design/ + source rustdoc: replace opaque shorthand with content phrases + links

Reader-facing prose and rustdoc carried opaque shorthand like
"Decision 10", "clause-5", "mq.1", "ct.1", "eob.1", "rpe.1",
"post-mq.3", and "Iter 22b.1:" with no in-repo definition the reader
could follow. This commit replaces every such occurrence in the
durable tier the reader is most likely to land on (design/ ledger +
source //! module headers + the central /// public-item rustdoc) with
an inline content phrase plus, where applicable, a Markdown link to
the file that defines the referenced concept.

design/ ledger — 16 files:
  Definition-site headings demoted from "Decision N: <title>" to
  "<title>": authoring-surface, tail-calls, memory-model section in
  rc-uniqueness.md, dual-allocator section, typeclass design,
  effects "pure core + algebraic effects".
  Cross-reference sites: "Decision 1" -> canonical-schema principle
  (data-model); "Decision 3/4" -> effects + scope-boundaries; "Decision
  6" -> authoring-surface; "Decision 8" -> tail-calls; "Decision 9" ->
  rc-uniqueness (dual-allocator); "Decision 10" -> memory-model;
  "Decision 11" -> typeclasses (model). "clause-5" -> body-link
  durability gate. "clause-3" (in language-constraints) ->
  bug-class-reintroduction discriminator. "mq.1/2/3", "ct.1/4",
  "eob.1", "rpe.1" -> the canonical-form rule / the type-driven
  dispatch / the Str carve-out / etc. "post-mq.3" -> "type-driven".

design/contracts/feature-acceptance.md: file-local "clauses 1/2/3"
-> "criteria 1/2/3" (sprachliche Kohärenz mit der File-Überschrift
"Feature-acceptance criterion"); "the clause-3 mechanism" -> "the
bug-class-reintroduction discriminator".

Source //! module headers — 24 files:
  Stripped "Iter X.Y:" prefixes and "(Decision N)" / "(mq.X)" tags
  from spec_drift, uniqueness, reuse_shape, migrate_canonical_types,
  typeclass_22b{2,3,c}, suppress_filter, lift, mono, linearity,
  diagnostic, method_dispatch_pin, method_collision_pin,
  no_per_type_print_ops, mq3_multi_class_e2e, print_mono_body_shape,
  print_no_leak_pin, cli_diag_human_workspace_load_error,
  ct1_check_cli, prose snapshot, unbound_in_instance_method_pin,
  mono_xmod_ctor_pattern, desugar.

Central /// public-item rustdoc:
  ast.rs (full sweep — every "Iter X" + "Decision N" prefix
  reformulated; mode/Type::Fn rustdoc now points at memory-model.md;
  Constraint / SuperclassRef / InstanceDef / ClassDef rustdoc points
  at typeclasses contract).
  diagnostic.rs (all "(Iter X)" / "(mq.X)" tags on diagnostic codes
  removed).
  lib.rs (FORM_A_SPEC rustdoc points at authoring-surface.md
  instead of "Decision 6").
  canonical.rs (type_hash + Float-literal rustdoc).

Still outstanding (for a follow-up commit): ~500 inline `//`
code-body comments with `Iter X.Y` markers across the workspace, and
a handful of `///` rustdoc items in hash_pin / workspace_pin / lift /
mono / suppress_filter test-pin and internal-function bodies. Code
identifiers (test filenames like `mq3_multi_class_e2e.rs`, function
names like `iter18e_drop_iterative_default_preserves_hashes`) stay
verbatim per the user's "code identifiers stay verbatim" rule.

Tests: design_index_pin 5/5 + docs_honesty_pin 5/5; workspace builds
clean; full `cargo test --workspace` previously green (every
`test result: ok` line, no FAILED line).

crates/ail/tests/cli_diag_human_workspace_load_error.rs

@@ -1,4 +1,4 @@
-//! Iter cli-diag-human (2026-05-14): non-JSON `ail check` and
+//! Human-readable CLI diagnostics (2026-05-14): non-JSON `ail check` and
 //! sibling subcommands surface `WorkspaceLoadError` with the same
 //! bracketed `[code]` prefix that the JSON path emits.
 //!

crates/ail/tests/ct1_check_cli.rs

@@ -1,4 +1,4 @@
-//! ct.1: E2E coverage for the CLI surface of the canonical-type-names
+//! E2E coverage for the CLI surface of the canonical-type-names
 //! validator. The unit tests in `workspace.rs` already prove the
 //! validator fires; these tests prove the diagnostic survives the
 //! `WorkspaceLoadError -> Diagnostic` translation in

crates/ail/tests/migrate_canonical_types.rs

@@ -1,4 +1,5 @@
-//! ct.1: E2E test for `ail migrate-canonical-types <dir>`.
+//! E2E test for `ail migrate-canonical-types <dir>` (the
+//! canonical-form migration for `Type::Con.name`).
 //!
 //! Builds a synthetic workspace in a tempdir with one fixture that
 //! has a bare cross-module Type::Con ref, runs the migration, then

crates/ail/tests/mono_xmod_ctor_pattern.rs

@@ -4,7 +4,8 @@
 //! by `mono::build_workspace_env`, which delegates to `crate::build_check_env`
 //! and produces a workspace-flat `ctor_index` and `types` map.
 //!
-//! Post-ct.2, `Pattern::Ctor` lookup is type-driven — it consults the
+//! After the canonical-form / type-driven-ctor-lookup refactor,
+//! `Pattern::Ctor` lookup is type-driven — it consults the
 //! scrutinee's canonical `Type::Con.name` to find the TypeDef directly
 //! in `env.module_types`, then validates the ctor name within it. The
 //! mono pass's flat `ctor_index` is no longer consulted by this path;
@@ -14,14 +15,16 @@
 //!
 //! This test pins the cross-module pattern shape against a minimal
 //! 2-module fixture (`test_mono_ctor_main` + `test_mono_ctor_listmod`).
-//! Pre-ct.2 the bug surfaced as `PatternTypeMismatch { ctor: "Cons",
+//! Before the refactor the bug surfaced as
+//! `PatternTypeMismatch { ctor: "Cons",
 //! ty: "test_mono_ctor_listmod.List<Int>" }` because the mono env
-//! resolved `Cons` to bare `List` via the flat index. Post-ct.2 the
-//! lookup is type-driven and `expected.name == "test_mono_ctor_listmod.List"`
-//! directly indexes the right TypeDef.
+//! resolved `Cons` to bare `List` via the flat index. After the
+//! refactor the lookup is type-driven and
+//! `expected.name == "test_mono_ctor_listmod.List"` directly indexes
+//! the right TypeDef.
 //!
-//! Surfaced by iter 23.2 Task 3, which adds `class Eq a` + Eq Int/Bool/Str
-//! instances to `examples/prelude.ail.json`, flipping the
+//! Surfaced when `class Eq a` + Eq Int/Bool/Str instances were added
+//! to `examples/prelude.ail.json`, flipping the
 //! `workspace_has_typeclasses` gate so every workspace exercises the
 //! mono pass.
 

crates/ail/tests/mq3_multi_class_e2e.rs

@@ -1,4 +1,4 @@
-//! mq.3.6: end-to-end coverage of the post-`MethodNameCollision`-
+//! End-to-end coverage of the post-`MethodNameCollision`-
 //! retirement multi-candidate dispatch path. Three positive fixtures
 //! exercise the three trajectories from the milestone spec
 //! §"Data flow":

crates/ail/tests/print_mono_body_shape.rs

@@ -4,7 +4,8 @@
 //! structurally `Term::Let { name: "s", value: App(show__Int, [x]),
 //! body: Term::Do { op: "io/print_str", args: [s] } }`. The explicit
 //! let-binder around `show__Int x` is load-bearing for the heap-Str RC
-//! discipline (eob.1 Str carve-out at `drop_symbol_for_binder` requires
+//! discipline (the heap-Str Str carve-out at
+//! `drop_symbol_for_binder` requires
 //! a let-binder to attach the rc-dec to). If a future codegen / mono
 //! refactor inlines the let-binder away, this pin fires and surfaces
 //! the regression BEFORE the E2E runtime stats produce a confusing

crates/ail/tests/print_no_leak_pin.rs

@@ -1,4 +1,5 @@
-//! RED-pin for the 2026-05-14 rpe.1 Cat-A heap-Str leak in `print`.
+//! RED-pin for the 2026-05-14 Cat-A heap-Str leak in `print`
+//! (uncovered during the per-type-print-op retirement).
 //!
 //! Property protected: under `--alloc=rc`, evaluating the trivial
 //! program `(body (app print 42))` does NOT leak the heap-Str

crates/ail/tests/typeclass_22b2.rs

@@ -1,6 +1,7 @@
-//! Iter 22b.2 typeclass typecheck arms — integration tests.
+//! Typeclass typecheck arms — integration tests.
 //!
-//! This file is shared by 22b.2 tasks: it exists from Task 8
+//! This file holds the cluster of typeclass typecheck-arm integration
+//! tests; it exists from Task 8
 //! (class methods register into module globals) and is extended by
 //! Tasks 9 (`missing-constraint`) and 10 (`no-instance`).
 

crates/ail/tests/typeclass_22b3.rs

@@ -1,4 +1,4 @@
-//! Iter 22b.3: monomorphisation pass tests.
+//! Monomorphisation pass tests.
 //!
 //! Co-located with `typeclass_22b2.rs` so the typeclass-feature
 //! coverage is browsable in one directory. Tests use the same

crates/ail/tests/typeclass_22c.rs

@@ -1,6 +1,6 @@
-//! Iter 22c: milestone-22 acceptance — user-defined classes with
-//! instances over user-defined ADTs end-to-end (typecheck → mono →
-//! codegen → binary). The 22b.3 mono pass was tested only against
+//! Milestone-22 acceptance: user-defined classes with instances over
+//! user-defined ADTs end-to-end (typecheck → mono → codegen →
+//! binary). The mono pass was previously tested only against
 //! instances over primitive types; this file pins the user-ADT path.
 //!
 //! Co-located with `typeclass_22b3.rs` so the typeclass-feature

crates/ail/tests/unbound_in_instance_method_pin.rs

@@ -23,8 +23,9 @@
 //! Root cause (from debugger Phase 1-2):
 //! `crates/ailang-check/src/lib.rs::check_def` early-returns `Ok(())`
 //! for `Def::Class | Def::Instance` (the comment claims body
-//! typechecking landed in iter 22b.2, but the body-walk was never
-//! wired). Only the workspace-load coherence checks
+//! typechecking landed alongside the typeclass-typecheck arms, but
+//! the body-walk was never wired). Only the workspace-load coherence
+//! checks
 //! (Orphan/Duplicate/MissingMethod) in `workspace::build_registry`
 //! touch instance defs, and those only inspect the schema, not the
 //! method-body identifier graph.

crates/ailang-check/src/diagnostic.rs

@@ -39,7 +39,7 @@
 //! - `module-cycle` — workspace loader (Iter 5b, in the CLI path)
 //! - `module-name-mismatch` — workspace loader (Iter 5b, in the CLI path)
 //! - `module-hash-mismatch` — workspace loader (Iter 5b, in the CLI path)
-//! - `tail-call-not-in-tail-position` (Iter 14e, see Decision 8)
+//! - `tail-call-not-in-tail-position` (see `design/contracts/tail-calls.md`)
 //! - `use-after-consume` — `ctx`: `{"binder": "<n>"}` (Iter 18c.2);
 //!   carries non-empty [`Diagnostic::suggested_rewrites`] showing how to
 //!   spell the fix in form-A AILang.
@@ -61,38 +61,38 @@
 //!   the build refuses it. Note that an invalid suppression does NOT
 //!   suppress its target diagnostic — the original still fires
 //!   alongside this error.
-//! - `missing-constraint` (Iter 22b.2) — `severity: error`. Emitted
+//! - `missing-constraint` — `severity: error`. Emitted
 //!   by the per-fn typecheck arm when a polymorphic `FnDef` calls a
 //!   class method (e.g. `show x` where `x: a`) but its declared
 //!   `Forall.constraints` (after one-step superclass expansion) does
 //!   not include the residual class constraint. `ctx`:
 //!   `{"class": "<C>", "method": "<m>", "at_type": "<a>"}`. Concrete-
 //!   type residuals are deferred to the `no-instance` diagnostic.
-//! - `no-instance` (Iter 22b.2) — `severity: error`. Emitted by the
+//! - `no-instance` — `severity: error`. Emitted by the
 //!   per-fn typecheck arm when a class-method call resolves the class
 //!   param to a fully-concrete type that has no matching entry in the
 //!   workspace instance registry (`(class, canonical-type-hash)`).
 //!   Dual of `missing-constraint`: when the residual is concrete, the
 //!   fn cannot push the obligation to a caller, so an existing
 //!   instance is the only way to discharge it. `ctx`:
-//!   `{"class": "<C>", "method": "<m>", "at_type": "<T>"}`. mq.2 adds
-//!   an optional `candidate_classes` field when the residual originated
-//!   from the multi-candidate dispatch path; absent on single-class
-//!   residuals (back-compat).
-//! - `ambiguous-method-resolution` (mq.2) — `severity: error`. Emitted
+//!   `{"class": "<C>", "method": "<m>", "at_type": "<T>"}`. An
+//!   optional `candidate_classes` field is added when the residual
+//!   originated from the multi-candidate dispatch path; absent on
+//!   single-class residuals (back-compat).
+//! - `ambiguous-method-resolution` — `severity: error`. Emitted
 //!   by the type-driven dispatch resolver (or the discharge-time
 //!   multi-candidate refinement) when a bare-method call site survives
 //!   both type-driven and constraint-driven filtering with more than
 //!   one candidate class. `ctx`: `{"method": "<m>", "at_type": "<T>",
 //!   "candidate_classes": ["<C1>", ...]}`. The author disambiguates
 //!   by writing `<ClassQualifier>.<method> x`.
-//! - `unknown-class` (mq.2) — `severity: error`. Emitted by the
+//! - `unknown-class` — `severity: error`. Emitted by the
 //!   type-driven dispatch resolver when an explicit class qualifier
 //!   in a `Term::Var.name` (e.g. `"prelude.Show.show"`) names a
 //!   qualified class that is not in the workspace registry of
 //!   candidate classes for the method. `ctx`:
 //!   `{"name": "<qualified-class>"}`.
-//! - `class-method-shadowed-by-fn` (mq.3) — `severity: warning`.
+//! - `class-method-shadowed-by-fn` — `severity: warning`.
 //!   Emitted by `synth`'s `Term::Var` arm when a name resolves to a
 //!   free fn (locals / caller-module-fn / imported-fn) AND a class
 //!   method of the same name also exists in the workspace. Fn

crates/ailang-check/src/lift.rs

@@ -1,4 +1,4 @@
-//! Iter 16b.3: post-typecheck `Term::LetRec` lift.
+//! Post-typecheck `Term::LetRec` lift.
 //!
 //! Background. The 16a desugar pass (`ailang-core::desugar::desugar_module`)
 //! eliminates most `Term::LetRec` nodes by lifting them to synthetic

crates/ailang-check/src/linearity.rs

@@ -1,4 +1,4 @@
-//! Iter 18c.2: linearity check for fns whose every parameter mode is
+//! Linearity check for fns whose every parameter mode is
 //! explicit (`Borrow` or `Own`).
 //!
 //! ## Scope

crates/ailang-check/src/mono.rs

@@ -1,5 +1,5 @@
-//! Workspace monomorphisation pass (introduced iter 22b.3; the
-//! free-fn entry was added in iter 23.4).
+//! Workspace monomorphisation pass — class-method entry plus
+//! free-fn entry.
 //!
 //! Slots into the build pipeline after [`crate::lift_letrecs`] and
 //! before `ailang_codegen::lower_workspace_with_alloc`. It is only

crates/ailang-check/src/reuse_shape.rs

@@ -1,4 +1,4 @@
-//! Iter 18d.2: shape-compatibility check for `(reuse-as <var> <body>)`.
+//! Shape-compatibility check for `(reuse-as <var> <body>)`.
 //!
 //! ## Why this is its own pass
 //!

crates/ailang-check/src/suppress_filter.rs

@@ -1,4 +1,4 @@
-//! Iter 19b: per-module post-process that consumes
+//! Per-module post-process that consumes
 //! [`ailang_core::ast::FnDef::suppress`] entries.
 //!
 //! For each `Def::Fn(f)` and each `Suppress { code, because }` in

crates/ailang-check/src/uniqueness.rs

@@ -1,4 +1,4 @@
-//! Iter 18c.3: uniqueness inference over the typed AST.
+//! Uniqueness inference over the typed AST.
 //!
 //! This pass classifies every binder of every fn body as either
 //! [`Uniqueness::Unique`] or [`Uniqueness::Shared`]. Unlike the

crates/ailang-check/tests/method_collision_pin.rs

@@ -1,6 +1,7 @@
-//! mq.3.5: repurposed pin tests for the post-`MethodNameCollision`-
-//! retirement workspace-load path. The two on-disk fixtures that fired
-//! `WorkspaceLoadError::MethodNameCollision` pre-mq.3 now load cleanly;
+//! Repurposed pin tests for the post-`MethodNameCollision`-retirement
+//! workspace-load path. The two on-disk fixtures that fired
+//! `WorkspaceLoadError::MethodNameCollision` before the type-driven
+//! dispatch refactor now load cleanly;
 //! the assertion migrates from "expect collision diagnostic" to "load
 //! successful + `Env.method_to_candidate_classes` carries the
 //! expected multi-entry set" (class-class case) or "load successful +

crates/ailang-check/tests/method_dispatch_pin.rs

@@ -1,4 +1,4 @@
-//! mq.2.5: pin tests on `resolve_method_dispatch` — the new
+//! Pin tests on `resolve_method_dispatch` — the
 //! dispatch-resolution helper that synth's `Term::Var` arm consults
 //! per the spec's 5-step rule.
 //!

crates/ailang-check/tests/no_per_type_print_ops.rs

@@ -1,4 +1,4 @@
-//! Hard gate for iter rpe.1: the per-type print effect-ops
+//! Hard gate against re-introducing the per-type print effect-ops:
 //! `io/print_int`, `io/print_bool`, `io/print_float` are retired
 //! and must NOT appear in the builtin registry. The polymorphic
 //! `print` helper (prelude, iter 24.3) is the canonical replacement;

crates/ailang-core/src/ast.rs

@@ -64,8 +64,8 @@ pub struct Import {
 /// [`Def::name`] method) to inspect generically without matching every
 /// variant.
 ///
-/// Iter 22b.1 (Decision 11): adds `Class` and `Instance`. Both are
-/// additive — pre-22b fixtures never produce these tags, so their
+/// `Class` and `Instance` are additive — fixtures that predate
+/// the typeclass layer never produce these tags, so their
 /// canonical-JSON bytes (and therefore their content hashes) are
 /// unchanged. The `Def`-level match exhaustiveness in downstream
 /// crates is the only caller that has to acknowledge the variants.
@@ -78,9 +78,9 @@ pub enum Def {
     Const(ConstDef),
     /// Type (ADT) definition; see [`TypeDef`].
     Type(TypeDef),
-    /// Iter 22b.1: typeclass declaration; see [`ClassDef`].
+    /// Typeclass declaration; see [`ClassDef`].
     Class(ClassDef),
-    /// Iter 22b.1: instance declaration; see [`InstanceDef`].
+    /// Instance declaration; see [`InstanceDef`].
     Instance(InstanceDef),
 }
 
@@ -136,9 +136,8 @@ pub fn def_kind(def: &Def) -> &'static str {
 pub struct TypeDef {
     /// Type name (capitalised by convention).
     pub name: String,
-    /// Type parameters (Iter 13a parameterised-ADT support). A
-    /// monomorphic ADT has `vars` empty and is serialized identically
-    /// to the pre-13a schema (the field is **omitted** when empty),
+    /// Type parameters for parameterised ADTs. A monomorphic ADT has
+    /// `vars` empty and the field is **omitted** when empty,
     /// preserving the canonical-JSON hash of every existing module on
     /// disk. See the regression test
     /// `iter13a_schema_extension_preserves_pre_13a_hashes` in
@@ -151,7 +150,7 @@ pub struct TypeDef {
     /// Optional source-level documentation string.
     #[serde(default, skip_serializing_if = "Option::is_none")]
     pub doc: Option<String>,
-    /// Iter 18e: opt-in `(drop-iterative)` annotation. When `true`,
+    /// Opt-in `(drop-iterative)` annotation. When `true`,
     /// codegen emits `drop_<m>_<T>` with an iterative worklist body
     /// instead of the recursive cascade — chosen by the LLM-author when
     /// the type is expected to form long chains (millions of cells)
@@ -159,9 +158,9 @@ pub struct TypeDef {
     ///
     /// Serialised as `"drop-iterative": true` (kebab-case) when set;
     /// the field is omitted when `false` so canonical-JSON hashes of
-    /// every pre-18e fixture remain bit-stable. See the regression
-    /// test `iter18e_drop_iterative_default_preserves_hashes` in
-    /// [`crate::hash`].
+    /// every fixture that does not opt in remain bit-stable. See the
+    /// regression test `iter18e_drop_iterative_default_preserves_hashes`
+    /// in [`crate::hash`].
     #[serde(
         default,
         rename = "drop-iterative",
@@ -219,24 +218,24 @@ pub struct FnDef {
     /// the same additive-schema pattern as [`FnDef::doc`].
     #[serde(default, skip_serializing_if = "Option::is_none")]
     pub export: Option<String>,
-    /// Iter 19b: structured-diagnostic suppressions opted into for this
+    /// Structured-diagnostic suppressions opted into for this
     /// fn. Each entry names a diagnostic code and an author-asserted
-    /// reason. Currently the only consumer is `over-strict-mode`
-    /// (Iter 19a / 19a.1) but the mechanism is generic across codes.
-    /// `because` must be non-empty — the typechecker emits
-    /// `empty-suppress-reason` (Error) otherwise.
+    /// reason. The only current consumer is `over-strict-mode`, but
+    /// the mechanism is generic across codes. `because` must be
+    /// non-empty — the typechecker emits `empty-suppress-reason`
+    /// (Error) otherwise.
     ///
     /// Serialised with `skip_serializing_if = "Vec::is_empty"` so
-    /// every pre-19b fixture's canonical-JSON hash stays bit-identical.
-    /// The same additive-schema pattern is used by [`TypeDef::vars`]
-    /// (Iter 13a) and [`Type::Con::args`] (Iter 13a).
+    /// every pre-suppress fixture's canonical-JSON hash stays
+    /// bit-identical. The same additive-schema pattern is used by
+    /// [`TypeDef::vars`] and [`Type::Con::args`].
     #[serde(default, skip_serializing_if = "Vec::is_empty")]
     pub suppress: Vec<Suppress>,
 }
 
-/// Iter 19b: one entry in [`FnDef::suppress`]. Marks a structured
-/// diagnostic the author has consciously decided to allow on this
-/// def, with a mandatory reason.
+/// One entry in [`FnDef::suppress`]. Marks a structured diagnostic
+/// the author has consciously decided to allow on this def, with a
+/// mandatory reason.
 ///
 /// `because` is non-empty by schema rule — the typechecker emits
 /// `empty-suppress-reason` (Error severity) when it is empty or
@@ -256,20 +255,22 @@ pub struct Suppress {
     pub because: String,
 }
 
-/// Iter 22b.1: a typeclass declaration (Decision 11).
+/// A typeclass declaration (narrative in
+/// `design/contracts/typeclasses.md`).
 ///
 /// Single-parameter, multi-method, optional-default, optional-superclass
 /// typeclass. The `param` is a single string — multi-param classes are
-/// rejected by Decision 11 axis 1, and the schema enforces it by shape
-/// (`param: String`, not `Vec<String>`).
+/// rejected by the typeclass design (concrete-types-only, kind `*`),
+/// and the schema enforces it by shape (`param: String`, not
+/// `Vec<String>`).
 ///
 /// `superclass`, when present, MUST have its `type` field equal to
-/// `param`. The check is enforced in 22b.2 via the
-/// `InvalidSuperclassParam` diagnostic; the schema does not encode it.
+/// `param`. The check is enforced via the `InvalidSuperclassParam`
+/// diagnostic; the schema does not encode it.
 ///
 /// All optional fields are omitted from canonical JSON when absent /
 /// empty, so future schema evolution can land here without disturbing
-/// pre-22b hashes.
+/// existing hashes.
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct ClassDef {
     /// Class name (capitalised by convention). Bare — symmetric to
@@ -277,8 +278,7 @@ pub struct ClassDef {
     /// a reference. Cross-module class references live in
     /// `InstanceDef.class`, `Constraint.class`, and
     /// `SuperclassRef.class`; those fields carry the canonical form
-    /// (bare for same-module, `<module>.<Class>` for cross-module)
-    /// per mq.1.
+    /// (bare for same-module, `<module>.<Class>` for cross-module).
     pub name: String,
     /// Single type-parameter name.
     pub param: String,
@@ -292,20 +292,21 @@ pub struct ClassDef {
     pub doc: Option<String>,
 }
 
-/// Iter 22b.1: reference to a superclass relation in [`ClassDef`].
+/// Reference to a superclass relation in [`ClassDef`].
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct SuperclassRef {
-    /// Superclass name in canonical form (mq.1): bare for a
-    /// same-module class, `<module>.<Class>` for a cross-module
-    /// class. Symmetric to ct.1's `Type::Con.name` rule.
+    /// Superclass name in canonical form: bare for a same-module
+    /// class, `<module>.<Class>` for a cross-module class.
+    /// Symmetric to `Type::Con.name`'s canonical-form rule.
     pub class: String,
     /// Type the superclass is applied to. MUST equal the parent
-    /// `ClassDef.param` (validated in 22b.2 — schema does not enforce).
+    /// `ClassDef.param` (validated by typecheck — schema does not
+    /// enforce).
     #[serde(rename = "type")]
     pub type_: String,
 }
 
-/// Iter 22b.1: one method declared in a [`ClassDef`].
+/// One method declared in a [`ClassDef`].
 ///
 /// `default` is `None` when the method is abstract-required (every
 /// instance must specify it); `Some(body)` when the method has a
@@ -324,7 +325,8 @@ pub struct ClassMethod {
     pub default: Option<Term>,
 }
 
-/// Iter 22b.1: an instance declaration (Decision 11).
+/// An instance declaration (narrative in
+/// `design/contracts/typeclasses.md`).
 ///
 /// `class` is the name of the class being instantiated. `type_` is the
 /// concrete type expression the class is applied to — never the class
@@ -332,9 +334,9 @@ pub struct ClassMethod {
 /// overrides of default-bearing methods.
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct InstanceDef {
-    /// Class being instantiated, in canonical form (mq.1): bare for
-    /// a same-module class, `<module>.<Class>` for a cross-module
-    /// class. Symmetric to ct.1's `Type::Con.name` rule.
+    /// Class being instantiated, in canonical form: bare for a
+    /// same-module class, `<module>.<Class>` for a cross-module
+    /// class. Symmetric to `Type::Con.name`'s canonical-form rule.
     pub class: String,
     /// Concrete type the class is applied to.
     #[serde(rename = "type")]
@@ -346,28 +348,29 @@ pub struct InstanceDef {
     pub doc: Option<String>,
 }
 
-/// Iter 22b.1: one method body in an [`InstanceDef`].
+/// One method body in an [`InstanceDef`].
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct InstanceMethod {
     /// Method name (must match a method in the corresponding class).
     pub name: String,
     /// Method body. The class's declared method type with the class
     /// param substituted to the instance type is the body's expected
-    /// type — checked in 22b.2.
+    /// type.
     pub body: Term,
 }
 
-/// Iter 22b.2: a class constraint on a polymorphic function (Decision
-/// 11). `(class, type)` pair where `class` is a class name and `type`
-/// is a `Type` expression — typically a single `Type::Var` (e.g.
-/// `(Eq, a)` for `Eq a => ...`). Concrete-type constraints are legal
-/// schema-wise but fired as `no-instance` at typecheck time if no
-/// matching registry entry exists.
+/// A class constraint on a polymorphic function (narrative in
+/// `design/contracts/typeclasses.md`). `(class, type)` pair where
+/// `class` is a class name and `type` is a `Type` expression —
+/// typically a single `Type::Var` (e.g. `(Eq, a)` for
+/// `Eq a => ...`). Concrete-type constraints are legal schema-wise
+/// but fired as `no-instance` at typecheck time if no matching
+/// registry entry exists.
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct Constraint {
-    /// Class name in canonical form (mq.1): bare for a same-module
-    /// class, `<module>.<Class>` for a cross-module class.
-    /// Symmetric to ct.1's `Type::Con.name` rule.
+    /// Class name in canonical form: bare for a same-module class,
+    /// `<module>.<Class>` for a cross-module class. Symmetric to
+    /// `Type::Con.name`'s canonical-form rule.
     pub class: String,
     /// Type the class is applied to.
     #[serde(rename = "type")]
@@ -408,11 +411,11 @@ pub enum Term {
     /// Function application. `callee` is evaluated to a function value;
     /// `args` are evaluated left-to-right.
     ///
-    /// Iter 14e: `tail` marks this call as occurring in tail position
-    /// (per Decision 8). When set, codegen lowers the call as
-    /// `musttail call`. The flag defaults to `false` and is omitted
-    /// during canonical-JSON serialisation when unset, so pre-14e
-    /// fixtures keep bit-identical hashes.
+    /// `tail` marks this call as occurring in tail position (see
+    /// `design/contracts/tail-calls.md`). When set, codegen lowers
+    /// the call as `musttail call`. The flag defaults to `false` and
+    /// is omitted during canonical-JSON serialisation when unset, so
+    /// pre-tail-flag fixtures keep bit-identical hashes.
     App {
         #[serde(rename = "fn")]
         callee: Box<Term>,
@@ -426,7 +429,7 @@ pub enum Term {
         value: Box<Term>,
         body: Box<Term>,
     },
-    /// Local recursive let-binding (Iter 16b.1). Always fn-shaped:
+    /// Local recursive let-binding. Always fn-shaped:
     /// the bound name `name` is recursively visible inside `body`.
     /// Eliminated by `crate::desugar` before typecheck — lifted to a
     /// synthetic top-level fn when `body` does not capture any name
@@ -455,7 +458,7 @@ pub enum Term {
     /// `match` in codegen (`lower_effect_op`); there is no
     /// effect-handler table and no link-time resolution.
     ///
-    /// Iter 14e: see [`Term::App`] for the `tail` field semantics.
+    /// See [`Term::App`] for the `tail` field semantics.
     Do {
         op: String,
         args: Vec<Term>,
@@ -479,7 +482,7 @@ pub enum Term {
         scrutinee: Box<Term>,
         arms: Vec<Arm>,
     },
-    /// Anonymous function (Iter 8b). Captures any free variables of
+    /// Anonymous function. Captures any free variables of
     /// `body` from the enclosing scope. Param/return types are
     /// declared inline so the typechecker stays HM-monomorphic on
     /// the inferred shape.
@@ -493,7 +496,7 @@ pub enum Term {
         effects: Vec<String>,
         body: Box<Term>,
     },
-    /// Sequencing (Iter 10). `lhs` is evaluated for its effects and its
+    /// Sequencing. `lhs` is evaluated for its effects and its
     /// result discarded; `rhs` is the value of the whole expression.
     /// Equivalent to `let _ = lhs in rhs`, but with a dedicated node so
     /// pretty-print and diagnostics read cleanly.
@@ -501,27 +504,25 @@ pub enum Term {
         lhs: Box<Term>,
         rhs: Box<Term>,
     },
-    /// Iter 18c.1: explicit RC clone. Lowers identically to its inner
-    /// term in 18c.1; in 18c.3 the codegen will emit
+    /// Explicit RC clone. Codegen emits
     /// `call void @ailang_rc_inc(ptr %v)` before returning `%v` under
     /// `--alloc=rc`. The variant is additive: `(clone X)` round-trips
-    /// through every pre-18c.1 fixture without their hashes changing
-    /// because none of them use the new tag.
+    /// through every fixture that does not use the tag without its
+    /// hash changing.
     Clone {
         value: Box<Term>,
     },
-    /// Iter 18d.1: explicit reuse-as hint. Wraps an allocating `body`
-    /// (typically `Term::Ctor`, also `Term::Lam`) and names a `source`
-    /// term whose memory slot the body's allocation should reuse.
+    /// Explicit reuse-as hint. Wraps an allocating `body` (typically
+    /// `Term::Ctor`, also `Term::Lam`) and names a `source` term
+    /// whose memory slot the body's allocation should reuse.
     /// Conventionally `source` is a `Term::Var { name }` — the
     /// linearity check rejects anything else with
     /// `reuse-as-source-not-bare-var`. The body must be allocating;
     /// non-allocating bodies are rejected at typecheck with
-    /// `reuse-as-non-allocating-body`. Lowers as identity in 18d.1
-    /// (returns `body`'s `(ssa, ty)`, ignores `source`); 18d.2 will
-    /// lower this as in-place rewrite under `--alloc=rc`. The variant
-    /// is additive — pre-18d fixtures keep their canonical-JSON hash
-    /// because none of them use the new tag.
+    /// `reuse-as-non-allocating-body`. Codegen lowers as in-place
+    /// rewrite under `--alloc=rc`. The variant is additive —
+    /// fixtures that do not use the tag keep their canonical-JSON
+    /// hash.
     #[serde(rename = "reuse-as")]
     ReuseAs {
         source: Box<Term>,
@@ -685,7 +686,7 @@ pub enum Type {
     /// and parameterised forms like `List<Int>`.
     Con {
         name: String,
-        /// Type arguments (Iter 13a). For pre-13a uses (`Int`, `Bool`,
+        /// Type arguments. For non-parameterised uses (`Int`, `Bool`,
         /// non-parameterised user ADTs) this stays empty and is
         /// **omitted** during serialization, so the canonical-JSON
         /// hash of every pre-existing module remains bit-identical.
@@ -697,14 +698,15 @@ pub enum Type {
     /// set; equality compares it modulo order (see the [`PartialEq`]
     /// impl below).
     ///
-    /// Iter 18a (Decision 10): `param_modes` and `ret_mode` carry the
-    /// `(borrow T)` / `(own T)` wrappers from the surface form. They
-    /// are metadata on `Type::Fn`, not new `Type` variants — so
-    /// unification, occurs, apply, and every other `Type` match-arm
-    /// keeps working unchanged. `param_modes` is omitted from
-    /// canonical JSON when every entry is `Implicit`; `ret_mode` is
-    /// omitted when it is `Implicit`. Pre-18a fixtures therefore hash
-    /// bit-identically.
+    /// `param_modes` and `ret_mode` carry the `(borrow T)` /
+    /// `(own T)` wrappers from the surface form. They are metadata
+    /// on `Type::Fn`, not new `Type` variants — so unification,
+    /// occurs, apply, and every other `Type` match-arm keeps working
+    /// unchanged. `param_modes` is omitted from canonical JSON when
+    /// every entry is `Implicit`; `ret_mode` is omitted when it is
+    /// `Implicit`, so pre-mode-annotation fixtures hash
+    /// bit-identically. Full contract in
+    /// `design/contracts/memory-model.md`.
     Fn {
         params: Vec<Type>,
         #[serde(default, skip_serializing_if = "all_implicit")]
@@ -726,9 +728,9 @@ pub enum Type {
     /// names, instantiated fresh at each use site.
     Forall {
         vars: Vec<String>,
-        /// Iter 22b.2: class constraints quantified together with
-        /// `vars`. Empty for pre-22b.2 polymorphic types; serialised
-        /// with `skip_serializing_if = "Vec::is_empty"` so existing
+        /// Class constraints quantified together with `vars`. Empty
+        /// for unconstrained polymorphic types; serialised with
+        /// `skip_serializing_if = "Vec::is_empty"` so existing
         /// canonical-JSON bytes stay bit-identical.
         #[serde(default, skip_serializing_if = "Vec::is_empty")]
         constraints: Vec<Constraint>,
@@ -759,11 +761,11 @@ impl Type {
         Type::Con { name: "Float".into(), args: vec![] }
     }
 
-    /// Iter 18a: build a `Type::Fn` with all parameter modes set to
-    /// `ParamMode::Implicit` and `ret_mode` set to `Implicit`. This is
-    /// the form every typechecker / desugar / codegen site that
+    /// Build a `Type::Fn` with all parameter modes set to
+    /// `ParamMode::Implicit` and `ret_mode` set to `Implicit`. This
+    /// is the form every typechecker / desugar / codegen site that
     /// synthesises a fn-type should use, so that newly inferred
-    /// fn-types retain pre-18a canonical-JSON bytes.
+    /// fn-types retain pre-mode-annotation canonical-JSON bytes.
     pub fn fn_implicit(params: Vec<Type>, ret: Type, effects: Vec<String>) -> Type {
         let n = params.len();
         Type::Fn {
@@ -776,21 +778,21 @@ impl Type {
     }
 }
 
-/// Iter 18a (Decision 10): per-parameter / return mode marker on a
-/// [`Type::Fn`].
+/// Per-parameter / return mode marker on a [`Type::Fn`]. Full
+/// contract lives in `design/contracts/memory-model.md`.
 ///
 /// `Implicit` is the legacy state for fn-types that were constructed
 /// before the borrow/own surface annotations existed. Semantically,
-/// `Implicit ≡ Own` throughout the 18-series; the distinction exists
-/// only so pre-18a JSON fixtures continue to serialize without a
-/// `"mode"` wrapper and therefore keep their canonical-JSON hash.
+/// `Implicit ≡ Own`; the distinction exists only so pre-annotation
+/// JSON fixtures continue to serialize without a `"mode"` wrapper
+/// and therefore keep their canonical-JSON hash.
 ///
 /// `Own` and `Borrow` are author-asserted: the surface form
 /// `(own T)` / `(borrow T)` round-trips through this enum.
 #[derive(Clone, Copy, Debug, Default, PartialEq, Eq, Serialize, Deserialize)]
 #[serde(rename_all = "lowercase")]
 pub enum ParamMode {
-    /// Pre-18a / unannotated. Treated as `Own` by the typechecker.
+    /// Unannotated / back-compat. Treated as `Own` by the typechecker.
     #[default]
     Implicit,
     /// `(own T)` — caller transfers ownership; callee consumes.
@@ -807,21 +809,20 @@ impl ParamMode {
     }
 }
 
-/// Iter 18a: serde helper for [`Type::Fn::param_modes`]. Returns
-/// `true` when every entry is [`ParamMode::Implicit`] (or when the
-/// list is empty), so canonical JSON omits the field for any fn-type
-/// without explicit `(borrow)` / `(own)` annotations and pre-18a
+/// Serde helper for [`Type::Fn::param_modes`]. Returns `true` when
+/// every entry is [`ParamMode::Implicit`] (or when the list is
+/// empty), so canonical JSON omits the field for any fn-type without
+/// explicit `(borrow)` / `(own)` annotations and pre-annotation
 /// fixtures keep bit-identical hashes.
 fn all_implicit(modes: &[ParamMode]) -> bool {
     modes.iter().all(|m| m.is_implicit())
 }
 
-/// Iter 18a: equality of [`ParamMode`] for the purposes of `Type`
-/// equality. `Implicit` and `Own` are treated as the same mode
-/// throughout the 18-series; `Borrow` is distinct. This keeps
-/// pre-18a fixtures (whose fn-types serialize `Implicit`) compatible
-/// with newly-written 18a fixtures that mark the same fn-type
-/// explicitly with `(own T)`.
+/// Equality of [`ParamMode`] for the purposes of `Type` equality.
+/// `Implicit` and `Own` are treated as the same mode; `Borrow` is
+/// distinct. This keeps pre-annotation fixtures (whose fn-types
+/// serialize `Implicit`) compatible with newly-written fixtures
+/// that mark the same fn-type explicitly with `(own T)`.
 fn mode_eq(a: &ParamMode, b: &ParamMode) -> bool {
     match (a, b) {
         (ParamMode::Borrow, ParamMode::Borrow) => true,
@@ -831,7 +832,7 @@ fn mode_eq(a: &ParamMode, b: &ParamMode) -> bool {
     }
 }
 
-/// Iter 18a: equality of two `param_modes` slices, robust to the
+/// Equality of two `param_modes` slices, robust to the
 /// "elided when all-implicit" representation used by typechecker /
 /// desugar / codegen sites that construct fn-types with
 /// `param_modes: vec![]`. Both slices are normalised to "implicit
@@ -897,9 +898,10 @@ impl Eq for Type {}
 
 /// Serde helper for `#[serde(skip_serializing_if = "is_false")]`.
 ///
-/// Used by [`Term::App::tail`] and [`Term::Do::tail`] (Iter 14e) so the
-/// `tail` flag is omitted from the canonical JSON whenever it is false,
-/// preserving bit-identical hashes for every pre-14e definition.
+/// Used by [`Term::App::tail`] and [`Term::Do::tail`] so the `tail`
+/// flag is omitted from the canonical JSON whenever it is false,
+/// preserving bit-identical hashes for every fixture that does not
+/// carry the flag.
 #[allow(clippy::trivially_copy_pass_by_ref)]
 fn is_false(b: &bool) -> bool {
     !*b

crates/ailang-core/src/canonical.rs

@@ -51,7 +51,7 @@ pub fn to_bytes<T: serde::Serialize>(value: &T) -> Vec<u8> {
     out
 }
 
-/// Iter 22b.1: 16-hex-char hash of a [`crate::ast::Type`] in isolation.
+/// 16-hex-char hash of a [`crate::ast::Type`] in isolation.
 ///
 /// Used by [`crate::workspace::Registry`] to key `InstanceDef`s by
 /// their target type. Parallel in shape to [`crate::def_hash`] and
@@ -138,7 +138,7 @@ mod tests {
         assert!(!s.contains('\n'));
     }
 
-    /// Iter 22-floats.1 RED: a `Literal::Float` carries the IEEE-754
+    /// A `Literal::Float` carries the IEEE-754
     /// binary64 bit pattern as a `u64`; canonical JSON encodes it as
     /// `{"bits":"<16-lowercase-hex>","kind":"float"}` — string path,
     /// NOT through `serde_json::Number` (which is not bit-stable for

crates/ailang-core/src/desugar.rs

@@ -29,14 +29,14 @@
 //!    re-evaluate effectful scrutinees per arm).
 //! 3. Build a chain of single-level matches via
 //!    `build_chain(s_var, arms, default)` where `default` is the
-//!    polymorphic bottom builtin `__unreachable__` (`forall a. a`,
-//!    Iter 16d) — codegen lowers it to LLVM `unreachable`. Valid
-//!    programs never reach it because the checker requires
-//!    exhaustiveness (the catch-all arm dominates the chain).
-//!    Pre-16d the default was a `Unit` literal, which forced any
-//!    match whose arms returned a non-Unit type to carry a
-//!    synthetic `_` arm dominating the terminator. The polymorphic
-//!    `__unreachable__` removes that workaround.
+//!    polymorphic bottom builtin `__unreachable__` (`forall a. a`)
+//!    — codegen lowers it to LLVM `unreachable`. Valid programs
+//!    never reach it because the checker requires exhaustiveness
+//!    (the catch-all arm dominates the chain). An earlier shape used
+//!    a `Unit` literal as the default, which forced any match whose
+//!    arms returned a non-Unit type to carry a synthetic `_` arm
+//!    dominating the terminator. The polymorphic `__unreachable__`
+//!    removes that workaround.
 //! 4. Each arm is lowered via `desugar_one_arm`:
 //!    - `Pattern::Wild` → arm body (catch-all; later arms are dropped).
 //!    - `Pattern::Var { name }` → `Let { name = scrutinee_var; body }`.

crates/ailang-core/src/lib.rs

@@ -107,11 +107,11 @@ pub type Result<T> = std::result::Result<T, Error>;
 /// loading fails with [`Error::SchemaMismatch`].
 pub const SCHEMA: &str = "ailang/v0";
 
-/// Iter 20f: complete LLM-targeted specification of Form-A — the
-/// canonical authoring surface (Decision 6). Embedded verbatim into
-/// any prompt that asks an LLM to produce or edit AILang code; in
-/// particular, `ail merge-prose` includes it in the round-trip
-/// prompt template. The string is the raw bytes of
+/// Complete LLM-targeted specification of Form-A — the canonical
+/// authoring surface (see `design/contracts/authoring-surface.md`).
+/// Embedded verbatim into any prompt that asks an LLM to produce or
+/// edit AILang code; in particular, `ail merge-prose` includes it in
+/// the round-trip prompt template. The string is the raw bytes of
 /// `specs/form_a.md`, co-located with this crate so the spec sits
 /// next to the AST it describes; a drift test
 /// (`tests/spec_drift.rs`) walks every AST enum variant and asserts

crates/ailang-core/tests/spec_drift.rs

@@ -1,4 +1,4 @@
-//! Iter 20f: drift detection between the AST and `specs/form_a.md`.
+//! Drift detection between the AST and `specs/form_a.md`.
 //!
 //! The spec is hand-curated, but it cannot silently fall behind the
 //! language. Every AST enum (`Term`, `Pattern`, `Type`, `Def`, `Literal`,

crates/ailang-prose/tests/snapshot.rs

@@ -1,4 +1,4 @@
-//! Iter 20a snapshot tests.
+//! Prose-projection snapshot tests.
 //!
 //! For each `examples/<name>.ail`, the renderer's output is compared
 //! against the committed `examples/<name>.prose.txt`. The snapshot