with the ADR-060 amendment. - 09 ADR-050 arithmetic-width references cited the removed TERM_VALUE_MAX_BYTES as the effective width ceiling; reframed to the WittLevel ceiling (WITT_LEVEL_MAX_BITS / carrier_inline_bytes) per ADR-060. - 09 ADR-051 audit-trail row annotated with the source-polymorphic carrier amendment. Remaining byte-width-cap / DefaultHostBounds occurrences are confined to the historical bodies of ADR-018/028/036/037 (each carrying an ADR-060 supersession banner) and ADR-060's own diagnosis/rejected-alternatives — preserved per the historical-record convention. No current-architecture chapter (2/4/5/6/8/10/11) or the OPM Conceptual Model references a removed constant as current fact. All eight validators (V1-V8) pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 22, 2026
  • ADR-060: source-polymorphic value carrier — remove the contrived 4096 byte-width caps and DefaultHostBounds Lands the foundation 0.5.0 architectural change diagnosing the 4096-byte carrier ceiling as a stable-Rust workaround (a pinned literal absent generic_const_exprs) whose application-overridability was fictional (foundation internals read the width through the hardcoded DefaultHostBounds path). Implementer-feedback-loop: ADR-037's byte-width-cap family proved unrealizable as committed, so ADR-060 supersedes it. ADR-060 (new, ch 9) commits: - TermValue becomes source-polymorphic TermValue<'a, const INLINE_BYTES: usize> (Inline / Borrowed / Stream). Inline width foundation-derived via carrier_inline_bytes::<B>() = max(Witt-literal width, fingerprint width, kappa-label ASCII width). Borrowed/Stream carriers have no byte-width ceiling; Stream(&'a dyn ChunkSource) admits unbounded payloads (multi-GB tensor data, large attestation envelopes, production canonical-JSON). - 12 byte-width capacity constants removed from HostBounds (TERM_VALUE_MAX_BYTES, AXIS_OUTPUT_BYTES_MAX, ROUTE_INPUT/OUTPUT_BUFFER_BYTES, 8 psi-stage *_OUTPUT_BYTES_MAX); per-psi-stage carrier widths derive from foundation const fns. HostBounds drops 26 -> 14 associated constants. - DefaultHostBounds removed; every application declares its own impl HostBounds. - 8 psi-stage resolver traits return a TermValue (not writer-style &mut [u8]); streaming-sigma-projection is now a foundation-level surface. - Wire format unchanged (kappa-label/trace/certificate byte-identical); no shim. Reconciled three internal inconsistencies in the source document against the wiki's authoritative ADR-037 starting count (26 = 4 pre-ADR-018 + 22 migrated): the removed set is 12 (not the doc's "11"), the retained set is 14 (not "13"), and carrier_inline_bytes is a 3-term max including the kappa-label width (the doc's 2-term max self-contradicts its own validation point 9, which admits the kappa-label can exceed a fingerprint-sized buffer). Propagation: supersession/amendment banners on ADR-037/018/028/036/041/051 (historical bodies preserved per convention); ADR-060 audit-trail row + strategic -decision row (ch 4); updated HostBounds/resolver-surface/carrier descriptions in ch 2/5/6/8/10; TR-10 marked resolved, TR-01/TR-18 updated, new TR-23 (carrier lifetime migration); glossary HostBounds/TermValue/Term::Literal/resolver/removed -constant entries + new TermValue/ChunkSource/carrier_inline_bytes entries. Fixed an AsciiDoc rendering break: the `*_OUTPUT_BYTES_MAX` / `*_carrier_bytes` backtick globs opened stray emphasis crossing the code span (interleaved <literal>/<emphasis> -> invalid DocBook XML, Pandoc failure on ch 02/04/08); converted to ++...++ passthrough. Domain-neutral: use cases (GGUF/ONNX/JSON/SLSA) kept illustrative; the downstream content-addressing crate is referred to generically, not named. All eight validators (V1-V8) pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 22, 2026
  • ADR-058 + ADR-059: κ-derivation as compression operator + Atlas/E₈ codomain; document kernel::convergence substrate vocabulary the wiki had omitted Lands the architect-provided ADR-058 (κ-derivation IS compression-to-canonical- form; three-tier closure-lossless taxonomy T1/T2/T3; cross-substrate compression universality as a TC-05 closure property) and ADR-059 (Atlas-image-inside-E₈ as the operator-geometry codomain of κ-derivation; Hopf convergence tower as coarse stratification; exceptional groups G₂/F₄/E₆/E₇/E₈ as finer codomain typing). Both are conceptual-reading commitments over existing constructs — audit-trail rows marked "No" wire-format change. Consistency repair before applying: ADR-059's prose asserts kernel::convergence is "existing substrate vocabulary per the building-block view in section 5; per the kernel-component glossary in section 12." That vocabulary was only referenced generically in the wiki ("the convergence machinery" / "convergence algebra") with no trait surface. Verified against the real foundation source (foundation/src/kernel/convergence.rs and spec/src/namespaces/convergence.rs, UOR ontology Amendment 66) that the module and its trait surface (ConvergenceLevel / HopfFiber / ConvergenceResidual / CommutativeSubspace / AssociativeSubalgebra; four levels R/C/H/O at dims {1,2,4,8}; Hopf fibers S⁰/S¹/S³/S⁷; identities existence/feedback/choice/self-reference; persistent β_{2^k−1}=1) are genuine substrate vocabulary the wiki had under-documented. Enriched the section 5 kernel-component description and the section 12 kernel entry with the authoritative trait surface, and added a "Hopf convergence tower" glossary entry — making ADR-059's section-5/12 citations precisely accurate. 08 concepts: two paragraphs after the ADR-057 paragraph commit the compression-operator reading (ADR-058) and the operator-geometry codomain (ADR-059) into the conceptual model, tied to the Hopf convergence tower. Domain-neutrality preserved: use cases enumerated as illustrative (JSON/XML/ ASN.1/S-expressions/proof-normalization/term-rewriting); no downstream consumer crate named. Atlas/exceptional-group code realization explicitly left as application-level / demand-driven Layer-3 per ADR-031, not committed to uor-foundation. All eight validators (V1–V8) pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 21, 2026
  • ADR-057: bounded recursive structural typing via ConstraintRef::Recurse + foundation shape-IRI registry Adds the foundation-declared ConstraintRef::Recurse { shape_iri, descent_bound } variant (and the LeafConstraintRef::Recurse parallel) for bounded recursive structural typing at the constraint-geometry level, completing the parallel to Term::Recurse at the Term level. Foundation publishes the shape_iri_registry module + register_shape! SDK macro for link-time shape-IRI registration; ψ_1 NerveResolver unrolls Recurse references against the registry up to the descent budget; const-time validation defers Recurse to runtime; partition_coproduct! / partition_product! macros admit recurse(<bound>):<T> operand markers; shapes carrying Recurse constraints saturate CYCLE_SIZE at u64::MAX. Wire format gains a Recurse variant tag carrying (shape_iri, descent_bound); wire-format version bumps. Unblocks bounded recursive structural typing for JSON, XML, code-module ASTs, S-expressions, ASN.1 / Protocol-Buffer message families, filesystem-like hierarchies, and categorical constructions (rose trees, bounded forests, etc.) without per-application const-eval-cycle-break invention. Architecture remains domain-neutral: the use-case scope enumerates illustrative domains; no downstream application crate is named in the substrate description. ADR-013 audit-trail table: new ADR-057 row marked wire-format-amending. ADR-040 future-BoundShape paragraph: replace "hypothetical MultiPrecisionFieldLessEqBound" speculative naming with policy phrasing — any totally-ordered value type with comparison semantics distinct from those the catalog realizes lands as its own distinct BoundShape primitive at the ADR-013/TR-08 amendment boundary. ADR-037 / MAX_RESOLVER_TUPLE_ARITY (3 sites): replace "currently 8, growing with future ADR-013/TR-08 substrate amendments" with structurally-bound phrasing — 8 per ADR-035's eight resolver-bound ψ-stages, tracking any ADR-013/TR-08 amendment that revises card(ResolverCategory). Variadic-generics carve-out: explicitly language-conditioned (Rust's current absence of variadic generics); migration follows the ADR-013/TR-08 rule if variadic generics stabilize. Glossary: ConstraintRef entry adds Recurse variant + foundation shape-IRI registry resolution semantics; LexicographicLessEqThreshold entry's "256^|target|" math notation broke a table cell — rewritten as "256^len(target)" with passthrough markup. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 18, 2026
  • LexicographicLessEqThreshold + TargetCommitment + ValueThresholdObservable: close ADR-040/048/049 catalog correspondence; refine ADR-035 scope cross-references to ADR-056 across chapters ADR-040: add Bound-shape-to-ObservablePredicate-impl correspondence — every BoundShape catalog primitive (ADR-040) has a foundation-published ObservablePredicate impl (ADR-049), with type:LexicographicLessEqBound realized by LexicographicLessEqThreshold under a new observable:ValueThresholdObservable taxonomy subclass. ADR-048: publish TargetCommitment = SingletonCommitment<LexicographicLessEqThreshold> as the canonical foundation-published search-cost commitment alias. ADR-049: extend foundation-published ObservablePredicate impls from 4 to 5 — add LexicographicLessEqThreshold (struct declaration, reference impl, accept_prob closed form, evaluate semantics under right-aligned big-endian comparison). New observable:ValueThresholdObservable taxonomy subclass parallel to SpectralObservable. Glossary: UOR observable surface 4 -> 5; add LexicographicLessEqThreshold, TargetCommitment, observable:ValueThresholdObservable entries; TypedCommitment entry notes TargetCommitment alias. ADR-013 audit table: ADR-040/048/049 rows reflect the new catalog-correspondence, type alias, and 5th impl. Cross-chapter ADR-056 scope refinement (02, 06, 08, 10, 11, 12): tighten "verb-body" framings of ADR-035's ψ-residuals discipline to "route body's syntactic surface", with compound-verb / axis-impl / resolver bodies explicitly outside scope per ADR-056. Eliminates the prior internal contradiction where chapters described discipline as forbidding emissions in all verb bodies while ADR-056 had refined it to only the prism_model!-declared route closure body. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 17, 2026
  • ADR-056: ψ-residuals discipline scope refinement — applies to route body's syntactic surface only Implementer report: "ADR-035/036 ψ-residuals (wiki constraint, not a foundation-sdk gap). concat/hash/le/lt/ge/gt rejection blocks SHA padding, HMAC, Merkle, tensor saturation. Closure requires wiki amendment." Wiki-internal contradiction surfaced by ADR-055's universal substrate-Term verb body discipline: every axis impl must decompose to substrate Terms, but ADR-035 (broad-scope reading) forbids Concat/Le/Lt/Ge/Gt/AxisInvocation in verb bodies — exactly the primitives SHA padding (Concat + Le/Lt), HMAC (Concat), Merkle (Concat), and tensor saturation (Le/Ge) require. Standard library unimplementable under the broad reading. ADR-056 refines ADR-035's scope. The ψ-residuals discipline applies to the **route body's syntactic surface** — the closure body the author writes directly in the `prism_model!`-declared `route` function. The discipline does NOT apply to: - Compound verb bodies declared via `verb!` per ADR-024 - Axis impl bodies per ADR-055's universal discipline - Resolver bodies per ADR-046 The four categories of "verb body" the wiki recognizes (route body, compound verb body, axis impl body, resolver body) now have explicit scope-boundary placement. The discipline's purpose — keeping admission structurally shaped at the typed-iso surface — is preserved at the route-body layer where it matters; the vocabulary restriction at the substrate layer is lifted. Coordinated updates: - ADR-035 RA7 marked "(*scope refined by ADR-056*)" with the refinement explanation preserving the original rejection rationale per the feedback-implementer-loop discipline. - ADR-046 decision section restructured to recognize all four verb-body categories (route bodies restricted; compound verb bodies + axis impl bodies + resolver bodies unrestricted). Discipline-scope-boundary paragraph in ADR-035 amended to point to ADR-056's refinement. - ADR-013/TR-08 audit-trail table extended with ADR-056 row. - Typed-iso surface consolidation extended with ADR-056. - §8 Closure Across Layers gains a paragraph explaining ADR-056's scope refinement and the canonical compound operations it enables. - §11 TR-14 mitigation posture updated to acknowledge ADR-056's scope refinement: authors restructure search-shaped logic by introducing compound verbs that internally use byte-comparison / byte-concat operations the route's syntactic surface cannot directly emit. - Glossary ψ-residuals discipline entry updated for route-body- syntactic-surface scope; canonical compound operations listed explicitly. Wire format unchanged. ADR-056 is not a substrate amendment — it refines the proc-macro discipline-walk scope; no primitives, Term variants, or wire-format fields change. All 8 validators pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 15, 2026
  • Wiki cleanup: restructure substrate-amendment audit trail; tighten three-fusion-layers duplication; cross-ref TR-14/TR-15 to ADR-055 Three high-impact cleanup items from the post-ADR-055 audit: (1) ADR-013/TR-08 substrate-amendment audit-trail restructured. The mega-paragraph at §9 line 190 (1,084 words classifying ADR-038..ADR-055 as substrate amendments or not) becomes a compact reference table with one row per ADR carrying the yes/no/wire-format-only verdict and a one-sentence summary. Cognitive load down ~70%; all architectural content preserved in scannable form. The ADR-037 reconciliation prose stays as introductory context; the table is the audit trail. (2) Three-fusion-layers verbatim duplication tightened. §4 Strategic Decisions row 55 and §8 Closure Across Layers concluding paragraph both carried the same 3-fusion-layers enumeration verbatim (macro-expansion / catamorphism / monomorphization, with identical sub-bullet phrasing). Both passages now cross-reference ADR-054 / ADR-055 in §9 as the canonical locus for the detailed mechanics; each chapter keeps its scope-appropriate framing (strategic commitment in §4; closure-across-layers reading in §8) without re-stating the full layer mechanics. ~800 words saved while preserving the chapter-cohesive architectural framings. (3) TR-14 (ψ-residuals discipline) and TR-15 (RESOLVER_ABSENT) gain post-ADR-055 cross-reference notes. TR-14's mitigation is now connected to the verb-body/resolver-body scope boundary per ADR-046 and the universal substrate-Term verb body discipline per ADR-055/TR-20. TR-15 clarifies that resolver-tuple selection is independent of axis bodies — resolver-absent retains its runtime mechanism while axis-body discipline moved to compile-time enforcement via ADR-055. TR-20 already references ADR-055 from prior commit; the cluster TR-14/TR-15/TR-20 now reads as a coherent post- ADR-055 surface. All 8 validators pass; wiki end-to-end consistent; auto cache-clear from prior commit kept the build pipeline clean throughout. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 15, 2026
  • Post-ADR-055 consistency: close 5 audit-surfaced gaps that contradicted or omitted universal fold-fusion Broader architectural audit after ADR-055 surfaced 5 places where the wiki text contradicted ADR-055's universal scope or omitted its type-system enforcement. Closed each: (1) ADR-031 'Canonical axis impl body discipline' subsection: removed the directly-contradicting clause "Application-author custom axis impls retain the arbitrary-Rust body permission per ADR-010; the discipline applies to standard-library canonical impls only" — replaced with ADR-055's universal scope and SubstrateTermBody supertrait enforcement. Retitled subsection to acknowledge both ADR-054 + ADR-055. (2) §6 Runtime View Scenario 1 Term::AxisInvocation fold-rule description: was "dispatches through AxisTuple selection to the right axis impl's kernel" — updated to "routes through AxisTuple selection to the right axis impl; per ADR-055's universal substrate-Term verb body discipline, the fold-rule recursively folds the axis impl's SubstrateTermBody::body_arena() static Term slice rather than calling an opaque kernel function." (3) §10 QS-06 Source citation: extended to include ADR-054 + ADR-055. The zero-cost typed-bandwidth admission's structural-enforceability claim now anchors on ADR-055's type-system enforcement explicitly. (4) SD2 Principal Data Path narrative C2 commitment: updated to reflect ADR-055's universal scope. The C2 "one Rust function per forward invocation" claim is now type-system-enforced; axis impls without a substrate-Term `body` clause fail to compile. (5) §12 Glossary `axis!` + `AxisExtension` entries: extended trait- contract description to include `SubstrateTermBody` sealed supertrait bound. The closure-check at axis! invocation now verifies substrate-Term decomposability alongside the prior "sealed, deterministic, no_std-clean, output-bounded" properties. Standard-library sub-crate entry: cross-reference updated from "(ADR-054)" to "(ADR-054 + ADR-055's universal scope)". Wire format unchanged. All 8 validators pass on clean rebuild (build/validate.sh + v2-arc42-build.sh now auto-clear caches per the prior commit). Wiki end-to-end consistent with ADR-055. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 15, 2026
  • ADR-055: universal substrate-Term verb body discipline — fold-fusion guaranteed for every axis impl via type-system enforcement Implementer experience proved the standard-library-only scope of ADR-054's canonical axis impl body discipline (ADR-054 RA2) load-bearing for prism's fold-fusion guarantee. An application-author custom axis with an arbitrary-Rust body produces a Term::AxisInvocation whose body the catamorphism cannot reach; the structural-inference guarantee per ADR-001 + ADR-019 fails at that axis's surface; C2's "one Rust function per forward invocation" claim is provisional, not universal. ADR-055 supersedes ADR-054 RA2 with a universal commitment enforced at compile time: - Foundation declares the sealed `SubstrateTermBody` supertrait on `AxisExtension`. Every axis impl (standard-library AND application- author custom) MUST provide a `body_arena() -> &'static [Term]` accessor returning the Term arena slice the kernel decomposes to. - The catamorphism's `Term::AxisInvocation` fold-rule per ADR-029 recursively folds the axis impl's `body_arena()` slice rather than calling an opaque kernel function. Fold-fusion's structural reach extends to the leaf level across every axis surface for every axis impl. - The `axis!` SDK macro per ADR-030 + ADR-052 extends with a normative `body` clause; the macro emits both `AxisExtension` and `SubstrateTermBody` impls together. Axes lacking a `body` clause fail to satisfy the supertrait bound and fail to compile. - ADR-010's arbitrary-Rust Hasher carve-out is amended: HashAxis impls (per ADR-031's HashAxis-from-prism-crypto framing) are axis impls and follow the universal discipline. Standard-library impls already satisfy this per ADR-054's canonical discipline; application-author custom Hashers provide a `body` clause too. Migration: pre-ADR-055 axes with arbitrary-Rust kernels have two paths per ADR-019's initial-algebra commitment: (1) decompose to substrate Terms via the closure-body grammar; (2) move the off-substrate kernel out of the axis surface to a host-boundary wrapper around `forward(input)`. No third option — operations that cannot be expressed in substrate Terms are not prism operations. Updates across the wiki for consistency: - §4 Strategic Decisions row updated to reflect universal scope and drop the standard-library-only framing. - §8 Closure Across Layers concluding paragraph updated similarly. - Conceptual Model intro C2 mechanism updated; "structurally guaranteed by the type system, not aspirational." - ADR-013/TR-08 substrate-amendment paragraph + typed-iso surface consolidation paragraph extended with ADR-055. - TR-20 rewritten: discipline regression is now a compile-time error, not a runtime / inclusion-criterion concern. Migration risk for pre-ADR-055 axes documented. - ADR-054's Rejected alternative 2 marked `(*superseded by ADR-055*)` with the supersession reasoning preserved per the feedback-implementer-loop discipline. - Glossary: new `SubstrateTermBody` trait entry; `Substrate-Term verb body` entry updated for universal scope. Wire format unchanged. All 8 validators pass on clean rebuild. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 15, 2026
  • Eradicate \`-escape rendering artifacts: AsciiDoc {empty} boundary attribute + passthrough conversion for angle-bracket spans Root cause: AsciiDoc smart-quote substitution + Pandoc DocBook→markdown backtick-escape interaction. When an AsciiDoc inline code-span `text` is immediately followed by a trigger character (letter, ASCII apostrophe, or ASCII double-quote), Pandoc escapes the opening backtick because the smart-quote substitution makes the code-span boundary ambiguous in markdown. Result: literal `\\`text` artifacts in the rendered wiki — the closing backtick is consumed by the entity, and the opening backtick appears as `\\`` text rather than a styled code span. Two-part fix: (1) Insert AsciiDoc `{empty}` attribute substitution between the closing backtick and the trigger character via a state-machine parser (`/tmp/fix_backtick_boundary.py`). `{empty}` is a zero-width attribute that breaks the AsciiDoc/Pandoc bonding without affecting rendered output. The parser tracks opening/closing backtick pairing using AsciiDoc constrained-span rules (opening flanked by start-of- line, whitespace, or specific opening punctuation; closing flanked by non-whitespace), so it only inserts at true closing positions. Applied 462 fixes across 14 .adoc source files (12 arc42 chapters plus 2 ISO 15288 lifecycle pages). (2) Convert 6 specific AsciiDoc inline code spans containing `<...>` angle brackets to `++...++` passthrough form. AsciiDoc inline code spans wrapping content with HTML-significant characters get the angle brackets escaped through the DocBook→markdown conversion, producing literal `\\<` and `\\>` plus the opening-backtick escape. Converting these to `++...++` passthrough preserves the content verbatim (the rendered output remains code-styled monospace). Fixed spans: ADR-026 G6 (match grammar, 4 spans); glossary's first_admit + fold_n + Term::FirstAdmit grammar forms. Build pipeline preserved; all 8 validators pass on clean rebuild (vendor/arc42-generator/build/ cleared); zero `\\` artifacts remain in any of the 22 rendered wiki pages. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 15, 2026
  • ADR-054: Fold-Fusion Principle + ADR-031 amendments + Conceptual Model C2 mechanism ADR-054 names fold fusion as the architectural mechanism that realizes C2 (zero runtime movement per inference per the Conceptual Model's cost commitments). Every prism transformation — verb-body, prism_model route, resolver body, standard-library canonical axis-impl body — is a folding operation, and the catamorphism's initial-algebra universal property (ADR-019) fuses composed folds into one Rust routine per substitution-axis selection. Three fusion layers stack: (1) Macro-expansion fusion — verb-call sites lower to Term subtrees at proc-macro expansion (ADR-024); the verb-call walker is absent at run time. (2) Catamorphism fusion — per-variant fold-rules (ADR-029) compose universally per the initial-algebra property; there is no choice about how to fuse them. (3) Monomorphization fusion — per-substitution-axis-selection Rust routine has no dynamic dispatch and no allocation (TC-01); LLVM inlining collapses per-variant dispatch with fold-rule bodies into straight-line code. Standard-library canonical axis impl body discipline: every published canonical impl carries a substrate-Term verb body, so the catamorphism's structural reach has no opaque axis-kernel boundary inside the published canonical surface. Application-author custom axis impls retain ADR-010's arbitrary-Rust permission. Substrate-Term realization examples cover the operation-space dichotomy: prism-numerics (algebraic/laws-structured — modexp_p, polyeval, gcd, ext_euclidean, field_<op><P>, fma, newton_step, horner) and prism-crypto (cryptographic/iteration-structured — SHA-2, SHA-3, BLAKE3 compression rounds as substrate-Term W32/W64 ops + rotate). Both reduce to fold compositions under the principle. ADR-031 amendments per the implementer patch: - "Application-neutral within domain" → "Application-neutral" with substrate-faithful framing (conventional groupings are presentation conventions, not architectural partitions). - Roster reframed as operational organization, not architectural commitment: reorganizing the roster is admissible operational policy so long as the disciplines (ADR-024 closure, ADR-017 content-addressing, ADR-031 conformance, ADR-054 body discipline) are preserved. - New "Conventional shape and axis naming" section: BigInt<MaxBits>, FixedPoint<I,F>, Tensor<Element,Shape>, Digest<N>, etc. are conventional aliases for substrate-level ConstrainedTypeShape constructions. - New "Canonical axis impl body discipline" section per ADR-054. - Deduped resolver-baselines paragraph; second occurrence updated to drop "application-neutral-within-domain" → "application-neutral". ADR-013 substrate-amendment paragraph + typed-iso consolidation extended with ADR-054 (not a substrate amendment — names a property the architecture already realizes; only substrate-adjacent change is the canonical axis impl body discipline on ADR-031's inclusion criteria). Conceptual Model intro's C2 commitment now points to the Fold-Fusion Principle as its structural mechanism, and SD2 Principal Data Path narrative carries the same cross-reference. Glossary entries added: Fold-Fusion Principle, Substrate-Term verb body. Folding-transformation entry updated to cross-reference ADR-054. Standard-library sub-crate entry updated to reflect the roster's operational (not architectural) framing. All 8 validators pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 15, 2026
  • Post-ADR-050..053 consistency pass: catalog count, Γ form, new risks Five remaining stale references closed after the ADR-050..053 additions: - §8 Concepts F-signature: PrimitiveOp catalog enumerated as 18 variants at the post-ADR-053 snapshot (was 15 pre-ADR-053). The six ring-axis primitives Γ = {Add, Sub, Mul, Div, Mod, Pow}, the four hypercube-axis ops {Xor, And, Or, Bnot}, the unary set, byte-level comparisons, and Concat. - §9 ADR-026 G12 disambiguation: lowercase PrimitiveOp identifier list updated to 18 variants (added div, mod, pow). - §11 TR-12 macro identifier-resolution: same 18-variant update. - §12 Categorical X regime split glossary: Γ corrected from the five-element form {+, −, ×, ÷, ^} to the six-element form {Add, Sub, Mul, Div, Mod, Pow} per ADR-053. Added cross-reference to ADR-050's full-Witt-tower evaluation discipline. - §11 gains two new risk entries: - TR-18 — width-parametric arithmetic cold-path may surprise authors at wide Witt levels. Per-op cost characterization: O(width) for additive ops, O(width^2) for Mul/Div/Mod, O(width^2 · log(exp)) for Pow. Mitigation via algorithmic optimizations within the strategy-selection lane (Karatsuba, Barrett, sliding-window — byte-output-equivalent to reference impl). - TR-19 — Div/Mod zero-divisor runtime path. ShapeViolation carrying https://uor.foundation/op/Div/nonZeroDivisor surfaces at evaluation time, not at proc-macro expansion. Mitigation: authors compose b > 0 precondition through Term::Match or first_admit on input metadata; foundation reference impl checks unconditionally at evaluation time. All 8 validators pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 15, 2026
  • ADR-050..053: width-parametric arithmetic + wide literal carrier + axis-generic SDK emission + Div/Mod/Pow catalog extension Four ADRs plus two inline amendments commit substrate arithmetic at the full Witt tower, widen the literal carrier to match, and admit parametric impl types in the SDK's axis-extension emission. Per the implementer document (prism repo implementing uor-foundation/SDK). - ADR-050 (width-parametric arithmetic fold-rules) — the ring-axis and hypercube-axis arithmetic PrimitiveOps (Neg/Bnot/Succ/Pred/ Add/Sub/Mul/Xor/And/Or, plus Div/Mod/Pow per ADR-053) evaluate at the full Witt tower with width = max(operand widths) for binary ops. The pre-ADR fold-rules silently truncated to u64, a Correction-4-class soundness gap. Operands wider than result_width are now rejected as ShapeViolation::CardinalityViolation; comparison primitives Le/Lt/Ge/Gt and Concat retain byte-level semantics. - ADR-051 (wide-value Term::Literal carrier) — Term::Literal widens from `value: u64` to `value: TermValue` (byte sequence sized to level.witt_length() / 8). Wide constants become single records instead of Concat-of-narrow chains. SDK helper literal_u64(value, level) preserves the narrow-construction surface. - ADR-052 (axis! macro generic-form emission) — the companion axis_extension_impl_for_<name>! macro accepts :ty (not just :ident), admitting parametric impl types like BigIntModularNumeric<BYTES> without hand-written AxisExtension boilerplate. Existing call sites continue to work; SDK seal discipline per ADR-022 D1 preserved. - ADR-053 (Div/Mod/Pow PrimitiveOp catalog extension) — catalog extends from 15 to 18 variants. Each is a folding-transformation per ADR-053's coined principle (one fold = one Term variant, not multiple Term variants the catamorphism walks). Div/Mod are Euclidean (b=0 → ShapeViolation with the nonZeroDivisor IRI); Pow is square-and-multiply modular exponentiation. Identity individuals DV_1..DV_4 + PW_1..PW_3 added to the Lean corpus commitment. GeometricCharacter gains Quotient/Remainder/ IteratedScaling. Inline amendments: - ADR-026 line ~670: Γ = {Add, Sub, Mul, Div, Mod, Pow} (six ring- axis primitives on Z/(2^n)Z) — corrected from the prior five- element form. Hypercube-axis ops {Xor, And, Or, Bnot} remain separate per the MetricAxis enum. - ADR-031 numerics sub-crate roster: substrate-native ring arithmetic is closed under PrimitiveOp::{Add,Sub,Mul,Div,Mod,Pow}; substrate-native hypercube arithmetic under {Xor,And,Or,Bnot}. BigIntAxis/FixedPointAxis/RingAxis dissolve into the substrate PrimitiveOp catalog at the appropriate Witt-level. FieldAxis retained for prime parametricity (PrimeFieldNumeric<P> instances). prism-numerics' role shifts to publishing shapes plus compound arithmetic verbs (modexp_p, polyeval, gcd, ext_euclidean, horner, newton_step, fma, field_<add|sub|mul|inv><P>). ADR-013 substrate-amendment paragraph + typed-iso surface consolidation paragraph updated. Glossary entries added for Folding-transformation (ADR-053's coined principle), literal_u64 (ADR-051's SDK helper), Term::Literal (wide-value carrier). PrimitiveOp glossary entry updated to 18-variant catalog at the post-ADR-053 snapshot. Gauge-tower Γ entry corrected to the six-element form. All 8 validators pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 15, 2026
  • Close remaining wiki gaps surfaced by post-ADR-047..049 review Final completeness pass over the wiki after the ADR-047..049 additions surfaced four classes of remaining items: - Five-parameter shape consistency: ADR-029 prose (the running catamorphism description) and §8 concepts (the closure-across- layers section) updated from "four substrate parameters" to "five substrate parameters" with cross-references to ADR-048. §2 architecture constraints table extended with a TypedCommitment row describing the fifth model-declaration parameter. - ADR-048 wire-format extension paragraph clarified: the CommitmentEvaluated trace event's variant-discriminator byte is foundation-assigned (next free after the existing five variants), not hard-coded; pipeline::run_route extends from five-type- parameter (<H, B, A, M, R>) to six (<H, B, A, M, R, C>). - §11 Technical Risks gains two entries: - TR-16 σ-Projection Hardening Principle non-conformance — structural attack surface a non-conforming Hasher impl exposes; mitigation via foundation CI running axis::cryptanalyze per ADR-049 plus third-party axis crates declaring U1–U6 conformance. - TR-17 deeply-nested AndCommitment compile-time cost — the architectural trade-off between runtime zero-cost and compile- time monomorphization work at large K; mitigation via shallower sub-commitment factoring at architectural boundaries. - §10 Quality Requirements gains QS-06 (zero-cost typed-bandwidth admission) with measurable response criteria: zero allocations attributable to C; bandwidth scaling matching α⁻¹·2^K within ±30% at N=100 and ±10% at N=10⁴, per Hardening Principle U6 (ADR-047). - §12 Glossary gains EmptyCommitment, SingletonCommitment<P>, and AndCommitment<A, B> entries (the three built-in TypedCommitment implementors per ADR-048). All 8 validators pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 14, 2026
  • Sharpen ADR-048/049 with reference-impl skeletons; add CommitmentEvaluated trace event To make ADR-048 and ADR-049 implementation-ready for the foundation team, both ADRs now carry a "Reference implementation surface" subsection with paste-ready Rust skeletons, explicit module paths, and the macro-syntax extension: - ADR-048: Foundation places TypedCommitment + 3 built-in impls in uor_foundation::commitment; the 5th model parameter lands on the pipeline::PrismModel trait with a sealed C: TypedCommitment bound defaulting to EmptyCommitment; pipeline::run_route extends with the &C value-reference parameter and the post-resolver C::evaluate(kappa_label) consultation. The prism_model! macro parses an optional 5th type-position clause in the impl block. - ADR-049: Foundation places the typed observables and ObservablePredicate trait in uor_foundation::observable::predicate; the cryptanalysis battery lives in uor_foundation::axis. Each observable carries a concrete evaluate() body with the p-adic valuation / Walsh–Hadamard parity / ultrametric distance / single-bit-parity computation. CryptanalysisReport carries StructuralReport (§A–§H) + CalibrationReport (§I–§J) with per-test critical thresholds at α = 0.001. - §8 Trace Wire Format: CommitmentEvaluated event variant added per ADR-048 with verifier-check semantics and the default-EmptyCommitment emission rule. The wiki commitments are now spec-complete for the foundation team to implement against. Five upstream items remain blocking prism-btc completion at foundation 0.4.5 (TypedCommitment + 3 impls; 5th PrismModel param; prism_model! macro extension; ObservablePredicate + 4 typed observables; axis::cryptanalyze battery); the wiki now specifies each at trait-signature + module-path + emission-rule detail. All 8 validators pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 14, 2026
  • ADR-047..049: generalize implementer-shipped zero-cost patterns + close upstream items Three coupled architectural commitments close the load-bearing implementer-surfaced gaps from a downstream implementation that shipped a typed-bandwidth admission surface, a UOR observable surface, and an empirical cryptanalysis battery — patterns the architecture must support in a generalized way so any implementor achieves zero-cost operations from the wiki alone. - ADR-047 (σ-Projection Hardening Principle) — commits six structural axioms (U1 marginal-uniformity, U2 joint-independence, U3 admission- orthogonality, U4 avalanche, U5 autocorrelation-flatness, U6 bandwidth-additivity) qualifying Hasher impls as UOR-hardened canonical hash axes per ADR-030. Property axiom on existing substrate constructs; no wire-format change. - ADR-048 (TypedCommitment + 5th substrate parameter) — commits the zero-cost typed-bandwidth admission surface: foundation-declared TypedCommitment: Copy trait with three built-in impls (EmptyCommitment, SingletonCommitment<P>, AndCommitment<A,B>); fifth model-declaration parameter C: TypedCommitment on PrismModel defaulting to EmptyCommitment; the catamorphism's ψ_9 dispatch consults C::evaluate(kappa_label) before sealing the success envelope. Substrate amendment under ADR-013/TR-08; trace gains CommitmentEvaluated event variant; wire-format version bumps. Pulls admission fully inside the catamorphism per the implementer's "residual upstream move" identification. - ADR-049 (UOR observable surface + cryptanalysis battery) — commits four foundation-declared typed observable primitives (Stratum<P>, WalshHadamardParity, UltrametricCloseTo<P>, AffineParity) under ADR-038's closed Observable taxonomy plus a new SpectralObservable subclass; commits axis::cryptanalyze::<H: Hasher>(samples) -> CryptanalysisReport as the substrate-level test primitive witnessing the Hardening Principle's U1-U6 axioms empirically. Substrate amendment under ADR-013/TR-08; new Observable subclasses bump the wire-format version. ADR-013 substrate-amendment paragraph and ADR-020 typed-iso surface consolidation updated with cross-references. Five glossary entries added (σ-Projection Hardening Principle, TypedCommitment, Typed- bandwidth admission, UOR observable surface, UOR Cryptanalysis Battery). All 8 validators pass. The wiki remains strictly domain-neutral: no downstream-implementer crate is named anywhere in src/ or rendered output. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 14, 2026
  • ADR-037: clarify HostBounds capacity-cap totals (22 migrated + 4 pre-existing = 26) The prior phrasing said "22 new associated constants" without disambiguating the 22 newly-migrated caps (14 free-standing data-shape caps + 8 ψ-stage resolver output ceilings) from the 4 pre-existing HostBounds fields per ADR-018 (FINGERPRINT_MIN_BYTES, FINGERPRINT_MAX_BYTES, TRACE_MAX_EVENTS, WITT_LEVEL_MAX_BITS). Read naively, implementations could miscount the surface or assume the pre-existing fields were superseded. Clarified phrasing makes both groups explicit and notes the 26-constant total. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 13, 2026
  • ADR-043..046: close implementer-surfaced gaps in resolver-body discipline + product-field metadata + Grounded::tag affordance Four new ADRs complete the wiki's coverage of architectural surfaces the prism-btc implementation depends on: - ADR-043 (Iterative-Resolution Discipline) — canonicalizes the resolver-internal computational pattern for resolver-bound ψ-stages: FreeRank protocol; bounded-domain enumeration over ConstrainedTypeShape::CYCLE_SIZE; convergence vs exhaustion termination; impossibility-witness emission per ADR-039; host- boundary retry. Not a substrate amendment — canonicalizes existing substrate constructs. - ADR-044 (PartitionProductFields trait) — amends ADR-033's partition_product! emission shape to a split-const form on a foundation-declared trait: FIELDS: &'static [(u32, u32)] + FIELD_NAMES: &'static [&'static str]. Macro-emitted impls and hand-written impls share the same metadata surface. - ADR-045 (Grounded::tag::<NewTag>()) — adds the zero-cost re-tagging consuming method to Grounded<T, Tag>. Preserves the seal regime; applications surface domain-distinct Grounded types without newtype boilerplate. - ADR-046 (Resolver-body discipline scope) — records the boundary between ADR-035's verb-body ψ-residuals discipline and ADR-043's resolver-body iterative-resolution discipline. Structural, not stylistic. Cross-references added to ADR-013 substrate-amendment paragraph and ADR-035 ψ-residuals discipline. Glossary entries added for the new concepts; FreeRank entry cross-references ADR-043. All 8 validators pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 13, 2026
  • Fix AsciiDoc parsing errors blocking rendered .md regeneration Three classes of parsing errors were preventing the arc42 build pipeline from regenerating chapters 06/08/09/12 from their AsciiDoc sources: - Table cell separators inside backticks: 04 line 40 contained `first_admit(<domain>, |<idx>| <predicate>)` and `fold_n(n, init, |state, idx| step)` — the `|` characters parsed as table cell separators, breaking a 3-column row into 5 cells. Escaped the closure-parameter bars with `\|`. - Rust lifetime apostrophes inside backticks: backtick-quoted spans containing patterns like `<'_, Output>` or `<'a, T>` produced interleaved emphasis/literal tags in the DocBook intermediate that Pandoc rejected. Converted ~39 such spans across 06/08/09/12 to AsciiDoc inline literal passthrough (`++...++`). - Asterisks inside backticks: `*Record<H>` and `#![no_std]` similarly produced overlapping emphasis/literal tag sequences. Converted to `++*Record<H>++` and `++#![no_std]++` passthrough. Source semantics are unchanged — passthrough renders identically to backtick monospace once the parsing path is unblocked. All 12 rendered chapter .md files now reflect the current source content (V1 placeholder resolutions in §6 and §8, ADR-040/041/042 in §9, glossary updates in §12). All 8 validators pass. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 13, 2026
  • Tighten + clean up: resolve ψ-residuals contradictions, split mega-paragraphs, align glossary ψ-residuals contradiction resolution: - Section 06 Scenario 1 step 9 + Properties bullet 4: drop "Routes whose closure body contains hash(input) invocations emit Term::AxisInvocation"; replace with canonical-hash-axis type-associated framing + ψ-residuals discipline note (resolver impls invoke canonical hash internally; verb bodies do not) - ADR-029 Consequences: rewrite from "model whose route uses hash(input) emits Term::AxisInvocation" to "verb body emits ψ-residual-free Term variants; resolver-bound ψ-Term fold-rules consult R" - ADR-029 Substitution-axis-realized verbs: reframe as substrate context (not verb-body context); explicitly note Term::AxisInvocation is forbidden in verb bodies per ψ-residuals discipline but remains substrate-present for resolver bodies + conformance corpus + trace replay - ADR-029 Catamorphism parameterization in (H, B, A) → (H, B, A, R): four substrate parameters; resolver-bound ψ-Term fold-rules consult R at evaluation time - ADR-030 Term enum's commitment: existing Hasher migration through Term::AxisInvocation removed; canonical hash axis is consumed at certificate-emission stage + internally by resolvers, not by verb bodies; hash(...) identifier remains reserved but verb-body emission is rejected by the ψ-residuals walk Mega-paragraph splits: - ADR-013 Consequences: split one ~10-sentence paragraph into six thematic paragraphs (TR-08 maintenance coupling / three-layer-closure framing / ADR-030+031 universal-extension scope / substrate amendments to date / ADR-035 commitment / ADR-036 supporting surface) - ADR-020 Consequences: split one ~8-sentence paragraph into four thematic paragraphs (compile-time trait commitments / four-parameter PrismModel signature + run_route call-site / Output declaration + value-payload carrier / typed-iso surface summary) Glossary alignment: - Hasher entry: drop "Existing wiki text and ADRs predating ADR-030 use 'Hasher' in the original sense" disclaimer; add ADR-036 + Hasher bound mechanism showing how resolver-bound substrate operations consume canonical hash axis type-associatedly via <A as Hasher>::initial() All 8 validators pass; Stage 5 idempotent. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 11, 2026
  • Audit-gap-closure pass: ADR-007 fourth-parameter framing, ADR-026 G19 ψ-residuals restriction, Section 8 enriched signature + Hashing Substrate Contract, TR-14 + TR-15 ADR-007: third-position bound strengthened from A: AxisTuple to A: AxisTuple + Hasher per ADR-036; fourth substrate parameter R: ResolverTuple defaulting to NullResolverTuple noted as parallel substrate parameter (not a substitution axis architecturally) with per-value content provision (instance-associated, value-reference threading) vs per-domain operation kernels (type-associated, type-parameter threading) ADR-026 G19: ψ-residuals discipline restriction surfaced at the G19 grammar form definition — Term::AxisInvocation is a ψ-residual; verb-body emissions are rejected at proc-macro expansion with a closure-violation error citing the verb's name; the variant remains in the substrate for resolver bodies, conformance-corpus generators, foundation-internal trace replay; authors who naturally needed hash(input) in verb bodies express the equivalent through resolver-bound ψ-Term variants whose resolver impls invoke the canonical hash axis internally Section 8 Foundation as a Signature Category — Signature Endofunctor: enriched-signature description extended with substitution-axis-invocation, field-access projection, bounded-search-with-short-circuit, and the nine ψ-chain operators; PrimitiveOp catalog reference updated to snapshot-specific (foundation 0.3.6 canonical catalog); ψ-residuals discipline scopes the verb-body sub-signature Section 8 Foundation as a Signature Category — Catamorphism: carrier parameterization updated to four parameters (H, B, A: AxisTuple + Hasher, R: ResolverTuple defaulting to NullResolverTuple); model declaration's resolver selection participates in carrier determination Section 8 Hashing Substrate Contract: preamble extended to thread the Hasher trait through ADR-030 (consumed via AxisTuple), ADR-031 (HashAxis subsumes the prior Hasher), and ADR-036 (A: AxisTuple + Hasher bound enables resolver-bound substrate operations to consume canonical hash axis type-associatedly) TR-14 (new): ψ-residuals discipline rejects naturally search-shaped verb bodies at proc-macro expansion — structured-error requirements (rejected variant + offending span + ψ-residual category + architectural reason + recommended restructuring path); the discipline is load-bearing for verifiability commitment per ADR-001 + ADR-019; foundation does not relax for application convenience TR-15 (new): RESOLVER_ABSENT runtime path is the only mechanism reporting resolver-absent at evaluation time — NullResolverTuple Null* impls return ShapeViolation carrying RESOLVER_ABSENT shape IRI + foundation-internal discriminator byte; mitigation posture names SDK proc-macro-time warning analysis (paralleling ψ-residuals walk) emitting at compile time when verb-body ψ-Term emissions require categories the model's R doesn't declare; the runtime mechanism is the load-bearing path, proc-macro warning is a diagnostic convenience All 8 validators pass; Stage 5 idempotent. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 11, 2026
  • ADR-035 + ADR-036 reframe: canonical k-invariants branch + ψ-residuals discipline + uniform byte-buffer resolver surface + R = NullResolverTuple default + no axes value parameter ADR-035 reframe (Block A): - Title extended: "Canonical ψ-pipeline + ψ-chain Term variants + ψ-residuals discipline on verb body composition" - Decision restructured into three coupled amendments: (1) nine ψ-Term variants, (2) canonical k-invariants branch ψ_1 → ψ_7 → ψ_8 → ψ_9 as the architectural composition for the verifiability commitment, (3) ψ-residuals discipline forbidding Term::FirstAdmit, Term::AxisInvocation, byte-comparison/concat PrimitiveOps in verb-body terms - Conceptual model paragraph: compiled binary IS the canonical circuit; data representation minimal (input bytes + κ-label); client conceptually larger than data; complete model fuses into a single reference - Closure check at verb!/prism_model! invocation extended with ψ-residuals walk - Composition example switched from `betti(homology_groups(chain_complex(nerve(input))))` to canonical `k_invariants(homotopy_groups(postnikov_tower(nerve(input))))` - Rejected alternatives 5-7 rewritten: partial-coverage / homology-branch / permit-ψ-residuals ADR-036 reframe (Block B): - PrismModel<H, B, A, R = NullResolverTuple> trait shape with R defaulting to NullResolverTuple; A: AxisTuple + Hasher bound (canonical hash axis bound) - ResolverTuple trait: drop per-category accessor methods (live in marker traits); add RESOLVER_ABSENT_DISCRIMINATOR byte const - Eight resolver traits: uniform writer-style byte-buffer surface `fn resolve(&self, input: &[u8], out: &mut [u8]) -> Result<usize, ShapeViolation>` (no typed *Record<H> returns) - resolver! macro emits all eight Has<Category>Resolver impls unconditionally (declared fields delegate; undeclared categories delegate to NullResolverTuple Null* impls) - pipeline::run_route<H, B, A, M, R>(input, resolvers: &R): five type params, no axes value parameter, all eight Has<Category>Resolver<A> bounds unconditionally - Application-side instantiation: no axes value (axis methods type-associated via <A as Hasher>::initial()) - MAX_RESOLVER_TUPLE_ARITY = 16 explicit (twice AxisTuple cap) - Replace GenericImpossibilityWitness with ShapeViolation + RESOLVER_ABSENT shape IRI + foundation-internal RESOLVER_ABSENT_DISCRIMINATOR byte - Rejected alternatives reorganized: AxisTuple-conflation / typed-records / ad-hoc-per-category-params / global / foundation-provided-impls Block C cross-references: - ADR-013 Consequences: ψ-residuals, canonical k-invariants branch, byte-buffer resolver surface, RESOLVER_ABSENT_DISCRIMINATOR - ADR-022 D4 comment block: <H, B, A: AxisTuple + Hasher, R = NullResolverTuple>; asymmetric threading rationale - ADR-022 D4 trait declaration: same shape - ADR-022 D5 run_route signature: drop axes value param, add unconditional Has<Category>Resolver<A> bounds - ADR-022 D5 forward body: run_route::<H, B, A, Self, R>(input, &resolvers) - ADR-026 G12 + TR-12 + ADR-030 closure-check: already at twenty reserved identifiers (no change) - ADR-029 fold-rule preamble: four substrate parameters; ψ-residuals discipline - ADR-029 ψ-Term fold-rules: byte-buffer resolver.resolve(input, out_buffer) signatures; KInvariants noted as canonical-branch terminal - ADR-029 variant-count statement: FirstAdmit noted as verb-body-forbidden - ADR-030 Consequences: ResolverTuple parallels AxisTuple with asymmetric threading - ADR-031 Consequences: resolver baselines parallel axis baselines Block D Section 4 + Section 8: - Strategic decision row updated: canonical k-invariants branch + ψ-residuals + R = NullResolverTuple + conceptual-model framing - Closure Across Layers paragraphs: canonical k-invariants branch + ψ-residuals + byte-buffer surface Block E glossary: - New: Canonical k-invariants branch (under C), NullResolverTuple, RESOLVER_ABSENT_DISCRIMINATOR, Resolver trait shape (under R/N), ψ-chain umbrella + ψ-residuals discipline (Greek-letter cluster) - Updates: ψ-chain resolver traits to byte-buffer surface; Has<Category>Resolver<H> to all-eight-emission pattern; PrismModel to new signature + run_route call; ResolverTuple drop per-category accessor mention; ResolverCategory annotated with ψ-stage numbers; MAX_RESOLVER_TUPLE_ARITY explicit value 16; resolver! to all-eight-emission pattern; nine Term:: variant entries annotated with canonical-branch / strictly-weaker positions Reconciliation in Sections 02/03/05/06/07/08/10 + 15288 docs: - TC-03 substitution axes table: R = NullResolverTuple default; A: AxisTuple + Hasher bound; byte-buffer signature; ψ-residuals - Section 03 author-input row: R optional with default - Section 05 pipeline component: run_route<H, B, A, M, R>(input, &resolvers) signature; no axes value param - Section 06 Scenario 1 steps 7-8: four substrate parameters, no axes value param; canonical k-invariants branch; ψ-residuals discipline - Section 07 deployment view executable: canonical resolver branch list; type-associated axis dispatch - Section 08 PrismModel trait + forward call: same shape - Section 10 QS-01: canonical-branch resolver list; byte-buffer surface; ψ-residuals - 15288 technical.adoc + technical-management.adoc: R optional with default; A: AxisTuple + Hasher; canonical k-invariants branch All 8 validators pass; Stage 5 idempotent. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 11, 2026
  • Cleanup: remove stray prism-btc / Bitcoin-mining domain examples; keep wiki domain-neutral The wiki accumulated domain-specific examples (prism-btc, Bitcoin proof-of-work, nonce_fiber_traversal, sha256d, MerkleRoot, etc.) as illustrative cases for ADRs that introduced architectural mechanisms. Per the framework's universal-model-compiler scope (ADR-031: prism-btc is one third-party Layer-3 implementation among many; the architecture commits to no specific domain over alternatives), the wiki should be domain-neutral. This commit replaces all such examples with generic framings. ADR-009 (consumer-class IL parameterization): - "Bitcoin Proof-of-Work consumer" -> "content-addressing consumer" - "Financial-substrate / Bitcoin proof-of-work consumers" row -> "Cryptographic-search consumers / High-cardinality bounded first-admission search applications" - "Bitcoin-substrate consumer's V&V plan" -> "cryptographic-search consumer's V&V plan" ADR-022: - D2 closure-body example: sha256d(serialize_header(input.template_prefix, input.nonce)) -> hash(concat(input.left, input.right)); MyHasher -> MyAxisTuple - D3 "prism-btc-shaped failure mode" -> generic "the failure mode the closure-bodied form is meant to avoid" - D4 axis-selection example: (Sha256dHasher,) -> (MyHashAxisImpl,) with hash-axis-as-content-addressing context ADR-024: - Layer-3 verb examples: Sha256Compression, MerkleRootDerivation, Sha256dProjection, HeaderSerialization, NonceFiberTraversal -> "each Layer-3 crate declares verbs whose names and semantics are domain-specific to that crate's substrate" - verb! example: pub fn sha256_compression(...) -> pub fn my_named_verb(...) - Rejected alternative 2: "Bitcoin-substrate implementation's verbs are not appropriate for an FHE-substrate implementation" -> generic Layer-3-vs-Layer-3 framing ADR-026: - "FHE-substrate consumer ... Bitcoin-substrate consumer (W32 fiber, SHA-256d hasher)" -> "FHE-substrate ... cryptographic-search (W32 with content-addressing axis) ... tensor-compute" examples ADR-031: - "Bitcoin, Ethereum, ZK applications" reuse example -> "content- addressing, signature-verification, ZK proof systems" reuse example - "prism-btc crate's Bitcoin-specific verbs" -> generic "third-party crate's application-domain-specific verbs" - "prism-btc depends on prism (consuming HashAxis::sha256 for SHA-256d composition; the prior Sha256dHasher impl migrates to a verb...)" -> generic protocol-specific Layer-3 application crate framing - Bitcoin-substrate / Bitcoin verbs framing in *Context* paragraph -> cryptographic-substrate framing - Consequences "Bitcoin pi-projection / SHA-256d-MAC" example -> "content-addressing search / keyed-MAC computation" ADR-032 Consequences: - prism-btc's nonce_fiber_traversal verb body declares first_admit(W32, |nonce| ...) -> "A model that declares first_admit(W32, |idx| ...) over a 32-bit search domain" ADR-033 Context + Consequences: - prism-btc's nonce_fiber_traversal field example -> generic "representative example: a verb that performs bounded first-admission search over a 32-bit candidate domain" - partition_product!(Prefix76, Target32) -> partition_product!(PrefixShape, ThresholdShape) - input.prefix / input.target / |nonce| -> input.left / input.threshold / |candidate| - "lines 102-104 of crates/prism-btc/src/verbs.rs at HEAD b233785" -> removed (specific cross-repo line reference) - prism-btc's MiningInput -> "Layer-3 crate's Input shape" - "the entire mining inference flows through pipeline::run_route" -> "the entire bounded-search inference flows through pipeline::run_route" - prism-btc's traverse_sequential / traverse_parallel retire -> "Layer-3 implementations that previously carried out-of-band traversal helpers can retire those helpers" ADR-034 Context + Decision + Consequences + Rejected alternatives: - "the prism-btc Bitcoin mining case at WittLevel::W32 requires ~6 hours wall-clock per call" -> "a WittLevel::W32 search domain at 2^32 = 4294967296 candidates would require iterating the full domain ... taking many hours of wall-clock" - prism-btc's ops/traversal::traverse_sequential to retire -> "Layer-3 implementations' bounded-search runtime helpers to retire" - partition_product!(MiningInputShape, prefix: Prefix76, target: Target32) -> partition_product!(MyInputShape, prefix: PrefixShape, threshold: ThresholdShape) (in field-access example) - "for example, nonce in the prism-btc Bitcoin mining verb" -> generic "candidate, key, index" identifier examples - TermValue capacity example "(the prism-btc Bitcoin mining case)" -> "a WittLevel::W32 search domain" - Consequences verb-body example: first_admit(W32, |nonce| hash(concat(input.prefix, nonce)) <= input.target) over MiningInputShape -> first_admit(W32, |candidate| hash(concat(input.left, candidate)) <= input.threshold) over MyInputShape - "Bitcoin mining predicate that calls HashAxis::sha256" -> "content-addressing search predicate" - Rejected alternative 2: "prism-btc's traverse_sequential cannot retire" -> "Layer-3 implementations' bounded-search runtime helpers cannot retire" Final scan: no remaining prism-btc, prism_btc, Bitcoin, sha256d, Sha256d, nonce_fiber, NonceFiber, MerkleRoot, HeaderSerial, Sha256Compression, MiningInput, template_prefix, traverse_sequential, or traverse_parallel references in src/, wiki-definition.md, or README.md. SHA-256 mentions remain only where legitimate (prism-crypto HashAxis families per ADR-031; Hasher contract examples in ADR-008 + section 8 + section 11; structurizr.war SHA-256 build pipeline checksum in 15288 technical-management). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 10, 2026
  • Audit-gap-closure: fix V2 build invalid-XML, complete twelve-variant + G1-G20 consistency Critical fix: previous commit cf3ce00 (ADR-034) and 8672dfd (ADR-034 audit) silently produced rendered output missing ADR-034 content. Root cause: two AsciiDoc syntax issues in ADR-034 produced invalid DocBook XML that pandoc rejected for 09_architecture_decisions.xml, leaving the rendered chapter staged at the pre-ADR-034 state. The underlying gap was visible only after re-running build.sh with a clean vendor/arc42-generator/build/ — then V2 surfaced "Pandoc failed for 09_architecture_decisions.xml: Invalid XML." AsciiDoc fixes: - ADR-034 *Mechanism 1* paragraph: removed [quote]/____ block (the first quote-block in the file; AsciiDoctor's DocBook output for this block produced unbalanced literal/emphasis tags). Inlined the fold-rule update prose into the surrounding paragraph. - ADR-031 *Decision*: replaced *closure under `uor-foundation`'s vocabulary* with *closure under `uor-foundation`* (per ADR-013). AsciiDoctor parsed the inner backtick-then-apostrophe construct inside the surrounding emphasis as overlapping literal/emphasis spans, producing invalid XML. After these fixes V2 succeeds; the rendered 09-Architecture-Decisions.md now has 65 ADR-034/FirstAdmit mentions (up from 0). Twelve-variant consistency completion: - Section 8 PrismModel concept entry: Term-enum listing extended from 10 variants ending at HasherProjection to 12 variants (..., AxisInvocation, ProjectField, FirstAdmit) with proper ADR attribution. G1-G20 consistency: - Section 5 BBV operation-declaration vocabulary description. - Section 8 Closure Across Layers Layer 2 description. - Glossary prism_model! entry. - Section 4 ADR-034 strategic decision row prose. Together these complete the wiki's consistency for ADR-034. The remaining "G1-G19" references all live within ADR-026 itself (describing its own G12-G19 additions, before ADR-033 added G20) and within the ADR-033 strategic decision row's historical description of the "prior G1-G19 baseline that ADR-033 extends" — those are accurate at their authoring point. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 10, 2026
  • ADR-032 + ADR-033: closure-body grammar completion + substrate-amendment reconciliation ADR-032 — CYCLE_SIZE associated const on ConstrainedTypeShape: compile-time domain-cardinality introspection for first_admit. The SDK proc-macro reads <DomainTy as ConstrainedTypeShape>::CYCLE_SIZE at expansion time and emits it as the Term::Literal measure value at the smallest WittLevel that fits. Per- shape-kind derivation: Witt-level = 2^N; product = saturating-product of factors; coproduct = saturating-sum + 1; output_shape! declared explicitly; ConstrainedTypeInput = 1. Replaces SDK 0.3.6's Literal(256) placeholder. ADR-033 — Field-access projection (G20) for product-of-shapes inputs and the eleventh Term variant Term::ProjectField. Closure-body grammar admits <expr>.<name> and <expr>.<index> on partition_product receivers, including chained access. partition_product! macro extends with a named-tuple form (prefix: Prefix76, target: Target32) emitting a FIELDS const that the proc- macro reads at expansion time. Term::ProjectField fold-rule slices [byte_offset .. byte_offset+byte_length] from the source's evaluation result; bounds verified at proc-macro expansion. Coproduct destructuring remains through Term::Match (G6); product destructuring is structurally a different operation. Substrate-amendment reconciliation in existing ADRs (the wiki catching up to foundation 0.3.5's TR-08 substrate amendments — Le/Lt/Ge/Gt/Concat — that landed before this commit): - ADR-022 D3 G3 split into G3a (identifier-call form) + G3b (binary-operator form +, -, *, ^, &, |, <=, <, >=, >). Both lower to identical Term::Application emissions. - ADR-026 G12 disambiguation references the canonical PrimitiveOp catalog at any committed snapshot rather than enumerating ten fixed identifiers. Reserved-identifier set extends with recurse, unfold, lift, project from named grammar forms, plus ADR-030's axis-trait method identifiers in the resolution order (step 3 of 5). - ADR-026 G16 extended with ADR-032 cross-reference: structural correctness vs runtime-strategy override now distinct. - ADR-029 Term::Application fold-rule references the snapshot catalog; Le/Lt/Ge/Gt return Literal(1)/Literal(0); Concat result length is sum of operand lengths bounded by TERM_VALUE_MAX_BYTES. - ADR-029 Term-enum-count statement: tenth (AxisInvocation) + eleventh (ProjectField). - ADR-013 Consequences extended with ADR-032 + ADR-033 + the PrimitiveOp catalog extension under the same TR-08 mechanism. - ADR-022 D4 trait body annotated with CYCLE_SIZE cross-reference. Section 4 strategic-decision row added; Section 8 Closure Across Layers paragraph extended. Glossary: 3 new entries (CYCLE_SIZE, G20, Term::ProjectField); PrimitiveOp entry updated to reference the snapshot catalog rather than enumerating fixed-by-architecture. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 9, 2026
  • ADR-030 + ADR-031 + Scientific Methods and Results partition (issue #7) ADR-030 — `axis!` SDK macro: universal substrate-extension declaration mechanism. Generalizes Hasher to one axis among many; introduces AxisExtension, AxisTuple, MAX_AXIS_TUPLE_ARITY, AXIS_OUTPUT_BYTES_CEILING. Term::HasherProjection becomes Term::AxisInvocation (Term enum stays at ten variants). Catamorphism's per-variant fold-rule dispatches through A::dispatch(axis_index, kernel_id, &evaluated_inputs). ADR-031 — `prism` is the standard library: a façade over `uor-foundation` plus standard-library Layer-3 sub-crates (prism-numerics, prism-crypto, prism-tensor, prism-fhe) published from the Prism repository. Inclusion criteria, layered consumption pattern, content-addressing invariance, conformance vetting. Repository allocation made explicit: architecture in UOR-Framework, standard library in Prism. Drops prior "blessed" framing. Cross-references: ADR-007 third-position generalization to A: AxisTuple; ADR-013 Consequences extension; ADR-024 axis-declaration mechanism alongside verb-declaration; ADR-026 G19 generalization to axis-trait method calls; ADR-029 fold-rule + Substitution-axis-realized verbs paragraph rewritten for AxisTuple. Section 4 strategic decision row + composition paragraph; Section 8 Closure Across Layers paragraph. Glossary: 11 new/updated entries (axis!, AxisExtension, AxisTuple, AXIS_ADDRESS, AXIS_OUTPUT_BYTES_CEILING, MAX_AXIS_TUPLE_ARITY, Standard-library Layer-3 sub-crate, prism-numerics, prism-crypto, prism-tensor, prism-fhe, Term::AxisInvocation rename). Scientific Methods and Results — new non-normative academic-documentation partition (Director of Research stream per issue #7). Adds src/scientific/Scientific-Methods-and-Results.adoc with the landing-page content; scripts/render-scientific.sh (Stage 3d, mirrors render-15288.sh with reconciliation-gap discipline); Stage 3d wiring in build.sh + Stage 4g in stage-to-wiki-root.sh + render_scientific in validate.sh; sidebar entry in src/wiki/Sidebar.adoc; partition specification in wiki-definition.md (Non-normative content partitions section, source-tree, build-stages, correctness-contract Property 7); SOP in README.md (Robin issues -> wiki maintainer authors AsciiDoc -> build.sh -> commit). Pin refresh: tools/versions.txt java line refreshed from openjdk 21.0.11 to openjdk 21.0.10 (the JDK 21.x actually installed on the build host). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 9, 2026
  • Three-layer algebraic closure + operator-set commitments (ADR-024..029) Architectural completion of the three-layer closure spine (substrate / prism / implementation), the operator sets at each layer, and the catamorphism evaluation semantics that realize "structural inference" at the value level. Six new ADRs: - ADR-024 Three-layer algebraic closure spine. Carriers, operator sets, closure checks, and the verb!/use_verbs! mechanism for the implementation layer. - ADR-025 Substrate operator set: composers (×, +) + endomorphism family after_op for op ∈ Γ = {+, −, ×, ÷, ^}, with the Categorical X regime split (T=3, O=8). Catalog and transition tables as conformance test vectors. - ADR-026 Prism operator set (seven operators: compose, parallel_compose, fold_n, tree_fold, first_admit, partition_product, partition_coproduct) + closure-body grammar extension G12-G19 (with G19 as the substitution-axis-realized verb form). Three-way runtime split for first_admit. - ADR-027 Custom Output shape declaration via output_shape! SDK macro. Closes the GroundedShape seal as a degenerate-typed-iso blocker. - ADR-028 Grounded<T> carries the output-value payload. ROUTE_OUTPUT_ BUFFER_BYTES ceiling; output_bytes() accessor. Brings application- level side-channel outputs into the foundation-sealed primary surface. - ADR-029 Catamorphism evaluation semantics. Per-variant fold-rules (Literal, Variable, Application, Lift, Project, Match, Recurse, Unfold, Try, HasherProjection) make pipeline::run a real catamorphism. ADR-023 reconciled with stable-Rust form (foundation-fixed ROUTE_INPUT_BUFFER_BYTES ceiling instead of per-impl-sized buffer requiring generic_const_exprs). Rejected alternative 4 documents the reasoning, paralleling ADR-022 D2's slice-form decision. Cross-references threaded through ADR-007, ADR-013, ADR-019, ADR-020, ADR-022 D4. Section 4 strategic decision row + section 8 "Closure Across Layers" subsection. Glossary additions: 18 new entries spanning the three closures (across / prism / substrate), the seven prism operators (compose, parallel_compose, fold_n, tree_fold, first_admit, partition_product, partition_coproduct), the substrate operator framework (Categorical X regime split, gauge-tower operator set Γ), the SDK macros (output_shape!, verb!, use_verbs!), the new Grounded accessor (output_bytes), the new ceiling (ROUTE_OUTPUT_BUFFER_BYTES), the new Term variant (Term::HasherProjection), the new threshold (FOLD_UNROLL_THRESHOLD), and the new coined term Verb. All 8 validators PASS, Stage 5 idempotence holds (0 divergent files), all tests PASS. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 7, 2026
  • ADR-022 reconciliation + ADR-023: input-binding gap closure ADR-022 + ADR-020 needed reconciliation with the 0fd0c03 implementation surface, plus closure of one architectural gap (M::Input → CompileUnit binding-table mechanism, ADR-023). ADR-022 updates: - D2: arena slice form is canonical (TermArena<CAP> on the trait would require generic_const_exprs); from_slice constructor remains for callers that explicitly want the arena form. - D3: closure-body grammar specified per-form (G1-G11) — replaces prior phrase-level enumeration. Adds Lift/Project (G4/G5, previously omitted), reserves name_index sentinels (0 for `input`, u32::MAX for wildcard / default-propagation). - D4: ConstrainedTypeShape is unparameterized; Input adds IntoBindingValue bound, Output adds GroundedShape bound. - D5: forward body calls run_route which reads arena_slice() (D6) and serializes the input via IntoBindingValue (ADR-023). - D6 (new): FoundationClosed trait surface — one method, fn arena_slice() -> &'static [Term]. ADR-020 updates: Decision text declares the FoundationClosed trait surface and removes "(proposed name)" hedges (both names ratified in source). Section 8 PrismModel Contract trait body matches the new D4 form. ADR-023 (new): Foundation declares IntoBindingValue trait that any ConstrainedTypeShape used as PrismModel::Input MUST implement. pipeline::run_route serializes the input value into a transient BindingEntry for Term::Variable { name_index: 0 }. Closes the gap that blocked correct catamorphism execution for non-empty inputs. Glossary: FoundationClosed entry reflects new method-on-trait surface; new IntoBindingValue entry; G.3 conditional skipped (no existing TermArena entry). Asciidoctor edge cases fixed: bare `_` inside backticks (parsed as italic emphasis, produced malformed XML pandoc choked on, silently dropping ADR-023 + (D2) Rejected); `prism's seal regime` and `forward's body invokes ... lower-level pipeline::run` (close- backtick + apostrophe-s + later open-backtick mis-parsed). Edits above plus the underscore-passthrough wrap (++`_`++) make the docbook well-formed throughout. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

    @afflom afflom committed May 7, 2026