tech-radar: 5 rows from Amara 8th-ferry recommendations (semantic hashing / LSH / HNSW / PQ / quantum illumination / Substrait)

docs/TECH-RADAR.md

@@ -54,6 +54,11 @@ ThoughtWorks-style radar for the technologies / research / papers
 | F\* extraction to F# | Assess | 35 | Successor path after the LiquidF# Hold. F\* is actively maintained and can extract to F#; a 2-3 week PoC on `FastCdc.fs` is the proposed next move for the off-by-one / bad-index bug class. See `docs/research/liquidfsharp-findings.md` §"Path A". |
 | Dafny / F* / Isabelle / Stainless / P#  | Assess | 18 | Enumerated in `docs/research/proof-tool-coverage.md`; each catches a different bug class. |
 | Category theory as code-contract grammar | Adopt | 12 | `docs/category-theory/` |
+| Semantic hashing | Assess | — | Hinton & Salakhutdinov — maps semantically similar documents to nearby binary-hash addresses. Proposed by Amara 8th ferry (PR #274) as real technical family for the "rainbow table" intuition; not the password kind. Candidate substrate for the provenance-aware-bullshit-detector research doc. See `docs/aurora/2026-04-23-amara-physics-analogies-semantic-indexing-cutting-edge-gaps-8th-ferry.md`. |
+| Locality-sensitive hashing (LSH) | Assess | — | Charikar — formal collision framework where similarity drives hash agreement. Sibling to semantic hashing; complementary mechanism. Proposed by Amara 8th ferry for the semantic-canonicalization research doc spine. |
+| HNSW (Hierarchical Navigable Small World) | Assess | — | Graph-based approximate nearest-neighbour index with logarithmic scaling + strong empirical performance. Candidate retrieval structure for the provenance-aware-bullshit-detector if a prototype lands. Proposed by Amara 8th ferry; `Trial` promotion contingent on prototype evidence. |
+| Product quantization | Assess | — | Compressed vector search at scale; memory-efficient large corpora. Optional compression layer under HNSW / ANN retrieval. Proposed by Amara 8th ferry. |
+| Quantum illumination (low-SNR sensing theory) | Assess | — | Lloyd 2008 + Tan et al. Gaussian-state 6 dB error-exponent advantage. Importable as **analogy for low-SNR software detection with retained-reference-path**, NOT as operational quantum-radar capability. 2024 engineering review (Amara 8th ferry) caps microwave QR range at <1 km typical — **Hold for long-range product claims**. Composes with SD-9 carrier-aware framing. See `docs/aurora/2026-04-23-amara-physics-analogies-semantic-indexing-cutting-edge-gaps-8th-ferry.md` §Quantum-radar-analogy-boundaries. |
 
 ### Tools / infra
 
@@ -78,6 +83,7 @@ ThoughtWorks-style radar for the technologies / research / papers
 | `../scratch` declarative-bootstrap harness (package manifests per ecosystem, profile/category composition, mise-unified runtimes, docker reproductions of GHA runners) | Assess | 39 | Named ethos-reference in two P1 BACKLOG entries (env-parity + CI meta-loop) without explicit pattern-inheritance contract. Time-budgeted ~1.5d research pass to classify patterns (already-in-Zeta / worth-porting / scratch-specific / flow-other-way) and produce `docs/research/scratch-zeta-parity.md`. See BACKLOG P1 "`../scratch` ↔ `Zeta` declarative-bootstrap parity". |
 | Declarative environment-parity stack (Argo CD / Flux / Kustomize / Helm / Pulumi / Crossplane / Tilt / Skaffold / Okteto / KCL / CUE / OPA-Gatekeeper / Kyverno candidates) | Assess | 39 | **Time-budgeted research pass.** Aaron 2026-04-20 ask: same declarative spec valid from dev-inner-loop (kind) through qa/dev/stage/prod, non-bespoke. Budget: 7 days split 1d landscape scan -> 3d shortlist deep-dive -> 2d env-parity finalist evaluation -> 1d synthesis ADR. Individual tools graduate to Trial/Adopt/Hold per finalist evaluation. See `docs/BACKLOG.md` P1 "Declarative parity across dev-inner-loop / qa / dev / stage / prod" for the scope; sibling P1 entry on CI meta-loop + retractable CD. |
 | bun + TypeScript (post-setup scripting default) | Trial | 43 | Round-43 adoption for post-`dotnet`-setup scripting. Replaces bash-per-script drift (the round-43 `tally.sh` pivot exposed the cost). First in-tree artefact: `tools/invariant-substrates/tally.ts` (bun Glob + `node:fs/promises`, no third-party runtime deps). Scaffold: `package.json` (pinned `bun@1.3.13`, bun-runtime only — `"type": "module"`, no `tsx`/`ts-node`), `tsconfig.json` cranked-to-11 (strict + `noUncheckedIndexedAccess` + `exactOptionalPropertyTypes` + `verbatimModuleSyntax` + `erasableSyntaxOnly` + `isolatedModules` + `noImplicitOverride`), `eslint.config.ts` strict-type-checked + stylistic-type-checked + sonarjs + `reportUnusedDisableDirectives`, `prettier` + `prettier-plugin-toml`, `markdownlint-cli2`. Excludes hardened across `tsconfig.json` / `bunfig.toml` / `eslint.config.ts` / `.prettierignore` to skip `references/upstreams/**` (~13 GB), `tools/lean4/.lake/**` (~7 GB), `tools/alloy/**`, `tools/tla/**`, `**/BenchmarkDotNet.Artifacts/**`, `**/TestResults/**`, `**/artifacts/**`, `**/bin/**`, `**/obj/**` — SQLSharp lesson: missing the upstream allowlist destroyed language-server perf until the `defaultRepoPathIgnorePatterns` helper shape was adopted. Latest-version audit done at ADR time (`docs/DECISIONS/2026-04-20-tools-scripting-language.md` §Latest-version audit). **Watchlist** (rails for the assumptions): (a) bun on Windows-from-source still maturing — Zeta maintainers are macOS/Linux today, re-audit if Windows becomes a CI target; (b) `erasableSyntaxOnly` presumes bun runs `.ts` without JS emit — re-audit if bun ever ships a JS-compilation pass; (c) `@types/bun 1.3.12` lags the `bun@1.3.13` runtime by one patch (DefinitelyTyped lag, not a bug), exception tracked inline in ADR. Graduates to Adopt once (1) a second in-tree `.ts` script lands, (2) no watchlist triggers fire for 5 rounds, (3) tally.ts + at least one other tool are invoked on round-close. |
+| Substrait (cross-language serialised relational-algebra plan format) | Assess | — | Answers a real P2 gap named in Amara 8th ferry (PR #274): Zeta's persistable query IR (IQbservable / Reaqtor-style Bonsai slim IR) sits at P2 only; Substrait + DataFusion Substrait-serde already exist. Strategic question — repo-local Bonsai vs Substrait interop target. **Stronger `Assess`** per 8th ferry (not `Trial` yet; 7-day research-pass scope like the declarative-env-parity row). Cross-reference Amara's cutting-edge-gaps catalogue. |
 
 ### Upstreams / prior art