diff --git a/CHANGELOG.md b/CHANGELOG.md
new file mode 100644
--- /dev/null
+++ b/CHANGELOG.md
@@ -0,0 +1,1372 @@
+# Hydra changelog
+
+All notable changes to the Hydra project are documented in this file.
+
+This changelog tracks changes across all Hydra implementations
+(Haskell, Java, Python, Scala, TypeScript, Clojure, Common Lisp, Emacs Lisp, Scheme)
+and supporting infrastructure.
+
+The format is inspired by [Keep a Changelog](https://keepachangelog.com/en/1.0.0/)
+and [Apache TinkerPop](https://github.com/apache/tinkerpop/blob/master/CHANGELOG.asciidoc).
+
+Pre-1.0 versions are listed below.
+Untagged versions correspond to in-repository version bumps that were not pushed as Git tags or Hackage releases;
+they are documented here for completeness.
+
+---
+
+## [0.17.0] - 2026-07-01
+
+Major themes: first-class support for effects and I/O in the kernel, TypeScript and
+Scala reaching publishable parity with the other hosts, and continued maturation of
+the published-host build model.
+
+### Highlights
+
+- **Effects and I/O** ([#286](https://github.com/CategoricalData/hydra/issues/286),
+  [#288](https://github.com/CategoricalData/hydra/issues/288),
+  [#494](https://github.com/CategoricalData/hydra/issues/494)): a new `effect` variant
+  on `hydra.core.Type` for representing effectful computations; new kernel modules
+  `hydra.file` (`FilePath`, `FileType`, `FileStatus`, `FileExtension`), `hydra.time`
+  (`Timespec`), and `hydra.error.file`; new primitive libraries `hydra.lib.effects`,
+  `hydra.lib.files`, and `hydra.lib.text` with byte-level I/O; and a new
+  `EffectfulTestCase` mechanism so the common test suite can exercise effectful
+  primitives across hosts.
+- **TypeScript is a complete implementation**
+  ([#462](https://github.com/CategoricalData/hydra/issues/462)): self-hosts and passes
+  the common test suite, with a per-package npm publish pipeline
+  ([#492](https://github.com/CategoricalData/hydra/issues/492)).
+- **Scala publishing**: per-package Maven Central publish infrastructure
+  ([#491](https://github.com/CategoricalData/hydra/issues/491)).
+- **Published-host build model**: published mode now compiles the kernel from the
+  co-generated `dist/haskell` rather than Hackage, removing the need for a
+  `haskell=local` shim ([#500](https://github.com/CategoricalData/hydra/issues/500));
+  the oil-and-water principle and cold-build workflow for `dist/haskell` are documented
+  ([#376](https://github.com/CategoricalData/hydra/issues/376)).
+- **Kernel hygiene**: `uniqueLabel` (apostrophe suffixes) replaced by `chooseUniqueLabel`
+  (numeric suffixes) in `hydra.names`
+  ([#436](https://github.com/CategoricalData/hydra/issues/436)).
+
+### Implementations
+
+Eight hosts self-host and pass the common test suite across five families: Haskell,
+Java, Python, Scala, and the four Lisp dialects (Clojure, Common Lisp, Scheme, Emacs
+Lisp), with TypeScript now joining as a complete host/target. Go remains a "head bud"
+(the coder still has emission bugs and the head's runtime is largely placeholder; it
+does not yet host the test suite).
+
+### New Features
+
+- Translingual cross-reference convention for `doc` strings: doc-escape tags render to
+  each host's native doc syntax (Javadoc, Haddock, Sphinx, TSDoc)
+  ([#433](https://github.com/CategoricalData/hydra/issues/433)).
+- Neo4j JSON-artifact validation demo, exercised across Java, Python, Haskell, and
+  TypeScript ([#510](https://github.com/CategoricalData/hydra/issues/510)).
+
+### Improvements
+
+- The Java/Maven publish set now covers every official Hydra target language, one coder
+  package per implementation family
+  ([#468](https://github.com/CategoricalData/hydra/issues/468)).
+- JVM Maven artifact groups are bifurcated to `net.fortytwo.hydra.java` and
+  `net.fortytwo.hydra.scala`, so the Java and Scala (and future Clojure) publish sets
+  never collide on coordinates
+  ([#519](https://github.com/CategoricalData/hydra/issues/519)).
+- Host-specific third-party integrations (the former `bindings/` tree) are folded into
+  `overlay/`, unifying hand-written host-native source under one tree
+  ([#511](https://github.com/CategoricalData/hydra/issues/511)).
+- The coder de-parameterizes context and error types, simplifying generated coder
+  signatures ([#518](https://github.com/CategoricalData/hydra/issues/518)).
+- TypeScript-to-Java generation now emits the `hydra.lib.*` aggregate namespace
+  interfaces (`Maps`, `Strings`, etc.)
+  ([#499](https://github.com/CategoricalData/hydra/issues/499)).
+
+---
+
+## [0.16.1] - 2026-06-16
+
+Major themes: the first release built by bootstrapping from previously published
+packages, the first full 9 × 9 cross-implementation bootstrapping sweep, and
+published-package self-containment fixes for the artifacts that bootstrap rests on.
+
+### Highlights
+
+- **Bootstrap from published packages**: the build links the published 0.16.0 hosts
+  (Hackage / Maven Central / PyPI) and uses them to build the next version of Hydra,
+  resting on 0.16.0's forward-compatibility guarantees — the published-host consume
+  model with a `--local-host` shim, used for the first time in a release
+  ([#370](https://github.com/CategoricalData/hydra/issues/370)). 0.16.1 still builds
+  against **local** Haskell and Python hosts (`hostOverrides`), because the published
+  0.16.0 artifacts it would otherwise consume are the very ones being repaired below.
+- **First full 9 × 9 bootstrapping sweep** — every implementation generating every
+  other; dozens of cross-host issues surfaced and fixed
+  ([#409](https://github.com/CategoricalData/hydra/issues/409), ongoing).
+- **Published-package self-containment**: relocate each host's hand-written runtime
+  into `overlay/<lang>/hydra-kernel/` so it ships inside the published `hydra-kernel`
+  artifact — Python `hydra.python.util` into the wheel
+  ([#461](https://github.com/CategoricalData/hydra/issues/461),
+  [#472](https://github.com/CategoricalData/hydra/issues/472)) and the Haskell
+  authoring DSL + `Libraries`/`Terms` into the sdist
+  ([#473](https://github.com/CategoricalData/hydra/issues/473)). Generalized to a
+  uniform `overlay/ = transform(packages) + copy(overlay)` model across all hosts
+  ([#434](https://github.com/CategoricalData/hydra/issues/434)).
+- **Kernel self-containment via generated `hydra.lib.*` registries**: primitive
+  names are derived from generated `PrimitiveDefinition` def-modules rather than
+  hand-written name strings, in every host
+  ([#473](https://github.com/CategoricalData/hydra/issues/473)).
+- **In-memory encode/decode synthesis**: derived `hydra.{encode,decode}.*` modules
+  are synthesized at build time and excluded from `dist/`, with the synthesizer (not
+  inference) as the authority for their annotations
+  ([#448](https://github.com/CategoricalData/hydra/issues/448),
+  [#476](https://github.com/CategoricalData/hydra/issues/476)).
+- **Release hardening**: Apache source-release compliance (signed source archive,
+  LICENSE/NOTICE bundling, KEYS) ([#441](https://github.com/CategoricalData/hydra/issues/441)),
+  per-package publish orchestrators ([#449](https://github.com/CategoricalData/hydra/issues/449)),
+  and packaging-boundary verification gated in CI and at publish time
+  ([#472](https://github.com/CategoricalData/hydra/issues/472)).
+
+### New features
+
+- Published-host consume model wired through `bin/sync.sh` for all three hosts
+  (`--published-host` default, `--local-host` shim), with the version basis keyed
+  off `hydra.json` and the `VERSION` file retired
+  ([#370](https://github.com/CategoricalData/hydra/issues/370),
+  [#347](https://github.com/CategoricalData/hydra/issues/347)).
+- Java and Python coder DSL sources made the sole source of truth; the legacy
+  Haskell DSL copies were deleted and cross-package references resolve by name
+  ([#346](https://github.com/CategoricalData/hydra/issues/346)).
+- Per-host `verify-distribution.sh` self-containment gates (Haskell stages the
+  per-package trio and `stack build`s it; Python installs the wheels into an
+  isolated venv and imports the kernel modules; Java resolves the published jars
+  from an offline consumer), run in CI and as a hard gate in `bin/prepare-release.sh`
+  ([#472](https://github.com/CategoricalData/hydra/issues/472)).
+- Apache-compliant canonical source archive: signed + checksummed `git archive`
+  tarball as the release of record, root `NOTICE`, `KEYS`, and a release-verify CI
+  workflow ([#441](https://github.com/CategoricalData/hydra/issues/441)).
+- Per-package publish orchestrators `publish-maven.sh` / `publish-pypi.sh` with
+  dependency-closure and leaves-first ordering checks
+  ([#449](https://github.com/CategoricalData/hydra/issues/449)).
+- `hydra-typescript` published as a target Java/Maven + PyPI artifact
+  ([#468](https://github.com/CategoricalData/hydra/issues/468)).
+- Java coder emits a fluent builder for every generated record type
+  ([#465](https://github.com/CategoricalData/hydra/issues/465)); Python coder emits
+  per-field `with_<field>` copy-update methods and a fluent builder
+  ([#466](https://github.com/CategoricalData/hydra/issues/466)).
+
+### Improvements
+
+- Drop the vestigial `InferenceContext` parameter from the `Primitive.implementation`
+  carrier across all hosts; add a graph-free `funT` coder and six `lists` default
+  implementations ([#446](https://github.com/CategoricalData/hydra/issues/446)).
+- Derive package routing from declared modules; split derived generation into broad
+  `mainDslModules` and scoped `mainEncodingModules`
+  ([#474](https://github.com/CategoricalData/hydra/issues/474),
+  [#475](https://github.com/CategoricalData/hydra/issues/475)).
+- Hash JSON content into the per-package input digest; add `digest-check
+  refresh-input` and a CI guard that force-regenerates and diffs
+  `dist/haskell/hydra-{java,python}` ([#469](https://github.com/CategoricalData/hydra/issues/469)).
+- Per-package inference seeds downstream packages from each package's full universe
+  while inferring only write targets, fixing `hydra-scala` inference within the CI
+  heap budget; declare the `hydra-java` dependency in `hydra-scala` for cold
+  ordering ([#470](https://github.com/CategoricalData/hydra/issues/470)).
+- Remove the vestigial `Lib/Defaults` registry and empty `defaultLibModules`
+  manifest field from all hosts ([#437](https://github.com/CategoricalData/hydra/issues/437)).
+- Bundle LICENSE + NOTICE into Java jars (`META-INF`) and Python wheels
+  (`dist-info`); declare Apache-2.0 in the Scala build
+  ([#441](https://github.com/CategoricalData/hydra/issues/441)).
+- Mark lazy parameters on the Scala and Common Lisp primitive registries so eager
+  coders thunk correctly ([#453](https://github.com/CategoricalData/hydra/issues/453),
+  [#477](https://github.com/CategoricalData/hydra/issues/477),
+  [#425](https://github.com/CategoricalData/hydra/issues/425)).
+- Staging-loop WIP-commit gate to keep `WIP:` commits off `origin/main`
+  ([#466](https://github.com/CategoricalData/hydra/issues/466)).
+
+### Bug fixes
+
+- **Python wheel self-containment**: the `hydra-kernel` wheel now ships
+  `hydra.python.util`; gated by a packaging-boundary smoke test in CI and
+  `publish-pypi.sh` ([#472](https://github.com/CategoricalData/hydra/issues/472),
+  [#461](https://github.com/CategoricalData/hydra/issues/461)).
+- **Haskell sdist self-containment**: track the generated `Hydra/Dsl/Terms.hs`
+  (was gitignored) so the committed dist exports `ToPrimName`, fixing the cold
+  build of the published `hydra-kernel`
+  ([#473](https://github.com/CategoricalData/hydra/issues/473)).
+- Decode synthesizer threads the fully-applied result type into wrap/record/union
+  decoder bodies, so derived decoders emit `DataRow<v>` rather than raw `DataRow`,
+  fixing Java and Scala generic decode compilation
+  ([#476](https://github.com/CategoricalData/hydra/issues/476)).
+- Scala coder emits explicit type arguments for polymorphic non-primitive
+  references under a type application (e.g. `requireField`), fixing `Any`-inference
+  of curried decoder lambda parameters in `decode/typed`
+  ([#434](https://github.com/CategoricalData/hydra/issues/434)).
+- Common Lisp and Emacs Lisp: move the hand-written runtime to a dialect-tier path
+  so it no longer collides with the generated `hydra.lib.*` def-modules
+  ([#434](https://github.com/CategoricalData/hydra/issues/434)).
+- Python coder uses the binding `TypeScheme` as the authoritative source for
+  parameter and return types ([#488](https://github.com/CategoricalData/hydra/issues/488));
+  fixes a duplicate `self` parameter in builders
+  ([#478](https://github.com/CategoricalData/hydra/issues/478)).
+- Sort term definitions by name in source-file generation so non-Haskell hosts emit
+  deterministic alphabetical Haskell output (host-independent of `Maps.elems`
+  iteration order) ([#489](https://github.com/CategoricalData/hydra/issues/489)).
+- Java coder serializes the `static` method modifier (previously a non-exhaustive
+  case) ([#465](https://github.com/CategoricalData/hydra/issues/465)); JavaDoc
+  `@link`/`@param` sanitization for strict doclint
+  ([#449](https://github.com/CategoricalData/hydra/issues/449)).
+- Scala host writer appends a trailing newline when missing, fixing a Lisp-target
+  byte mismatch ([#463](https://github.com/CategoricalData/hydra/issues/463)).
+- Clojure host: silent wrapping `bigint_to_int{8,16,32,64}` casts per the kernel
+  contract ([#483](https://github.com/CategoricalData/hydra/issues/483)); load-order
+  and `--kernel-only` filter fixes after kernel renames
+  ([#425](https://github.com/CategoricalData/hydra/issues/425)); load the
+  `hydra-haskell` baseline for Clojure→Haskell rebuilds
+  ([#487](https://github.com/CategoricalData/hydra/issues/487),
+  [#460](https://github.com/CategoricalData/hydra/issues/460)).
+- Lisp-dialect bootstrap fixes: emit JSON object entries as 2-element `Pair` lists
+  rather than dotted cons ([#452](https://github.com/CategoricalData/hydra/issues/452),
+  also fixing [#16](https://github.com/CategoricalData/hydra/issues/16)); read
+  module names via the correct struct accessor
+  ([#471](https://github.com/CategoricalData/hydra/issues/471)); drop stale
+  surplus flag arguments from `generate-source-files` calls
+  ([#457](https://github.com/CategoricalData/hydra/issues/457),
+  [#458](https://github.com/CategoricalData/hydra/issues/458),
+  [#464](https://github.com/CategoricalData/hydra/issues/464)); rename eta-wrapped
+  projection/unwrap parameters to avoid kernel shadowing
+  ([#428](https://github.com/CategoricalData/hydra/issues/428)); propagate a
+  non-zero exit from the Scheme test runner on failures and fix Scheme float
+  scientific-notation formatting for `|x| >= 1e7`
+  ([#447](https://github.com/CategoricalData/hydra/issues/447)).
+- Cross-host target dispatch: Common Lisp and Clojure hosts dispatch Scala and
+  TypeScript targets ([#482](https://github.com/CategoricalData/hydra/issues/482),
+  [#483](https://github.com/CategoricalData/hydra/issues/483)); Scala/Java host
+  target-dispatch completions ([#5](https://github.com/CategoricalData/hydra/issues/5),
+  [#7](https://github.com/CategoricalData/hydra/issues/7),
+  [#17](https://github.com/CategoricalData/hydra/issues/17)).
+- TypeScript test runtime reads the kernel JSON field names `body`/`signature`
+  (not legacy `term`/`typeScheme`); `lists.nub` dedups by canonical JSON
+  ([#485](https://github.com/CategoricalData/hydra/issues/485)).
+- Sync robustness: always reassemble `dist/python` under `HYDRA_IN_SYNC` to heal
+  partial-dist state ([#480](https://github.com/CategoricalData/hydra/issues/480));
+  persist the run dir on bootstrap prep-phase failures
+  ([#481](https://github.com/CategoricalData/hydra/issues/481)); complete per-host
+  static deps so the bootstrap demo finds fresh dist
+  ([#445](https://github.com/CategoricalData/hydra/issues/445)).
+- Pin the Python target test interpreter to Python ≥ 3.12
+  ([#486](https://github.com/CategoricalData/hydra/issues/486)); Emacs Lisp / Scheme
+  target test-runner re-copy robustness
+  ([#456](https://github.com/CategoricalData/hydra/issues/456)).
+- JavaDoc Pages publish rewritten to build a combined site from the CI-generated
+  `dist/java/` artifact, fixing the failure mode where `pages.yml` ran against a
+  bare checkout with no `dist/java/` present
+  ([#450](https://github.com/CategoricalData/hydra/issues/450)).
+
+### Documentation
+
+- Document the `overlay/` self-containment model, the published-host consume model
+  for all three hosts, the `hydra.json` version / host-version basis, and a
+  migration-shims recipe ([#434](https://github.com/CategoricalData/hydra/issues/434),
+  [#370](https://github.com/CategoricalData/hydra/issues/370),
+  [#347](https://github.com/CategoricalData/hydra/issues/347)).
+- Document the kernel self-containment / `hydra.lib.*` def-module model and the
+  derived-module no-inference invariant
+  ([#473](https://github.com/CategoricalData/hydra/issues/473),
+  [#466](https://github.com/CategoricalData/hydra/issues/466)); generated Java and
+  Python record builders documented in the DSL guides
+  ([#465](https://github.com/CategoricalData/hydra/issues/465),
+  [#466](https://github.com/CategoricalData/hydra/issues/466)).
+- Use absolute GitHub URLs in package READMEs so links resolve when published
+  standalone ([#454](https://github.com/CategoricalData/hydra/issues/454)); bundle a
+  README into each Python package dir for sdist metadata
+  ([#418](https://github.com/CategoricalData/hydra/issues/418)).
+
+---
+
+## [0.16.0] - 2026-06-09
+
+Major themes: primitive-metadata reification, self-hosting coder DSLs with all
+eight hosts passing the common test suite, finalized `hydra.packaging`, and a
+deep kernel-vocabulary cleanup.
+
+### Highlights
+
+- **`PrimitiveDefinition` and per-namespace registry modules** with host-independent
+  specs for every standard-library primitive ([#156](https://github.com/CategoricalData/hydra/issues/156), [#319](https://github.com/CategoricalData/hydra/issues/319)).
+- **Self-hosting coder DSLs; all eight hosts pass the common test suite**, including
+  the completed TypeScript head ([#344](https://github.com/CategoricalData/hydra/issues/344), [#126](https://github.com/CategoricalData/hydra/issues/126)).
+- **Finalized `hydra.packaging`** and a unified `Module.dependencies` field ([#369](https://github.com/CategoricalData/hydra/issues/369), [#354](https://github.com/CategoricalData/hydra/issues/354)).
+- **Kernel-vocabulary cleanup**: `maybe`→`optional` across kernel and hosts, plus
+  order-preserving JSON objects ([#401](https://github.com/CategoricalData/hydra/issues/401), [#398](https://github.com/CategoricalData/hydra/issues/398)).
+
+### New features
+
+- Reify `PrimitiveDefinition` in `hydra.packaging` with per-namespace `hydra.lib.<sub>` registries ([#156](https://github.com/CategoricalData/hydra/issues/156)).
+- Host-independent `comments` specifications for all standard-library primitives ([#319](https://github.com/CategoricalData/hydra/issues/319)).
+- Kernel-oriented style checks in `hydra.validate.*` ([#321](https://github.com/CategoricalData/hydra/issues/321), [#351](https://github.com/CategoricalData/hydra/issues/351), [#352](https://github.com/CategoricalData/hydra/issues/352)), on a reshaped validation API ([#320](https://github.com/CategoricalData/hydra/issues/320)).
+- Promote the JSON Schema coder to a full DSL coder ([#350](https://github.com/CategoricalData/hydra/issues/350)).
+- Transpose coder sources into host-native DSLs for self-hosting ([#344](https://github.com/CategoricalData/hydra/issues/344), [#358](https://github.com/CategoricalData/hydra/issues/358)).
+- `Term`-typed annotations replacing `Map<Name, Term>` ([#386](https://github.com/CategoricalData/hydra/issues/386)).
+- Language-specific EDSL features on the kernel `Language` ([#233](https://github.com/CategoricalData/hydra/issues/233)).
+- Primitive-metadata-driven laziness in eager coders ([#391](https://github.com/CategoricalData/hydra/issues/391)).
+- `/test` skill and `bin/test.sh` for target-language test validation ([#387](https://github.com/CategoricalData/hydra/issues/387)).
+- Automatic differentiation: a `grad` transform ([#324](https://github.com/CategoricalData/hydra/issues/324)).
+- `hydra.show.error.pg` for `InvalidGraphError` ([#374](https://github.com/CategoricalData/hydra/issues/374)).
+
+### Improvements
+
+- Finalize `hydra.packaging`: `Binding`→`TypeDefinition`, `Module`/`Package` `comments`, `moduleName` cleanup ([#369](https://github.com/CategoricalData/hydra/issues/369), [#396](https://github.com/CategoricalData/hydra/issues/396), [#402](https://github.com/CategoricalData/hydra/issues/402), [#404](https://github.com/CategoricalData/hydra/issues/404)).
+- Merge `Module.termDependencies`/`typeDependencies` into one `dependencies` field ([#354](https://github.com/CategoricalData/hydra/issues/354)).
+- Resolve the `maybe`/`optional` ambiguity across kernel and all hosts ([#401](https://github.com/CategoricalData/hydra/issues/401)).
+- Order-preserving JSON objects; order-sensitive `Value` equality ([#398](https://github.com/CategoricalData/hydra/issues/398)).
+- Rename `Namespace`→`ModuleName` ([#316](https://github.com/CategoricalData/hydra/issues/316)) and the phantom DSL types to `Typed*` ([#397](https://github.com/CategoricalData/hydra/issues/397)).
+- Move from a type-class enum to type-class names ([#275](https://github.com/CategoricalData/hydra/issues/275)).
+- Repackage "eval lib" modules as kernel lib modules; promote remaining sources ([#260](https://github.com/CategoricalData/hydra/issues/260), [#355](https://github.com/CategoricalData/hydra/issues/355), [#356](https://github.com/CategoricalData/hydra/issues/356)).
+- Reshape `hydra.context` into `InferenceContext` ([#368](https://github.com/CategoricalData/hydra/issues/368)).
+- Streamline the Haskell, Java, Python, and Scala syntax models, re-pinning Java to SE 21 and Python to the 3.14 PEG grammar ([#297](https://github.com/CategoricalData/hydra/issues/297)).
+- Adopt persistent immutable collections across hosts ([#359](https://github.com/CategoricalData/hydra/issues/359), [#360](https://github.com/CategoricalData/hydra/issues/360), [#361](https://github.com/CategoricalData/hydra/issues/361), [#362](https://github.com/CategoricalData/hydra/issues/362)).
+- Migrate underscored kernel definition names to camelCase ([#348](https://github.com/CategoricalData/hydra/issues/348)).
+- Tidy qualified/unqualified imports in the Haskell DSL sources ([#308](https://github.com/CategoricalData/hydra/issues/308)).
+- Eliminate the `Hydra.Module.Compat` shim ([#315](https://github.com/CategoricalData/hydra/issues/315)).
+- Homogenize writer conventions across target languages ([#339](https://github.com/CategoricalData/hydra/issues/339)).
+- Disambiguate format-version field naming across JSON metadata ([#412](https://github.com/CategoricalData/hydra/issues/412)).
+- Per-package DSL→JSON build with per-package incremental inference ([#381](https://github.com/CategoricalData/hydra/issues/381), [#395](https://github.com/CategoricalData/hydra/issues/395)).
+- Prune stale `dist/` outputs; self-heal digests; exclude digests from VCS; reconcile drift; fail loud ([#357](https://github.com/CategoricalData/hydra/issues/357), [#393](https://github.com/CategoricalData/hydra/issues/393), [#405](https://github.com/CategoricalData/hydra/issues/405), [#378](https://github.com/CategoricalData/hydra/issues/378), [#379](https://github.com/CategoricalData/hydra/issues/379), [#392](https://github.com/CategoricalData/hydra/issues/392), [#414](https://github.com/CategoricalData/hydra/issues/414)).
+- Accept all valid (host, target) pairs in the generation drivers ([#421](https://github.com/CategoricalData/hydra/issues/421)).
+- Move Gradle build infrastructure into `heads/java/`; remove top-level `pixi.toml` ([#384](https://github.com/CategoricalData/hydra/issues/384), [#385](https://github.com/CategoricalData/hydra/issues/385)).
+- Eliminate post-generation test patches; fix loader/runner kludges; re-enable test timing ([#307](https://github.com/CategoricalData/hydra/issues/307), [#309](https://github.com/CategoricalData/hydra/issues/309), [#311](https://github.com/CategoricalData/hydra/issues/311)).
+- Performance/benchmarks and usability assessments ([#277](https://github.com/CategoricalData/hydra/issues/277), [#283](https://github.com/CategoricalData/hydra/issues/283)).
+- Split the Hackage distribution into per-package distributions ([#418](https://github.com/CategoricalData/hydra/issues/418)).
+
+### Bug fixes
+
+- Encode `uint32` JSON literals as numbers, not strings ([#431](https://github.com/CategoricalData/hydra/issues/431)).
+- Satisfy or suppress warnings in generated Java ([#349](https://github.com/CategoricalData/hydra/issues/349)).
+- Restore JSON parser test cases (polymorphic case handling, ordering fallback) ([#336](https://github.com/CategoricalData/hydra/issues/336)).
+- Keep annotations through `termAlternatives`/`typeAlternatives` adaptation ([#353](https://github.com/CategoricalData/hydra/issues/353)).
+- Scala host decodes module JSON containing term definitions ([#342](https://github.com/CategoricalData/hydra/issues/342)).
+- Java coder: fix exponential `let`-bound `cases` cost ([#372](https://github.com/CategoricalData/hydra/issues/372)) and explicit type args on `PersistentMap.ofEntries` ([#394](https://github.com/CategoricalData/hydra/issues/394)).
+- Python-host codegen: drop extra `pyStringToPyStrings` wrapping ([#367](https://github.com/CategoricalData/hydra/issues/367)).
+- `escape_literal_char` no longer mangles RDF triples ([#363](https://github.com/CategoricalData/hydra/issues/363)).
+- Sync robustness: stale coder JSON, omitted native modules, dead helper, bad source dirs ([#406](https://github.com/CategoricalData/hydra/issues/406), [#400](https://github.com/CategoricalData/hydra/issues/400), [#411](https://github.com/CategoricalData/hydra/issues/411), [#408](https://github.com/CategoricalData/hydra/issues/408)).
+- Cross-host self-hosting fixes (Scala, TypeScript, and the four Lisp dialects) ([#422](https://github.com/CategoricalData/hydra/issues/422), [#423](https://github.com/CategoricalData/hydra/issues/423), [#424](https://github.com/CategoricalData/hydra/issues/424), [#429](https://github.com/CategoricalData/hydra/issues/429), [#444](https://github.com/CategoricalData/hydra/issues/444), [#366](https://github.com/CategoricalData/hydra/issues/366), [#389](https://github.com/CategoricalData/hydra/issues/389), [#407](https://github.com/CategoricalData/hydra/issues/407), [#425](https://github.com/CategoricalData/hydra/issues/425), [#426](https://github.com/CategoricalData/hydra/issues/426), [#427](https://github.com/CategoricalData/hydra/issues/427), [#432](https://github.com/CategoricalData/hydra/issues/432), [#438](https://github.com/CategoricalData/hydra/issues/438), [#439](https://github.com/CategoricalData/hydra/issues/439), [#443](https://github.com/CategoricalData/hydra/issues/443)).
+
+### Removed
+
+- Removed the `bigfloat` literal type from `FloatType` and `FloatValue` ([#330](https://github.com/CategoricalData/hydra/issues/330)).
+  Arbitrary-precision floating-point values were inconsistently supported
+  across host languages (Java/Scala in particular lacked a true IEEE 754
+  arbitrary-precision implementation), and nothing in the codebase relied on
+  them. `FloatType` and `FloatValue` now have only `float32` and `float64`
+  variants. Removed library primitives: `bigfloatToBigint`, `bigfloatToFloat32`,
+  `bigfloatToFloat64`, `bigintToBigfloat`, `float32ToBigfloat`,
+  `float64ToBigfloat`, `readBigfloat`, `showBigfloat`, `roundBigfloat`. New
+  primitives `float32ToFloat64` and `float64ToFloat32` (the latter lossy)
+  replace the removed bigfloat-routed conversions.
+- Removed `heads/haskell/bin/assemble-hackage-sdist.sh` ([#418](https://github.com/CategoricalData/hydra/issues/418)), the 0.15-era
+  bridge that flattened the multi-source-dir Haskell head into one self-contained
+  `hydra` sdist. Superseded by the per-package assembler (see Improvements).
+
+### Documentation
+
+- Comprehensive documentation refresh for 0.16 ([#383](https://github.com/CategoricalData/hydra/issues/383))
+  spanning nine tracks across READMEs, `docs/`, the wiki, CLAUDE.md, and
+  in-source kernel comments.
+- New `docs/json-format.md` reference describing the canonical JSON wire format
+  (tagged-union duality, optional-field rules, IEEE 754 sentinels, integer
+  threshold).
+- New wiki pages: `Design` (rationale behind each major choice) and `Inference`
+  (HM type inference with class constraints).
+- Cross-checked every prose doc against current code; corrected drift in
+  primitive-checking sections of `docs/recipes/maintenance.md`, the `Primitive`
+  shape in `docs/implementation.md` and `docs/recipes/adding-primitives.md`,
+  and the build-system narrative.
+- Wove motivation ("the why") through the existing doc surface rather than
+  collecting it into separately-titled sidebars; landed an inference-overview
+  centerpiece on the wiki ([#377](https://github.com/CategoricalData/hydra/issues/377)).
+- Aligned doc vocabulary with the LambdaGraph paper and the literature review
+  for the design space Hydra inhabits.
+- Net-new READMEs for `packages/hydra-kernel`, `packages/hydra-ext`,
+  `packages/hydra-coq`, and `packages/hydra-wasm`; top-level `README.md`
+  now carries a canonical packages table.
+- Audited and refreshed the Claude surface: CLAUDE.md "Where to look up X"
+  table extended with rows for `docs/getting-started.md`, `docs/json-format.md`,
+  and the `Design` / `Inference` wiki pages; hard-rules block reordered by
+  violation frequency; `claude/pitfalls.md` superseded entries folded into
+  `docs/troubleshooting.md` / recipes.
+- Project-descriptor strings reviewed and aligned across 24 build / packaging
+  files (Hackage synopsis, Maven Central description, PyPI / conda-forge
+  classifiers, Hydra-format `package.json` files).
+- `docs/index.html` (the GitHub Pages landing page) refreshed with audience
+  hops into wiki Concepts / Design, `docs/getting-started.md`, and the GitHub
+  repo, plus the full published-artifact list across Hackage, Maven Central,
+  PyPI, and conda-forge.
+- In-source kernel-code rationale comments added at key Sources sites covering
+  injection / pair / either / unit-void duality, error taxonomy, variant
+  records, inference-context threading, validation profiles, reduction state,
+  lexical filters, name mangling, scoping, variable handling, defaults
+  integration, and Tarjan SCC choice rationale. JSON encode/decode modules
+  cross-reference `docs/json-format.md`.
+- Filed `Primitive.implementation` carrier-simplification follow-up
+  ([#446](https://github.com/CategoricalData/hydra/issues/446))
+  to drop the unused `InferenceContext` and `Graph` parameters once the
+  `Lib/Defaults/` integration ([#437](https://github.com/CategoricalData/hydra/issues/437))
+  lands.
+
+---
+
+## [0.15.0] - 2026-04-29
+
+A structural-cleanup release.
+The repository is reorganized into a three-tree
+`packages/` (DSL sources) / `heads/` (per-host runtime) / `dist/` (generated output)
+layout ([#290](https://github.com/CategoricalData/hydra/issues/290)), the kernel is simplified ([#251](https://github.com/CategoricalData/hydra/issues/251), [#292](https://github.com/CategoricalData/hydra/issues/292), [#332](https://github.com/CategoricalData/hydra/issues/332)), and Java + Python now
+ship as per-package publishable artifacts on Maven Central / PyPI / conda-forge
+([#305](https://github.com/CategoricalData/hydra/issues/305)).
+Incremental, content-hash-based caches accelerate the sync pipeline ([#247](https://github.com/CategoricalData/hydra/issues/247)).
+Three new generation-only targets — Coq ([#326](https://github.com/CategoricalData/hydra/issues/326)), WebAssembly ([#325](https://github.com/CategoricalData/hydra/issues/325)), and an
+automatic-differentiation demo ([#324](https://github.com/CategoricalData/hydra/issues/324)) — are in progress as Claude collaborations.
+
+### Highlights
+
+- **Three-tree repository layout**: `packages/` (DSL), `heads/` (host runtime),
+  `dist/` (generated) ([#290](https://github.com/CategoricalData/hydra/issues/290)).
+- **Per-package distributions** for Java (4 Maven artifacts) and Python (5 wheels):
+  `hydra-kernel`, `hydra-pg`, `hydra-rdf`, `hydra-{java,python}`, plus `hydra-ext`
+  on PyPI ([#305](https://github.com/CategoricalData/hydra/issues/305)).
+- **Kernel simplifications**: `Function`/`Elimination` removed ([#332](https://github.com/CategoricalData/hydra/issues/332)),
+  `FunctionPrimitive` removed ([#251](https://github.com/CategoricalData/hydra/issues/251)), `Hydra.Ext.*` prefix retired ([#331](https://github.com/CategoricalData/hydra/issues/331)),
+  `Context` error wrapper gone ([#292](https://github.com/CategoricalData/hydra/issues/292)).
+- **Incremental caches** across the sync pipeline; ~70x speedup on no-op resync ([#247](https://github.com/CategoricalData/hydra/issues/247)).
+- **JSON kernel format v1** with formal wire-format spec and four-field rename ([#343](https://github.com/CategoricalData/hydra/issues/343)).
+- **Three new generation-only targets** (in progress): Coq ([#326](https://github.com/CategoricalData/hydra/issues/326)), WebAssembly ([#325](https://github.com/CategoricalData/hydra/issues/325)),
+  automatic-differentiation demo ([#324](https://github.com/CategoricalData/hydra/issues/324)).
+
+### New features
+
+- Per-package distributions ([#305](https://github.com/CategoricalData/hydra/issues/305)): standalone Maven / PyPI / conda-forge builds,
+  generated `build.gradle` / `pyproject.toml` per package, transitive dep resolution.
+- JSON kernel format v1 ([#343](https://github.com/CategoricalData/hydra/issues/343)): formal spec at `docs/json-format.md`,
+  four-field rename to `typeScheme`, `Module` field reorder, `formatVersion` stamp.
+- `decimal` literal type in the kernel ([#338](https://github.com/CategoricalData/hydra/issues/338)) across all hosts.
+- Type-directed JSON encoder with idiomatic optional encoding ([#314](https://github.com/CategoricalData/hydra/issues/314), [#318](https://github.com/CategoricalData/hydra/issues/318)).
+- Incremental type inference plus content-hash caches ([#247](https://github.com/CategoricalData/hydra/issues/247)).
+- Coq generation target ([#326](https://github.com/CategoricalData/hydra/issues/326), in progress): `hydra-coq` package, 131/131 `.v` files pass `coqc`.
+- WebAssembly target ([#325](https://github.com/CategoricalData/hydra/issues/325), in progress): `hydra-wasm` package, M4 closure mechanism.
+- Automatic-differentiation demo ([#324](https://github.com/CategoricalData/hydra/issues/324), in progress): symbolic source-to-source.
+- `hydra.show.error` module ([#265](https://github.com/CategoricalData/hydra/issues/265)): consolidated error-message builders.
+- `hydra.lib.maybes.toList` primitive ([#257](https://github.com/CategoricalData/hydra/issues/257)).
+- NaN / Inf round-tripping through JSON and per-language serdes ([#312](https://github.com/CategoricalData/hydra/issues/312), [#318](https://github.com/CategoricalData/hydra/issues/318), [#330](https://github.com/CategoricalData/hydra/issues/330)).
+- Typeclass inference test group ([#274](https://github.com/CategoricalData/hydra/issues/274)).
+- Term and type linters ([#232](https://github.com/CategoricalData/hydra/issues/232)).
+- Validation for modules and graphs ([#155](https://github.com/CategoricalData/hydra/issues/155)).
+- User-defined-function (UDF) support and DSL-vs-primitive overrides
+  ([#62](https://github.com/CategoricalData/hydra/issues/62), [#63](https://github.com/CategoricalData/hydra/issues/63), [#102](https://github.com/CategoricalData/hydra/issues/102), [#158](https://github.com/CategoricalData/hydra/issues/158)).
+- Stale-source detection across sync entrypoints ([#228](https://github.com/CategoricalData/hydra/issues/228)).
+
+### Improvements
+
+- Sync infrastructure: matrix tool, per-package orchestrator, per-language wrappers ([#290](https://github.com/CategoricalData/hydra/issues/290)).
+- TestGraph post-generation patches eliminated; DSL emits `TestEnv` refs directly.
+- Minimized standard imports in generated Haskell ([#161](https://github.com/CategoricalData/hydra/issues/161)); improved auto-aliases ([#322](https://github.com/CategoricalData/hydra/issues/322)).
+- Removed kernel term dependencies from the interpreter ([#257](https://github.com/CategoricalData/hydra/issues/257)).
+- Removed `Graph` and `Context` arguments from primitive implementations ([#266](https://github.com/CategoricalData/hydra/issues/266)).
+- Refactored kernel term modules: split rewriting/schemas, renamed show.meta → codegen,
+  moved `Coder` to `coders`, merged `extract.helpers` ([#221](https://github.com/CategoricalData/hydra/issues/221)).
+- Removed Aeson and HsYAML dependencies; YAML coder moved to `hydra-ext` ([#261](https://github.com/CategoricalData/hydra/issues/261)).
+- Lisp recursive let bindings via SCC ([#341](https://github.com/CategoricalData/hydra/issues/341)); parenthesis cleanup ([#84](https://github.com/CategoricalData/hydra/issues/84)).
+- Everything-to-everything bootstrapping demo ([#254](https://github.com/CategoricalData/hydra/issues/254)).
+- Re-enabled skipped Python tests; `disabledForPython` retired ([#263](https://github.com/CategoricalData/hydra/issues/263)).
+- Promoted host-specific code in `packages/` into per-head locations ([#337](https://github.com/CategoricalData/hydra/issues/337)).
+- Migrated `hydra.common` Java module out ([#10](https://github.com/CategoricalData/hydra/issues/10)).
+- Test infrastructure unification under `UniversalTestCase` ([#246](https://github.com/CategoricalData/hydra/issues/246)).
+- Default property graph → RDF mapping ([#296](https://github.com/CategoricalData/hydra/issues/296)).
+- 13 unsafe partial primitives replaced with `Maybe`-returning alternatives ([#201](https://github.com/CategoricalData/hydra/issues/201)).
+- Java generated code uses standard `java.util` collection interfaces ([#313](https://github.com/CategoricalData/hydra/issues/313)).
+- `Term.union` renamed to `Term.inject` ([#334](https://github.com/CategoricalData/hydra/issues/334)).
+- Inconsistencies in domain DSLs cleaned up ([#219](https://github.com/CategoricalData/hydra/issues/219)).
+- Coder packages moved out of hydra-haskell ([#182](https://github.com/CategoricalData/hydra/issues/182)).
+- Java package-naming scheme revised ([#223](https://github.com/CategoricalData/hydra/issues/223)).
+- Decimal used (instead of bigfloat) for JSON numbers ([#340](https://github.com/CategoricalData/hydra/issues/340)).
+- Hydra kernel moved out of hydra-haskell ([#282](https://github.com/CategoricalData/hydra/issues/282)).
+
+### Bug fixes
+
+- `substInTypeScheme` capture-unsafe substitution under nested forall binders.
+- Java `Float64` arithmetic primitives corrected for current `PrimitiveFunction` API.
+- Clojure decimal handling: value-based equality, `BigInteger` return from `decimal_to_bigint`.
+- Scheme runtime R7RS imports for `sets.scm` / `maps.scm` / `eithers.map_set`.
+- Python coder defaults corrected from `Py.Name` to `Py.Expression`.
+- Scala host JSON-decode failure on modules containing term definitions ([#342](https://github.com/CategoricalData/hydra/issues/342)).
+- Lisp coder mutually-recursive let bindings for Clojure ([#341](https://github.com/CategoricalData/hydra/issues/341)).
+
+### Documentation
+
+- Documentation refresh for the 0.15 packaging restructure ([#290](https://github.com/CategoricalData/hydra/issues/290), [#331](https://github.com/CategoricalData/hydra/issues/331)); CLAUDE.md,
+  READMEs, recipes, demos, and seven wiki pages updated.
+- Code-generation recipe rewritten for the per-package dist layout ([#282](https://github.com/CategoricalData/hydra/issues/282)).
+- New `maintenance.md` recipe consolidates non-source-file scans, stale generated
+  detection, design-violation checks, and freshness checks.
+- Documentation style guide added at `docs/documentation-style-guide.md`.
+- Cross-worktree messaging protocol documented in CLAUDE.md.
+- Decimal type documented in the lexicon and per-language READMEs ([#338](https://github.com/CategoricalData/hydra/issues/338)).
+- JSON encoding format docs updated for idiomatic optional encoding ([#314](https://github.com/CategoricalData/hydra/issues/314)).
+
+---
+
+## [0.14.1] - 2026-03-30
+
+Patch release.
+Regenerates all implementations after a fix to the group eval primitive ([#281](https://github.com/CategoricalData/hydra/issues/281)),
+and relocates the property-graph and RDF/SHACL "ext" modules into hydra-java and hydra-python.
+
+### Improvements
+
+- Moved PG, RDF, and SHACL ext modules from hydra-ext into hydra-java and hydra-python.
+
+### Bug fixes
+
+- Group eval primitive fix; all targets regenerated ([#281](https://github.com/CategoricalData/hydra/issues/281)).
+- JavaDoc HTML-entity escaping in generated `DestructuringPattern.java`.
+
+---
+
+## [0.14.0] - 2026-03-29
+
+Major release adding four new complete Hydra implementations — Scala ([#273](https://github.com/CategoricalData/hydra/issues/273)),
+Clojure ([#278](https://github.com/CategoricalData/hydra/issues/278)), Common Lisp, and Scheme — bringing the total to seven bootstrapping hosts
+(plus Emacs Lisp, which passes the test suite but does not yet participate in bootstrapping).
+Also includes kernel simplifications, an Avro bidirectional coder ([#301](https://github.com/CategoricalData/hydra/issues/301)),
+eval primitives ([#281](https://github.com/CategoricalData/hydra/issues/281)), new demos, and significant performance work on the Scheme host.
+
+### Highlights
+
+- **Four new complete implementations**: Hydra-Scala ([#273](https://github.com/CategoricalData/hydra/issues/273)), Hydra-Clojure ([#278](https://github.com/CategoricalData/hydra/issues/278)),
+  Hydra-Common Lisp, and Hydra-Scheme pass the common test suite and can serve as
+  bootstrapping hosts.
+  The four Lisp dialects share a single coder and serializer.
+- **Kernel simplifications**: removed `RowType`/`WrappedType` ([#82](https://github.com/CategoricalData/hydra/issues/82)),
+  unified `Graph`/`TypeContext`/`InferenceContext` ([#192](https://github.com/CategoricalData/hydra/issues/192)),
+  replaced `Module.elements` with `Module.definitions` ([#214](https://github.com/CategoricalData/hydra/issues/214)),
+  consolidated error types ([#268](https://github.com/CategoricalData/hydra/issues/268)), and promoted all staging modules to DSL sources ([#267](https://github.com/CategoricalData/hydra/issues/267)).
+- **Avro bidirectional coder** ([#301](https://github.com/CategoricalData/hydra/issues/301)) with property graph pipeline demo.
+- **Scheme bootstrap performance**: 58 minutes → 5.7 minutes via O(1) vhash data structures
+  and IEEE 754 float precision fixes.
+
+### Breaking changes
+
+- Removed legacy adapter framework; replaced by the staging-promoted variant ([#236](https://github.com/CategoricalData/hydra/issues/236), [#295](https://github.com/CategoricalData/hydra/issues/295)).
+- Replaced legacy Flow monad with Either ([#245](https://github.com/CategoricalData/hydra/issues/245)).
+- Removed `Module.elements`; use `Module.definitions` ([#214](https://github.com/CategoricalData/hydra/issues/214)).
+- Removed `RowType` and `WrappedType` ([#82](https://github.com/CategoricalData/hydra/issues/82)).
+- Renamed `hydra.accessors` to `hydra.paths` ([#271](https://github.com/CategoricalData/hydra/issues/271)).
+- Renamed `hydra.tarjan` and merged into `hydra.sorting` ([#220](https://github.com/CategoricalData/hydra/issues/220)).
+- Merged `hydra.constraints` into `hydra.query` ([#272](https://github.com/CategoricalData/hydra/issues/272)).
+- Moved `hydra.workflow` to hydra-ext ([#270](https://github.com/CategoricalData/hydra/issues/270)).
+- Consolidated `hydra.compute` into `hydra.util` ([#269](https://github.com/CategoricalData/hydra/issues/269)).
+- Removed `hydra.describe` ([#216](https://github.com/CategoricalData/hydra/issues/216)) and types from `hydra.mantle` ([#217](https://github.com/CategoricalData/hydra/issues/217)).
+- Replaced `OtherError` with task-specific error types ([#268](https://github.com/CategoricalData/hydra/issues/268)).
+- Java: lowered minimum Java version to 11 ([#249](https://github.com/CategoricalData/hydra/issues/249)).
+- Java: `Pair` replaces `Tuple2` ([#252](https://github.com/CategoricalData/hydra/issues/252)); `Comparable` implemented for generated classes ([#131](https://github.com/CategoricalData/hydra/issues/131)).
+
+### New features
+
+- Hydra-Scala ([#273](https://github.com/CategoricalData/hydra/issues/273)): complete implementation, sbt/Scala 3, 3043 tests.
+- Hydra-Lisp: Clojure ([#278](https://github.com/CategoricalData/hydra/issues/278)), Common Lisp, Emacs Lisp, Scheme — four dialects,
+  one coder. CI workflows for Clojure, Common Lisp, and Scheme.
+- Avro bidirectional coder ([#301](https://github.com/CategoricalData/hydra/issues/301)) with demo.
+- Eval primitives ([#281](https://github.com/CategoricalData/hydra/issues/281)): Logic, Equality, Math, collections, and Groups.
+- Generated DSL modules ([#180](https://github.com/CategoricalData/hydra/issues/180)): `hydra.dsls` generates DSL helpers from type modules.
+- New demos: GraphQL+JSON ([#279](https://github.com/CategoricalData/hydra/issues/279)), PG validation ([#284](https://github.com/CategoricalData/hydra/issues/284)), SHACL ([#294](https://github.com/CategoricalData/hydra/issues/294)), Avro+PG.
+- Packaging module ([#290](https://github.com/CategoricalData/hydra/issues/290)): `hydra.packaging` with `Package` type and validation.
+- New primitives: `lists.foldr` ([#280](https://github.com/CategoricalData/hydra/issues/280)), `maybes.toList`, float rounding ([#264](https://github.com/CategoricalData/hydra/issues/264), [#285](https://github.com/CategoricalData/hydra/issues/285)).
+- Cross-language benchmarking suite ([#234](https://github.com/CategoricalData/hydra/issues/234)).
+- Limited typeclass inference ([#164](https://github.com/CategoricalData/hydra/issues/164)).
+- Standardized validation patterns across modules ([#291](https://github.com/CategoricalData/hydra/issues/291)).
+
+### Improvements
+
+- Promoted all staging modules to DSL sources ([#267](https://github.com/CategoricalData/hydra/issues/267)).
+- Removed legacy adapters, JSON/YAML coders, grammar framework ([#236](https://github.com/CategoricalData/hydra/issues/236), [#295](https://github.com/CategoricalData/hydra/issues/295)).
+- Java: efficient immutable data structures ([#193](https://github.com/CategoricalData/hydra/issues/193)); flattened term naming.
+- Java: hidden specialized collection classes in type-level files ([#313](https://github.com/CategoricalData/hydra/issues/313)).
+- Inference error messages improved ([#231](https://github.com/CategoricalData/hydra/issues/231)).
+- Combined nested let bindings in the Haskell coder ([#248](https://github.com/CategoricalData/hydra/issues/248)).
+- Simplified `hydra.adapt.simple` ([#255](https://github.com/CategoricalData/hydra/issues/255)).
+- Removed Haskell operators dependency from tests ([#288](https://github.com/CategoricalData/hydra/issues/288)).
+- Regression tests for non-Haskell/Java coders ([#85](https://github.com/CategoricalData/hydra/issues/85)).
+- Regex support added ([#293](https://github.com/CategoricalData/hydra/issues/293)).
+- `hydra.annotations.isNativeType` revisited ([#214](https://github.com/CategoricalData/hydra/issues/214)).
+- Migrated grammar-generated models ([#295](https://github.com/CategoricalData/hydra/issues/295)).
+- `<<= -shouldBe-> <<~` operator change ([#276](https://github.com/CategoricalData/hydra/issues/276)).
+- Minimize Generation.hs ([#225](https://github.com/CategoricalData/hydra/issues/225)).
+- Implemented missing Flows primitives ([#183](https://github.com/CategoricalData/hydra/issues/183)).
+- Removed JavaDocs from the repository ([#262](https://github.com/CategoricalData/hydra/issues/262)).
+- Floating-point test portability fixes for Linux CI ([#264](https://github.com/CategoricalData/hydra/issues/264), [#285](https://github.com/CategoricalData/hydra/issues/285)).
+- Moved YAML coder to hydra-ext ([#224](https://github.com/CategoricalData/hydra/issues/224)).
+- Minimized native Python and Java primitives in favor of DSL implementations ([#244](https://github.com/CategoricalData/hydra/issues/244)).
+- Removed the `disabledForPython` test tag ([#250](https://github.com/CategoricalData/hydra/issues/250)).
+
+### Release tooling
+
+- 0.14 release tracking ([#304](https://github.com/CategoricalData/hydra/issues/304)).
+
+### Documentation
+
+- Thorough cleanup of user documentation ([#302](https://github.com/CategoricalData/hydra/issues/302), [#303](https://github.com/CategoricalData/hydra/issues/303)).
+- Coding style guide added to wiki.
+- Updated 'new implementation' recipe with learnings from Clojure head ([#278](https://github.com/CategoricalData/hydra/issues/278)).
+
+---
+
+## [0.13.0] - 2026-02-27
+
+Major release completing Hydra-Python ([#66](https://github.com/CategoricalData/hydra/issues/66)) and Hydra-Java ([#166](https://github.com/CategoricalData/hydra/issues/166)) as self-hosting Hydra implementations
+and demonstrating mutual self-hosting across all three languages.
+Significant improvements to type inference, type checking, rewriting, and adaptation.
+New language features include first-class Either types, typeclass inference, binary data support,
+and a native JSON parser/writer.
+Comprehensive tooling for cross-implementation bootstrapping
+and provisional support for Go, Rust, and JavaScript targets.
+
+### Highlights
+
+- **Mutual self-hosting across three implementations**: Hydra is now fully self-hosting
+  in Haskell ([#179](https://github.com/CategoricalData/hydra/issues/179)), Java, and Python, and can cross-generate between every combination
+  of host and target language (9 paths total).
+- **Hydra-Java is complete** ([#166](https://github.com/CategoricalData/hydra/issues/166)): all kernel and generation tests pass.
+- **Hydra-Python is complete** ([#66](https://github.com/CategoricalData/hydra/issues/66)): 100% test parity with Haskell.
+- **Language coders promoted into the Hydra kernel** ([#176](https://github.com/CategoricalData/hydra/issues/176)): Haskell, Java, and Python
+  coders are now defined as Hydra DSL modules, making them self-hosting.
+- **Either type support** ([#210](https://github.com/CategoricalData/hydra/issues/210)) with full inference, checking, and library support.
+- **Native JSON parser and writer** ([#188](https://github.com/CategoricalData/hydra/issues/188), [#243](https://github.com/CategoricalData/hydra/issues/243), [#253](https://github.com/CategoricalData/hydra/issues/253)): Hydra's own JSON parser and writer
+  replace the Aeson dependency; bootstrap-from-JSON architecture for cross-implementation
+  code generation.
+
+### Breaking changes
+
+- Renamed `Optional` to `Maybe` throughout the codebase ([#204](https://github.com/CategoricalData/hydra/issues/204)):
+  `hydra.lib.optionals` → `hydra.lib.maybes`; all `optional` variants renamed to `maybe`.
+- Removed unlabeled product and sum types; replaced with pair types ([#212](https://github.com/CategoricalData/hydra/issues/212)).
+- Renamed `hydra.core.TypedTerm` to `hydra.core.TypeApplicationTerm`.
+- Renamed `hydra.json` to `hydra.json.model`.
+- Renamed `hydra.lib.maps.remove` to `hydra.lib.maps.delete`.
+- Removed deprecated `hydra.decoding` module.
+- Removed old `hydra.mantle.Either` type (replaced by core `Either`).
+- Removed `hydra.lib.tuples` library.
+- Removed C# coder stub (C# syntax model remains).
+- DSL syntax migration: removed `OverloadedStrings` for term expressions ([#238](https://github.com/CategoricalData/hydra/issues/238)).
+- Removed `Hydra.Dsl.ShorthandTypes` ([#222](https://github.com/CategoricalData/hydra/issues/222)).
+
+### New features
+
+- Self-hosting in Haskell ([#179](https://github.com/CategoricalData/hydra/issues/179)) and full mutual cross-generation (Haskell ↔ Java ↔ Python).
+- New bootstrapping demo validating all 9 paths across 249 modules.
+- Promoted Haskell, Java, and Python coders from staging code into Hydra DSL sources ([#176](https://github.com/CategoricalData/hydra/issues/176)).
+- JSON-as-source-of-truth bootstrapping architecture ([#243](https://github.com/CategoricalData/hydra/issues/243), [#253](https://github.com/CategoricalData/hydra/issues/253)):
+  new executables and scripts enabling `bootstrap-from-json` for cross-implementation
+  code generation.
+- Generated term encoders and decoders replace legacy hand-coded modules ([#47](https://github.com/CategoricalData/hydra/issues/47)).
+- Hydra-Python completion ([#66](https://github.com/CategoricalData/hydra/issues/66)): restructured project layout with `src/main` and
+  `src/gen-main` ([#191](https://github.com/CategoricalData/hydra/issues/191)); full library implementations including `hydra.lib.flows`;
+  `@lru_cache` optimization for nullary bindings; PyPy compatibility.
+- Hydra-Java completion ([#166](https://github.com/CategoricalData/hydra/issues/166)): all kernel and generation tests pass; deep changes
+  to Java coder; `Lazy` utility class for delayed evaluation; array-type support
+  in Java serde; tail-call optimization in the Java coder; `Tuple` implements `Comparable`.
+- Either type ([#210](https://github.com/CategoricalData/hydra/issues/210)): `EitherType`, `either` term constructor, `hydra.lib.eithers` library
+  with `either`, `map`, `mapList`, `mapMaybe`, `bimap`, `bind`, `fromLeft`, `fromRight`,
+  `isLeft`, `isRight`, `lefts`, `rights`, `partitionEithers`.
+- Pair type as a core construct ([#211](https://github.com/CategoricalData/hydra/issues/211)).
+- Void type ([#237](https://github.com/CategoricalData/hydra/issues/237)): uninhabited `void` type as the dual of `unit`; type-level only.
+- Native JSON parser and writer ([#188](https://github.com/CategoricalData/hydra/issues/188), [#242](https://github.com/CategoricalData/hydra/issues/242)):
+  bidirectional JSON encoder/decoder; replaces the Aeson dependency;
+  cross-checked via a special test runner.
+- Typeclass inference ([#164](https://github.com/CategoricalData/hydra/issues/164)):
+  type schemes with typeclass constraints via `TypeVariableMetadata`;
+  typeclass metadata on primitive definitions; Haskell coder uses typeclass info.
+- Binary/`ByteString` support ([#172](https://github.com/CategoricalData/hydra/issues/172)): binary data via `ByteString` instead of `String`;
+  Haskell coder uses `Data.ByteString`; new `binaryToBytes` primitive.
+- Parser combinators: new `hydra.parsing` module with combinator types and DSL.
+- Higher-order primitive interpreter ([#198](https://github.com/CategoricalData/hydra/issues/198)): all primitives are fully interpretable
+  across all three implementations; legacy `requiresInterp` tag removed.
+- Let hoisting and flattening: hoisting polymorphic let terms;
+  refined hoisting for monomorphic bindings inside polymorphic ones;
+  more efficient `liftLambdaAboveLet` ([#202](https://github.com/CategoricalData/hydra/issues/202));
+  case-statement hoisting ([#241](https://github.com/CategoricalData/hydra/issues/241)); subterm hoisting helpers.
+- Provisional language targets: Go ([#65](https://github.com/CategoricalData/hydra/issues/65)), Rust (with experimental type-level coder),
+  JavaScript (syntax model, serde, DSL, generated sources), and Java/Python syntax DSLs.
+- New adapter framework ([#236](https://github.com/CategoricalData/hydra/issues/236)): all external coders refactored
+  (Haskell, YAML, Scala, JSON Schema, C++, PDL, Protobuf, GraphQL).
+- Compact labeled records and variants ([#122](https://github.com/CategoricalData/hydra/issues/122)).
+- New primitives:
+  `hydra.lib.lists.find`, `hydra.lib.lists.partition`,
+  `hydra.lib.flows.withDefault`, `hydra.lib.flows.foldl`,
+  `hydra.lib.math.{max,min,abs,pred,signum,succ,even,odd}`,
+  `hydra.lib.pairs.bimap`, `hydra.lib.eithers.bind`,
+  `hydra.lib.literals.binaryToBytes`, complete `readXxx`/`showXxx` families,
+  `hydra.lexical.chooseUniqueName`, `neg` renamed to `negate`,
+  full floating-point math primitives in Python ([#208](https://github.com/CategoricalData/hydra/issues/208)).
+- DeepCore DSL: a layer one level deeper than the Meta DSLs,
+  for programs that construct programs that construct terms.
+- `typeOf` implementation ([#168](https://github.com/CategoricalData/hydra/issues/168)).
+- Implement `Comparable` for generated Java classes ([#131](https://github.com/CategoricalData/hydra/issues/131)); update of Hydra-Java ([#166](https://github.com/CategoricalData/hydra/issues/166)).
+- Compact labeled records exploration ([#122](https://github.com/CategoricalData/hydra/issues/122)).
+- Comparison primitives without typeclasses ([#187](https://github.com/CategoricalData/hydra/issues/187)).
+- Investigated core support for Either and Flow monads ([#200](https://github.com/CategoricalData/hydra/issues/200)).
+
+### Improvements
+
+- Type system:
+  - Refactored `hydra.inference` to create new module `hydra.checking`.
+  - Inference checks after unification to prevent schema names from being unified
+    with inferred type variables.
+  - Type application terms preserved when preparing application terms for Python.
+  - Sanity-check tuple projection index against arity during inference.
+  - Bind lost type variables in term annotations.
+  - Made `extendTypeContextForLet` tolerant of untyped bindings.
+- Code generation:
+  - Python: support for deeply-nested match statements; more complete `let` support;
+    Pythonic syntax for polymorphic function definitions; refinements for complex
+    case-hoisting scenarios; performance improvements ([#209](https://github.com/CategoricalData/hydra/issues/209), [#239](https://github.com/CategoricalData/hydra/issues/239), [#240](https://github.com/CategoricalData/hydra/issues/240)); inline
+    let/case rewriting for Python targets ([#203](https://github.com/CategoricalData/hydra/issues/203)).
+  - Java: method declaration style consistent with calling style.
+  - Haskell: standardized typeclass names.
+  - Increased maximum trace depth to support more complex sources.
+- Common test suite promotion: all tests promoted into a shared, implementation-independent
+  test kernel ([#213](https://github.com/CategoricalData/hydra/issues/213)).
+- Statically-compiled test suite in Haskell ([#207](https://github.com/CategoricalData/hydra/issues/207)).
+- Included all Hydra kernel types in test schema ([#205](https://github.com/CategoricalData/hydra/issues/205)).
+- Reorganized Haskell DSL definitions; standardized imports.
+- Added TODO comments to all unsafe primitives ([#201](https://github.com/CategoricalData/hydra/issues/201)).
+- New `rewriteAndFoldTerm` utility for simultaneous rewriting and folding.
+- Lexical helper for dereferencing schema types through aliases.
+- Made JSON parser lazy with new `lazy` parser combinator.
+- More efficient substitution helpers for empty substitutions.
+- Precomputed type/inference context in kernel test runner for performance.
+- Property graph encoding and decoding modules promoted into DSL.
+- `hydra.tabular` and `hydra.pg.graphson` modules promoted into DSL.
+- `hydra.show.*` modules generated into Java and Python.
+- Aligned epsilon encoding with the LambdaGraph spec ([#89](https://github.com/CategoricalData/hydra/issues/89)).
+- Fixed asymmetry of introductions and eliminations ([#86](https://github.com/CategoricalData/hydra/issues/86), [#134](https://github.com/CategoricalData/hydra/issues/134)).
+- Minimized inferred type annotations ([#114](https://github.com/CategoricalData/hydra/issues/114)).
+- Investigated combining products with records, sums with unions ([#196](https://github.com/CategoricalData/hydra/issues/196), [#212](https://github.com/CategoricalData/hydra/issues/212)).
+- Refined type reduction in adapters ([#144](https://github.com/CategoricalData/hydra/issues/144)).
+- Investigation of unification issues ([#163](https://github.com/CategoricalData/hydra/issues/163)).
+- Refactored RowType ([#82](https://github.com/CategoricalData/hydra/issues/82)).
+- LLM-assisted PG schemas and mappings ([#171](https://github.com/CategoricalData/hydra/issues/171)).
+- Simplified unit-valued variants in Python ([#206](https://github.com/CategoricalData/hydra/issues/206)).
+- Checked for inconsistencies between Term/TTerm and Type/TType DSLs ([#218](https://github.com/CategoricalData/hydra/issues/218)).
+- GenPG demo: refactored to support Haskell, Python, and Java; Python runner;
+  Java generation; finalized output format.
+- Testing: extended test runners for let hoisting (Haskell, Python, Java);
+  floating-point precision tests; ordering tests for `hydra.lib.maps`/`hydra.lib.sets`;
+  inference tests for inferred System F terms; case-hoisting test cases; new JSON coder tests;
+  thorough test cases for all primitives ([#199](https://github.com/CategoricalData/hydra/issues/199)).
+
+### Bug fixes
+
+- Fixed interpreter bugs causing test failures ([#235](https://github.com/CategoricalData/hydra/issues/235)).
+- Fixed bug in type checking in connection with dead code.
+- Fixed bug in `removeTypesFromTerm`.
+- Fixed bug in `typeOfMap`.
+- Fixed 32-bit max int value in `hydra.constants`.
+- Fixed numeric precision in Python.
+- Added `BigDecimal` support and fixed `uint8` (short) support in Java.
+- Fixed Haskell coder with respect to imports for binary literals.
+- Excluded `Prelude.encodeFloat` and `Prelude.decodeFloat` from generated Haskell
+  due to name collisions.
+- Fixed consistency issues in Python primitives:
+  `hydra.lib.sets.toList`, `hydra.lib.maps.union` (precedence),
+  `hydra.lib.strings.lines`, `hydra.lib.strings.readString` (dequoting),
+  `hydra.lib.lists.apply`, `hydra.lib.maps.toList`.
+- Fixed Haskell type signature of `hydra.lists.span`.
+- Fixed issue with eta expansion of typed terms.
+- Fixed shadowing issue in generated encoding modules.
+- Added missing alternatives in `hydra.rewriting.rewriteTermM`.
+
+### Documentation
+
+- Updated main README with complete implementation status.
+- Added Documentation sections to all implementation READMEs.
+- Fixed outdated Code-organization.md (Python now uses `src/gen-main`).
+- Fixed incorrect code paths in Implementation.md.
+- Added Java section to Testing wiki.
+- Comprehensive developer recipes in `docs/recipes/`.
+- Documentation for tail-call optimization implementation.
+- Added Haddock comments to Haskell primitives.
+- Created wiki page on the release process ([#194](https://github.com/CategoricalData/hydra/issues/194)).
+
+### Community
+
+- Accepted babeloff's `isTrivialTerm` changes.
+
+---
+
+## [0.12.0] - 2025-08-28
+
+A consolidation release: significant Python progress, kernel-level cleanup,
+and a wave of bookkeeping closures for Hydra-Java module ports.
+
+### Breaking changes
+
+- Removed `hydra.decoding` module ([#190](https://github.com/CategoricalData/hydra/issues/190)).
+- Flattened dependency tiers; eliminated tier-1, tier-2, tier-3 organization ([#184](https://github.com/CategoricalData/hydra/issues/184)).
+- Removed `typed` term variant ([#173](https://github.com/CategoricalData/hydra/issues/173), [#162](https://github.com/CategoricalData/hydra/issues/162)).
+- Module naming changes: moved tabular and relational modules into kernel ([#152](https://github.com/CategoricalData/hydra/issues/152)).
+- Standardized on uncurried helper functions in term-level DSLs ([#174](https://github.com/CategoricalData/hydra/issues/174)).
+- Renamed `hydra.lib.flows.traverseOptional` to `hydra.lib.flows.mapOptional`.
+- Eliminated camel-cased namespace parts from the kernel ([#152](https://github.com/CategoricalData/hydra/issues/152)).
+- Replaced list and optional eliminations with primitive functions ([#150](https://github.com/CategoricalData/hydra/issues/150)).
+- Moved `Hydra.Ext` to hydra-ext ([#178](https://github.com/CategoricalData/hydra/issues/178)); moved 'ext' sources into hydra-ext ([#189](https://github.com/CategoricalData/hydra/issues/189)).
+
+### New features
+
+- Significant progress toward Hydra-Python completion:
+  generated all kernel types and terms into Python;
+  added `Decimal` support for bigfloat values;
+  environment tracking for variable scoping;
+  generated `hydra.languages` into Python;
+  `__hash__` and `__eq__` on serializable generated classes ([#160](https://github.com/CategoricalData/hydra/issues/160)).
+- Designated `unit` term and type variants ([#186](https://github.com/CategoricalData/hydra/issues/186)).
+- `hydra.lib.flows.mapElems` and `hydra.lib.flows.mapKeys` utilities.
+- Promoted JSON coder into the DSL ([#181](https://github.com/CategoricalData/hydra/issues/181)).
+- Promoted JSON utilities into the DSL ([#181](https://github.com/CategoricalData/hydra/issues/181)).
+- C# coder ([#139](https://github.com/CategoricalData/hydra/issues/139)).
+- GQL model and parser ([#140](https://github.com/CategoricalData/hydra/issues/140)).
+- Tabular adapters ([#142](https://github.com/CategoricalData/hydra/issues/142)).
+- Type-level C++ coder ([#170](https://github.com/CategoricalData/hydra/issues/170)).
+- Variable-types-to-Haskell-typedefs transformation ([#70](https://github.com/CategoricalData/hydra/issues/70)).
+- A newline at the end of each generated file ([#154](https://github.com/CategoricalData/hydra/issues/154)).
+- Refactor kernel modules and primitives by connectivity ([#177](https://github.com/CategoricalData/hydra/issues/177)).
+
+### Improvements
+
+- Condensed repeated elements in error traces ([#165](https://github.com/CategoricalData/hydra/issues/165)).
+- Fixed transitive schema dependencies ([#185](https://github.com/CategoricalData/hydra/issues/185)).
+- Improved adapter solution with respect to literals.
+- Better handling of type variables in Python coder.
+- Enhanced term-level Python generation with proper ordering to minimize forward references.
+- Topo-sort helper function for dependency management.
+- Updated all Python libraries with latest changes.
+- Distinguished between deannotation and detyping of terms.
+- Adapted primitives when applying language constraints to a graph.
+- Use unqualified names where possible in generated code ([#153](https://github.com/CategoricalData/hydra/issues/153)).
+- Eliminated `Data.Graph` (containers) dependency ([#167](https://github.com/CategoricalData/hydra/issues/167)).
+- "Close the loop" demonstration ([#175](https://github.com/CategoricalData/hydra/issues/175)).
+- Implemented `typeOf` ([#168](https://github.com/CategoricalData/hydra/issues/168)).
+
+### Bug fixes
+
+- Fixed System F terms for records and case statements ([#168](https://github.com/CategoricalData/hydra/issues/168)).
+- Fixed issue with transitive type-level module dependencies ([#185](https://github.com/CategoricalData/hydra/issues/185)).
+- Corrected handling of variable references in Python application terms.
+- Fixed nullary functions in Python (added empty parens).
+- Added missing alternatives in `hydra.rewriting.rewriteTermM`.
+
+### Bookkeeping
+
+- Closed long-standing Hydra-Java module-port issues
+  ([#9](https://github.com/CategoricalData/hydra/issues/9), [#11](https://github.com/CategoricalData/hydra/issues/11), [#15](https://github.com/CategoricalData/hydra/issues/15), [#16](https://github.com/CategoricalData/hydra/issues/16), [#18](https://github.com/CategoricalData/hydra/issues/18), [#19](https://github.com/CategoricalData/hydra/issues/19), [#20](https://github.com/CategoricalData/hydra/issues/20), [#21](https://github.com/CategoricalData/hydra/issues/21), [#22](https://github.com/CategoricalData/hydra/issues/22), [#23](https://github.com/CategoricalData/hydra/issues/23), [#25](https://github.com/CategoricalData/hydra/issues/25), [#44](https://github.com/CategoricalData/hydra/issues/44), [#45](https://github.com/CategoricalData/hydra/issues/45));
+  most of the underlying work shipped in earlier releases.
+- Closed Hydra-Java DSL-port issues ([#24](https://github.com/CategoricalData/hydra/issues/24), [#26](https://github.com/CategoricalData/hydra/issues/26), [#27](https://github.com/CategoricalData/hydra/issues/27), [#28](https://github.com/CategoricalData/hydra/issues/28), [#29](https://github.com/CategoricalData/hydra/issues/29)) and
+  the Scala/Avro tracking issues ([#77](https://github.com/CategoricalData/hydra/issues/77), [#81](https://github.com/CategoricalData/hydra/issues/81)); these had been resolved
+  by earlier releases (Scala in 0.14.0, Avro bidi in 0.14.0).
+
+### Documentation
+
+- Updated Hydra-Haskell README.
+- Updated Hydra-Java README with corrected links.
+- Updated JavaDocs.
+
+---
+
+## [0.11.0] - 2025-03-16
+
+Untagged in-repository version.
+A type-inference rebase release: major restructuring of inference internals
+to align with Algorithm W, plus a sweeping removal of explicit type annotations
+across the kernel.
+
+### Highlights
+
+- **Inference rebased on Algorithm W** ([#118](https://github.com/CategoricalData/hydra/issues/118)): unify-early strategy ([#146](https://github.com/CategoricalData/hydra/issues/146)),
+  infinite-type checks restored ([#116](https://github.com/CategoricalData/hydra/issues/116)), recursive-element inference ([#90](https://github.com/CategoricalData/hydra/issues/90)),
+  unified handling of elements and let bindings ([#112](https://github.com/CategoricalData/hydra/issues/112)), minimized manual
+  type annotations ([#119](https://github.com/CategoricalData/hydra/issues/119)).
+- **`hydra.inference` module** introduced for inference type definitions;
+  substitution and unification moved up a level.
+
+### Breaking changes
+
+- Argument order of `Logic.ifElse` changed ([#147](https://github.com/CategoricalData/hydra/issues/147)).
+
+### New features
+
+- Inference test cases promoted into the common test suite ([#148](https://github.com/CategoricalData/hydra/issues/148));
+  added "kernel examples" inference tests; more recursion / mutual recursion cases.
+
+### Improvements
+
+- Removed type annotations across all kernel sources ([#119](https://github.com/CategoricalData/hydra/issues/119)).
+- Migrated to fully-applied primitive library DSLs ([#157](https://github.com/CategoricalData/hydra/issues/157)).
+- Aligned `Graph` and `Element` with `LetBinding` ([#159](https://github.com/CategoricalData/hydra/issues/159)).
+- `hydra.core.Unit` recognized by the Haskell coder.
+- Refactored inference tests.
+
+### Bug fixes
+
+- Fix for underdetermined type in `hydra.decode`.
+
+---
+
+## [0.10.0] - 2025-02-19
+
+Untagged in-repository version.
+Adds Hydra-Python (initial completion), a GraphSON coder,
+provisional Hydra-Go support, and dot-separated namespace conventions.
+
+### Highlights
+
+- **Hydra-Python initial completion** ([#80](https://github.com/CategoricalData/hydra/issues/80), [#66](https://github.com/CategoricalData/hydra/issues/66)): Python coder, generated kernel
+  modules in Python, and full library implementations across kernel namespaces.
+- **GraphSON coder** ([#79](https://github.com/CategoricalData/hydra/issues/79)).
+- **Provisional Hydra-Go** ([#65](https://github.com/CategoricalData/hydra/issues/65)).
+- **Namespace formatting** standardized to dot-separated form ([#151](https://github.com/CategoricalData/hydra/issues/151)).
+
+### Breaking changes
+
+- All namespaces use dot-separated form ([#151](https://github.com/CategoricalData/hydra/issues/151)).
+
+### New features
+
+- Hydra-Python: Python coder; generated `hydra.coreEncoding`, `hydra.strip`,
+  `hydra.literals` and other kernel modules in Python; tuple-expression support;
+  updated `hydra.constants` in Python.
+- GraphSON coder ([#79](https://github.com/CategoricalData/hydra/issues/79)).
+- Provisional Hydra-Go ([#65](https://github.com/CategoricalData/hydra/issues/65)).
+- `string-to-char-list` and `char-list-to-string` primitives ([#149](https://github.com/CategoricalData/hydra/issues/149)).
+- Restored normalization for generated term variables ([#145](https://github.com/CategoricalData/hydra/issues/145)).
+
+### Improvements
+
+- Re-leveled dependency tiers ([#135](https://github.com/CategoricalData/hydra/issues/135)).
+- Updated all Java, Haskell, hydra-ext, and Python primitives to current sources.
+
+---
+
+## [0.9.0] - 2025-01-06
+
+Untagged in-repository version.
+Bootstraps Hydra-Python and the JSON Schema coder; introduces a fluent Java DSL for flows.
+
+### Highlights
+
+- **JSON Schema coder** ([#141](https://github.com/CategoricalData/hydra/issues/141)).
+- **Python coder scaffold**: serde, language constraints, and generated Haskell
+  sources for Python.
+- **Fluent-style Java DSL for flows** ([#143](https://github.com/CategoricalData/hydra/issues/143)).
+
+### New features
+
+- JSON Schema coder ([#141](https://github.com/CategoricalData/hydra/issues/141)).
+- Python coder scaffold and serde.
+- Provisional Python language constraints module.
+- Fluent-style Java DSL for flows ([#143](https://github.com/CategoricalData/hydra/issues/143)); moved Flows DSL into `hydra.dsl`.
+- New `hydra.coders` DSL.
+- New DSL functions for core variants.
+
+### Improvements
+
+- Generalized `Namespaces` so it can be used for languages other than Haskell.
+- Local-name-only Python type aliases (the directory tree provides the namespace).
+- Coders DSL used to simplify Java language-constraint sources.
+- Removed redundant productions in the Python grammar.
+- Added unit tests for fluent flows.
+
+### Documentation
+
+- Updated JavaDocs for 0.9.0.
+
+---
+
+## [0.8.1] - 2024-09-24
+
+Patch release.
+
+### Improvements
+
+- Generalized the property graph merging utility to accept type systems other than Hydra Core.
+
+---
+
+## [0.8.0] - 2024-09-09
+
+C# support ([#139](https://github.com/CategoricalData/hydra/issues/139)), Graphviz DOT coder, term accessors, and a small breaking change in Java
+to support Spark.
+
+### Breaking changes
+
+- Made `FlowException` serializable for Spark support (may affect existing error handling).
+
+### New features
+
+- C# syntax module based on the Microsoft ANTLR grammar ([#139](https://github.com/CategoricalData/hydra/issues/139)).
+- Graphviz DOT coder ([#136](https://github.com/CategoricalData/hydra/issues/136)):
+  full support for lambdas and recursive `let` statements;
+  compact 'accessor graph' visualization option;
+  customizable label styles;
+  highlighted let-bound terms; edge labels.
+- Term accessor type to facilitate lenses and path-aware transformations.
+- OpenGQL grammar and generated Haskell for the OpenGQL model ([#140](https://github.com/CategoricalData/hydra/issues/140)).
+
+### Improvements
+
+- Added `</>` as alternative to `@@` application operator in Haskell DSLs.
+- Utilities for working with term accessors.
+- More specific exception class for JSON decoding.
+- Additional convenience methods on `JsonDecoding.java`.
+- Used transitive dependencies in hydra-java and hydra-ext.
+- Upgraded ANTLR to address a vulnerability.
+- Moved TinkerPop utilities from `hydra/ext/org/apache/tinkerpop` into `hydra/pg`.
+
+---
+
+## [0.7.0] - 2024-08-21
+
+Major refactoring of module organization and namespace management.
+Establishes the DNS-style namespace convention in hydra-ext
+and lifts property-graph models into a top-level `hydra/pg` namespace.
+
+### Breaking changes
+
+- Namespace reorganization: migrated to DNS-based module naming in hydra-ext ([#138](https://github.com/CategoricalData/hydra/issues/138)):
+  `hydra/ext/avro` → `hydra/ext/org/apache/avro`,
+  `hydra/ext/graphql` → `hydra/ext/org/graphql`,
+  `hydra/ext/yaml` → `hydra/ext/org/yaml`,
+  `hydra/ext/json/decoding` → `hydra/ext/org/json/decoding`,
+  `hydra/ext/rdf` → `hydra/ext/org/w3/rdf`,
+  `hydra/ext/shacl` → `hydra/ext/org/w3/shacl`,
+  `hydra/ext/tinkerpop` → `hydra/ext/org/apache/tinkerpop`.
+- Property graphs: promoted property graph modules from
+  `hydra/ext/org/apache/tinkerpop` to `hydra/pg`.
+- Project structure: renamed `hydra-extensions` to `hydra-ext`;
+  renamed `hydra/langs` to `hydra/ext` ([#138](https://github.com/CategoricalData/hydra/issues/138));
+  made hydra-ext into a Gradle subproject alongside hydra-java;
+  added top-level Gradle build.
+
+### Improvements
+
+- Moved TinkerPop modules from hydra-haskell into hydra-ext for better separation.
+- Moved miscellaneous models to `Hydra.Ext.Other` ([#138](https://github.com/CategoricalData/hydra/issues/138)).
+- Moved XML Schema, SQL, ShEx, and OWL models to hydra-ext.
+- Moved KQL, Parquet, and Python modules to hydra-ext.
+- Separated JavaDocs for hydra-java and hydra-ext.
+- Updated publishing configuration in `build.gradle`.
+
+---
+
+## [0.6.0] - 2024-08-19
+
+Type-system improvements (System F type abstraction/application; eliminated polymorphic
+types from Hydra Core), expanded language support (Graphviz DOT model, generated Java
+field-name constants), and revamped annotation handling.
+
+### Breaking changes
+
+- Eliminated `InferenceContext` helper type ([#103](https://github.com/CategoricalData/hydra/issues/103)).
+- Changed typing environment from `Map Name Type` to `Map Name TypeScheme` in graphs ([#76](https://github.com/CategoricalData/hydra/issues/76)).
+- Updated annotations and flows to use `map<Name, Term>` instead of `map<string, Term>` ([#133](https://github.com/CategoricalData/hydra/issues/133)).
+- Unified `Name` and `FieldName`; aliased to `string` ([#121](https://github.com/CategoricalData/hydra/issues/121)).
+- Renamed `hydra/core.UnitType` to `hydra/core.Unit` ([#123](https://github.com/CategoricalData/hydra/issues/123)).
+- Eliminated annotation classes; Hydra has a single built-in notion of annotations ([#113](https://github.com/CategoricalData/hydra/issues/113)).
+- Unified the `Nominal` type with `RowType` ([#115](https://github.com/CategoricalData/hydra/issues/115)).
+- Removed the `extends` parameter for row types ([#132](https://github.com/CategoricalData/hydra/issues/132)).
+- Replaced `hydra/Kv` with `hydra/Annotations` (rename revisited).
+
+### New features
+
+- System F support: `typeAbstraction` and `typeApplication` term constructors.
+- Python3 syntax model based on the official Python BNF ([#80](https://github.com/CategoricalData/hydra/issues/80)).
+- Graphviz DOT model and initial support ([#136](https://github.com/CategoricalData/hydra/issues/136)).
+- Field-name constants generated in Java code generation ([#137](https://github.com/CategoricalData/hydra/issues/137)).
+- Provisional Algorithm W implementation in Haskell (originally from @wisnesky), kept for reference.
+
+### Improvements
+
+- Primitive organization: Hydra primitives organized into libraries for namespace management.
+- Code generation:
+  - Escaped field name constants and `with_` methods in Java coder.
+  - Generated field name constants in Java ([#137](https://github.com/CategoricalData/hydra/issues/137)).
+  - Avoided duplicated comments in wrapper classes in generated Java.
+  - Removed superfluous newtype comments in Haskell coder.
+- Updated Delta Parquet model to follow the Java API more closely.
+- Enriched OpenCypher features module with the complete list of standard Cypher functions.
+- Added convenience methods to `JsonEncoding.java`.
+- Restored `Hydra.Inference` subdirectory.
+- Allowed arbitrary case for element names ([#7](https://github.com/CategoricalData/hydra/issues/7)).
+- Corrected `uint8` representation in Java to `Short` ([#120](https://github.com/CategoricalData/hydra/issues/120)).
+
+### Bug fixes
+
+- Fixed handling of type annotations in term adapters.
+- Fixed decoding of encoded terms annotated with a type.
+- Corrected encoding of universal types in `showTerm` ([#117](https://github.com/CategoricalData/hydra/issues/117)).
+
+---
+
+## [0.5.3] - 2024-08-06
+
+Patch release.
+Preserves order of map key/value pairs and set elements during JSON serialization.
+
+### Bug fixes
+
+- Preserved ordering of key/value pairs and set elements in JSON output.
+
+### New features
+
+- Polymorphic `equal` primitive in Haskell and Java.
+- Generated Hydra type definitions together with native instantiations of Hydra types
+  in Haskell (a stepping stone toward generated coders, [#47](https://github.com/CategoricalData/hydra/issues/47)).
+
+---
+
+## [0.5.1] - 2024-08-01
+
+Patch release.
+Dependency updates and minor improvements following the 0.5.0 Hydra Core overhaul.
+
+### Improvements
+
+- Updated dependencies.
+
+---
+
+## [0.5.0] - 2024-07-26
+
+Untagged in-repository version.
+Major Hydra Core overhaul: eliminates polymorphic types in favor of explicit
+TypeScheme; consolidates the annotation system; revamps `Let` term encoding.
+
+### Breaking changes
+
+- Eliminated polymorphic types from Hydra Core in Haskell ([#125](https://github.com/CategoricalData/hydra/issues/125)):
+  removed type parameters from `hydra/core` types in Hydra-Java; replaced type schemes
+  with forall types ([#76](https://github.com/CategoricalData/hydra/issues/76)).
+- Revamped core encoding and `Let`: bindings are now a list of
+  `(name, term, optional type scheme)` triplets instead of a map.
+- Unified typed-term types and moved `TypeScheme` into Hydra Core.
+- Eliminated `AnnotationClass`; Hydra has a single built-in notion of annotations now ([#113](https://github.com/CategoricalData/hydra/issues/113)).
+- Renamed and refactored phantom-type wrappers ([#128](https://github.com/CategoricalData/hydra/issues/128)); removed deprecated `Reference`
+  type from `hydra/phantoms`.
+- Removed incomplete support for stream types and stream terms ([#88](https://github.com/CategoricalData/hydra/issues/88)).
+
+### New features
+
+- `LiteralTypes` DSL module so literals can be built independently of terms.
+- `Hydra.Minimal` module to facilitate collaboration on Hydra application prototypes.
+- `TypeConstraint` type added to the kernel.
+- `foldl` primitive in Haskell and Java; DSL uses `foldl` instead of the built-in
+  list elimination term by default.
+
+### Improvements
+
+- Improvements to Java and JSON coders; improved untyped JSON coder.
+- Pre-order and post-order term-traversal test cases added.
+- Refactored type inference tests.
+- Restored generated Java (Core through Tier 3) with only a few manual tweaks.
+- Element names allow arbitrary case ([#7](https://github.com/CategoricalData/hydra/issues/7)).
+
+---
+
+## [0.4.0] - 2024-07-10
+
+Untagged in-repository version.
+Adds null checks to generated Java (breaking, but justified — Hydra-Java disallows nulls),
+the Gremlin model, and Protobuf annotation refinements.
+
+### Breaking changes
+
+- Null checks added to constructors and `withXXX()` methods in generated Java;
+  null values are no longer permitted in Hydra-Java.
+- Removed Java 8 restriction (raised minimum Java version).
+
+### New features
+
+- Gremlin model ([#127](https://github.com/CategoricalData/hydra/issues/127)); generated Haskell for Gremlin.
+- `hydra/lib/lists.safeHead` primitive in Haskell (necessary in Java due to eager evaluation).
+- Support for Protobuf `deprecated=true` annotation; non-string Protobuf options.
+- Java options in generated Protobuf.
+
+### Improvements
+
+- Used `java.util.Objects.requireNonNull` instead of a custom null check in generated Java.
+- Added 'automatically generated file' disclaimer to generated Java files.
+- Added autogen comments and null checks to (most) generated Java.
+- Added necessary schema-level dependencies for built-in modules.
+- Added null checks to the Flows DSL in Java.
+
+---
+
+## [0.3.0] - 2024-06-05
+
+Untagged in-repository version.
+Adds an OpenCypher parser based on an ANTLR grammar ([#124](https://github.com/CategoricalData/hydra/issues/124)),
+a KQL (Kusto Query Language) model and serializer,
+a Delta Parquet model,
+and an early property graph queries model in the style of Cypher.
+
+### New features
+
+- OpenCypher parser based on the OpenCypher M23 ANTLR grammar ([#124](https://github.com/CategoricalData/hydra/issues/124));
+  ANTLR added as a build dependency;
+  large Cypher test suite drawn from the Cypher Manual.
+- Cypher-to-PG-query transformer (Java; partially implemented).
+- Property graph queries DSL and generated property graph query classes in Java.
+- Property graph queries model (Cypher / future GQL style).
+- KQL (Kusto Query Language) initial model and basic serializer.
+- Delta Parquet model with generated sources.
+
+### Improvements
+
+- Java options and `deprecated=true` support in generated Protobuf.
+- Convenience methods in Java.
+
+---
+
+## [0.2.0] - 2024-01-10
+
+Hydra-Java 0.2.0 release.
+Significant Java tooling and DSL build-out, Protobuf bidirectional coder,
+Python3 syntax model ([#80](https://github.com/CategoricalData/hydra/issues/80)), property graph validation ([#100](https://github.com/CategoricalData/hydra/issues/100)),
+JSON bidirectional support in Java ([#104](https://github.com/CategoricalData/hydra/issues/104)),
+OpenCypher initial model ([#108](https://github.com/CategoricalData/hydra/issues/108)),
+Cypher unification, and the start of the typed-DSL story in Haskell.
+
+### New features
+
+- Property graph validation: validator type, optionality on property types,
+  customizable id/property handling in the merging adapter, fine-tuned validation,
+  comprehensive unit tests in Java ([#100](https://github.com/CategoricalData/hydra/issues/100)).
+- Bidirectional JSON serialization/deserialization in Java ([#104](https://github.com/CategoricalData/hydra/issues/104)).
+- Checkstyle configuration based on Google style guide ([#111](https://github.com/CategoricalData/hydra/issues/111)).
+- Python3 syntax model based on the official Python BNF ([#80](https://github.com/CategoricalData/hydra/issues/80)).
+- Initial OpenCypher model and parser scaffold ([#108](https://github.com/CategoricalData/hydra/issues/108)).
+- Protobuf coder ([#99](https://github.com/CategoricalData/hydra/issues/99)): type encoder, field numbering, formatting, deprecated annotation,
+  Java options, non-string options.
+- Adapter framework tolerance of recursive types ([#58](https://github.com/CategoricalData/hydra/issues/58)); default branch for case
+  statements ([#60](https://github.com/CategoricalData/hydra/issues/60)).
+- Topological sort in type inference ([#67](https://github.com/CategoricalData/hydra/issues/67)).
+- `hydra/query` model ([#73](https://github.com/CategoricalData/hydra/issues/73)).
+- Records-to-TinkerPop-elements coder ([#64](https://github.com/CategoricalData/hydra/issues/64)).
+- Equality primitives for all literal types ([#92](https://github.com/CategoricalData/hydra/issues/92)); polymorphic logic primitives ([#95](https://github.com/CategoricalData/hydra/issues/95)).
+- `mapKeys` primitive ([#93](https://github.com/CategoricalData/hydra/issues/93)).
+- Tuple support in Java ([#94](https://github.com/CategoricalData/hydra/issues/94)).
+- Java 8 as an optional target ([#96](https://github.com/CategoricalData/hydra/issues/96)); optionally generate Java classes as `Serializable` ([#97](https://github.com/CategoricalData/hydra/issues/97)).
+- Bidirectional "merged vertex" and "merged edge" coders ([#106](https://github.com/CategoricalData/hydra/issues/106)).
+- Nullability and parameterization in the tabular model ([#107](https://github.com/CategoricalData/hydra/issues/107)).
+- Term-level / type-level dependency distinction for modules ([#109](https://github.com/CategoricalData/hydra/issues/109)).
+- Flatten nested `let` terms for transformation into Java ([#110](https://github.com/CategoricalData/hydra/issues/110)).
+- Various `Flow` utilities in Java: `bind3`; convenience methods for consuming results;
+  `fromFlow` variants including one that throws.
+
+### Improvements
+
+- Hydra-Java module ports ([#10](https://github.com/CategoricalData/hydra/issues/10), [#12](https://github.com/CategoricalData/hydra/issues/12), [#13](https://github.com/CategoricalData/hydra/issues/13), [#14](https://github.com/CategoricalData/hydra/issues/14), [#17](https://github.com/CategoricalData/hydra/issues/17), [#30](https://github.com/CategoricalData/hydra/issues/30), [#31](https://github.com/CategoricalData/hydra/issues/31), [#32](https://github.com/CategoricalData/hydra/issues/32), [#33](https://github.com/CategoricalData/hydra/issues/33), [#34](https://github.com/CategoricalData/hydra/issues/34), [#35](https://github.com/CategoricalData/hydra/issues/35),
+  [#38](https://github.com/CategoricalData/hydra/issues/38), [#39](https://github.com/CategoricalData/hydra/issues/39), [#41](https://github.com/CategoricalData/hydra/issues/41), [#43](https://github.com/CategoricalData/hydra/issues/43), [#48](https://github.com/CategoricalData/hydra/issues/48), [#71](https://github.com/CategoricalData/hydra/issues/71), [#72](https://github.com/CategoricalData/hydra/issues/72), [#75](https://github.com/CategoricalData/hydra/issues/75), [#83](https://github.com/CategoricalData/hydra/issues/83)):
+  `Common`, `CoreEncoding`, `CoreLanguage`, `Kernel`, `Flows`, `Lists`, `Literals`,
+  `Math`, `Sets`, `Strings`, `Optionals`, `Maps`, `Flows` primitives, `Sets` primitives.
+- Common DSLs in Haskell: Flow support ([#68](https://github.com/CategoricalData/hydra/issues/68)), Let support ([#69](https://github.com/CategoricalData/hydra/issues/69)).
+- Renamed annotation type parameter `m` to `a` ([#41](https://github.com/CategoricalData/hydra/issues/41)).
+- Updated wrapper, element, and variable types ([#75](https://github.com/CategoricalData/hydra/issues/75)).
+- Removed `schema` field from `Element` ([#83](https://github.com/CategoricalData/hydra/issues/83)).
+- GraphQL coder ([#78](https://github.com/CategoricalData/hydra/issues/78)).
+- Code quality:
+  - Updated main and test sources to conform to Checkstyle configuration ([#111](https://github.com/CategoricalData/hydra/issues/111)).
+  - Extended line-length limit from 100 to 120 characters.
+  - Disabled overly restrictive indentation and import-order rules.
+  - Disabled `RightCurlyAlone` for compact inline map definitions.
+- Streamlined property-graph element merging to unify properties with identical
+  keys and types ([#106](https://github.com/CategoricalData/hydra/issues/106)).
+- Convenience functions for printing literals and literal types in the Java DSL.
+- Minor addition to JSON decoding.
+- DSL convenience: alphabetic-case flexibility for element names ([#7](https://github.com/CategoricalData/hydra/issues/7)).
+
+### Bug fixes
+
+- Fixed type annotations for the Sets Java port ([#48](https://github.com/CategoricalData/hydra/issues/48)).
+- Various minor fixes during the Java module ports.
+
+---
+
+## [0.1.1] - 2022-12-04
+
+Patch release the same day as 0.1.0.
+README polish only.
+
+---
+
+## [0.1.0] - 2022-12-04
+
+First packaged release of Hydra (Hackage).
+This release contains the complete foundation of Hydra.
+
+### Core language
+
+- Hydra's core type and data languages (`hydra.core`).
+- Core models for graphs and modules.
+- Computation model with the `Flow` monad.
+- BNF grammars support.
+- Phantom types.
+- Basic operations on types and terms.
+
+### Type system
+
+- Hindley–Milner-style type inference.
+- Type schemes with polymorphism.
+- Type/term validation.
+
+### Transformations
+
+- Adapter system: type-to-type rewriting and transformation.
+- Coders:
+  Haskell coder (types and terms),
+  Java coder (types and terms),
+  partial Scala coder (terms only),
+  Avro coder,
+  JSON coder,
+  PDL (Pegasus Data Language) coder,
+  RDF + SHACL coder,
+  YAML coder.
+
+### Language support
+
+- Models for GraphQL, OWL (Web Ontology Language), ShEx (Shape Expressions),
+  and TinkerPop-style property graphs.
+
+### Developer tools
+
+- Type-construction DSL and term-construction DSL.
+- QuickCheck property-based tests.
+
+### Implementations
+
+- Hydra-Haskell: bootstrapping implementation with full kernel.
+- Hydra-Java: Java implementation with mature tooling.
+
+### Community
+
+- Set up the LambdaGraph Discord server ([#2](https://github.com/CategoricalData/hydra/issues/2)).
+
+---
+
+## Contributing
+
+We welcome contributions! Please see:
+
+- [Developer recipes](https://github.com/CategoricalData/hydra/tree/main/docs/recipes)
+- [LambdaGraph Discord](https://bit.ly/lg-discord)
+
+## Release process
+
+See the [Release policy](https://github.com/CategoricalData/hydra/wiki/Release-policy)
+wiki page and [docs/release-workflow.md](docs/release-workflow.md)
+for information on how Hydra releases are versioned, built, and published.
diff --git a/LICENSE b/LICENSE
new file mode 100644
--- /dev/null
+++ b/LICENSE
@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
diff --git a/NOTICE b/NOTICE
new file mode 100644
--- /dev/null
+++ b/NOTICE
@@ -0,0 +1,3 @@
+Hydra
+Copyright 2021-2026 The Hydra authors
+https://github.com/CategoricalData/hydra
diff --git a/hydra-jvm.cabal b/hydra-jvm.cabal
new file mode 100644
--- /dev/null
+++ b/hydra-jvm.cabal
@@ -0,0 +1,38 @@
+cabal-version: 1.12
+
+-- This file has been generated from package.yaml by hpack version 0.39.1.
+--
+-- see: https://github.com/sol/hpack
+
+name:           hydra-jvm
+version:        0.17.0
+synopsis:       Shared JVM support for Hydra's Java, Scala, and Clojure coders
+description:    Hydra is an implementation of the LambdaGraph data model, which takes advantage of an isomorphism between labeled hypergraphs and typed lambda calculus: in Hydra, "graphs are programs, and programs are graphs". Common JVM support for Hydra (shared by hydra-java, hydra-scala, hydra-clojure)
+category:       Data
+homepage:       https://github.com/CategoricalData/hydra#readme
+bug-reports:    https://github.com/CategoricalData/hydra/issues
+author:         Joshua Shinavier <josh@fortytwo.net>
+maintainer:     Joshua Shinavier <josh@fortytwo.net>
+license:        Apache-2.0
+license-file:   LICENSE
+build-type:     Simple
+extra-source-files:
+    LICENSE
+    NOTICE
+    CHANGELOG.md
+
+source-repository head
+  type: git
+  location: https://github.com/CategoricalData/hydra
+
+library
+  exposed-modules:
+      Hydra.Jvm.Serde
+  other-modules:
+      Paths_hydra_jvm
+  hs-source-dirs:
+      src/main/haskell
+  build-depends:
+      base >=4.19.0 && <4.22
+    , hydra-kernel ==0.17.0
+  default-language: Haskell2010
diff --git a/src/main/haskell/Hydra/Jvm/Serde.hs b/src/main/haskell/Hydra/Jvm/Serde.hs
new file mode 100644
--- /dev/null
+++ b/src/main/haskell/Hydra/Jvm/Serde.hs
@@ -0,0 +1,50 @@
+-- Note: this is an automatically generated file. Do not edit.
+-- | Common JVM serialization helpers: Java/Scala string and character escaping
+
+module Hydra.Jvm.Serde where
+import qualified Hydra.Overlay.Haskell.Lib.Lists as Lists
+import qualified Hydra.Overlay.Haskell.Lib.Logic as Logic
+import qualified Hydra.Overlay.Haskell.Lib.Strings as Strings
+import Prelude hiding  (Enum, Ordering, decodeFloat, encodeFloat, fail, map, pure, sum)
+
+-- | Escape a single character for use in a Java/Scala string or character literal
+escapeJavaChar :: Int -> String
+escapeJavaChar c =
+    Logic.ifElse (c == 34) "\\\""
+    (Logic.ifElse (c == 92) "\\\\"
+    (Logic.ifElse (c == 10) "\\n"
+    (Logic.ifElse (c == 13) "\\r"
+    (Logic.ifElse (c == 9)  "\\t"
+    (Logic.ifElse (c == 8)  "\\b"
+    (Logic.ifElse (c == 12) "\\f"
+    (Logic.ifElse (c >= 32 && c < 127) (Strings.fromList [c])
+    (javaUnicodeEscape c))))))))
+
+-- | Escape a string for use in a Java/Scala string literal
+escapeJavaString :: String -> String
+escapeJavaString s = Strings.cat (Lists.map escapeJavaChar (Strings.toList s))
+
+-- | Convert an integer in [0,15] to the ASCII code of the corresponding hex digit
+hexDigit :: Int -> Int
+hexDigit n = Logic.ifElse (n < 10) (n + 48) (n - 10 + 65)
+
+-- | Produce a \\uXXXX Unicode escape for the given code point
+javaUnicodeEscape :: Int -> String
+javaUnicodeEscape n =
+    Logic.ifElse (n > 65535)
+      (let n' = n - 65536
+           hi = 55296 + (n' `div` 1024)
+           lo = 56320 + (n' `mod` 1024)
+       in "\\u" ++ padHex4 hi ++ "\\u" ++ padHex4 lo)
+      ("\\u" ++ padHex4 n)
+
+-- | Pad an integer to 4 hex digits
+padHex4 :: Int -> String
+padHex4 n =
+    let d3 = n `div` 4096
+        r3 = n `mod` 4096
+        d2 = r3 `div` 256
+        r2 = r3 `mod` 256
+        d1 = r2 `div` 16
+        d0 = r2 `mod` 16
+    in Strings.fromList [hexDigit d3, hexDigit d2, hexDigit d1, hexDigit d0]