Research-Stack/4-Infrastructure/shim/compression_signal_shaping_synthesis_receipt.json

{
  "approach_taxonomy": [
    {
      "approach": "PAQ_style_context_mixing",
      "promotion_gate": "only byte measurement and exact decode count",
      "risk": "context/model bytes can silently exceed gain",
      "shapes": "probability context",
      "use": "long-range sparse contexts, context mixing, arithmetic-coding style evidence"
    },
    {
      "approach": "decision_diagram_route_search",
      "promotion_gate": "route cost < incumbent with decoder/residual/witness counted",
      "risk": "route explosion without admissible lower bounds",
      "shapes": "candidate route space",
      "use": "enumerate transform routes with lower bounds and prune dominated branches"
    },
    {
      "approach": "T16_candidate_pipeline",
      "promotion_gate": "event must become an executable route with exact rehydration",
      "risk": "signal analogy mistaken for compression evidence",
      "shapes": "weak event detection",
      "use": "detect residual-collapse events in noisy candidate forests"
    },
    {
      "approach": "phi_response_family_selection",
      "promotion_gate": "held-out fit beats simple baselines",
      "risk": "Phi gain treated as universal law",
      "shapes": "response curve",
      "use": "choose log/saturating/Hill/low-exponent response by measured error"
    },
    {
      "approach": "nonlinear_compressed_sensing",
      "promotion_gate": "structure and regularity conditions explicit",
      "risk": "nonlinear manifold route without bounds",
      "shapes": "regular nonlinear measurement map",
      "use": "guide structured recovery when RIP-like or separability conditions exist"
    },
    {
      "approach": "generative_compressed_sensing",
      "promotion_gate": "latent + residual + uncertainty bytes beat baseline",
      "risk": "generator becomes hidden payload or biased source substitute",
      "shapes": "latent proposal manifold",
      "use": "replace plain sparsity with learned low-dimensional priors"
    },
    {
      "approach": "invertible_generative_inverse",
      "promotion_gate": "invertibility guard, support check, residual closure",
      "risk": "approximate invertibility treated as lossless",
      "shapes": "invertible/flow chart",
      "use": "reduce representation error and expose uncertainty in inverse route charts"
    },
    {
      "approach": "holographic_fractional_recursive_fold",
      "promotion_gate": "decoded hash closes and memory/kernel bytes counted",
      "risk": "boundary/bulk split hides payload",
      "shapes": "boundary descriptor and bounded memory",
      "use": "short descriptor plus exact residual, graph harmonics, bounded history"
    },
    {
      "approach": "signal_invariant_roots",
      "promotion_gate": "features only choose routes; bytes decide",
      "risk": "feature score promoted without codec trial",
      "shapes": "signal morphology feature space",
      "use": "route chunks by spectral, transient, autocorrelation, DCT, phase, and similarity roots"
    },
    {
      "approach": "semantic_topology_regimes",
      "promotion_gate": "round-trip loss and contradiction/torsion receipts",
      "risk": "smooth story over torn semantics",
      "shapes": "fold/prune/tear decision",
      "use": "avoid false merges; classify beautiful/ugly/horrible compression regimes"
    },
    {
      "approach": "llm_control_plane_compression",
      "promotion_gate": "source bytes, retained bytes, quality delta, provenance",
      "risk": "lossy summary sold as exact compression",
      "shapes": "prompt/logogram/control representation",
      "use": "prune prompts, use symbolic cells, use compressed proxy views"
    },
    {
      "approach": "docmd_static_payload_strategy",
      "promotion_gate": "built-site payload measurement with exact plugin config",
      "risk": "architecture reduction confused with content compression",
      "shapes": "runtime payload",
      "use": "pre-render static HTML, omit heavy framework runtime, gate plugins, externalize search index"
    }
  ],
  "claim_boundary": "This synthesis proposes testable route-shaping experiments. It is not a Hutter Prize result, not proof of a new compressor, and not a guarantee that signal features will improve wiki8.",
  "failure_rules": [
    "feature score treated as byte gain -> invalid",
    "sidecar, witness, residual, decoder, or container bytes omitted -> invalid receipt",
    "latent/generative prior used as hidden source payload -> invalid",
    "semantic merge without round-trip or contradiction check -> hold",
    "history kernel unbounded or uncounted -> fail closed",
    "docmd-style staticization reported as Hutter compression -> overclaim",
    "negative controls omitted from new route claim -> weak claim"
  ],
  "immediate_experiment_ladder": [
    {
      "action": "extract per-chunk signal features and compare feature clusters to bz2/zstd outcomes",
      "name": "wiki8_signal_feature_baseline",
      "step": "E1",
      "success": "feature clusters predict which chunks benefit from which existing codec route"
    },
    {
      "action": "choose among existing routes only: raw, bz2, zstd, xml_token+bz2, tokenbook+bz2 if available",
      "name": "route_classifier_without_new_codec",
      "step": "E2",
      "success": "classifier beats always-bz2 after classifier sidecar bytes"
    },
    {
      "action": "apply semantic/topology guards before tokenbook merge",
      "name": "topology_guard_tokenbook",
      "step": "E3",
      "success": "bad merges fall while byte gain remains non-negative"
    },
    {
      "action": "export a small tiddler/article slice to static pages plus external index",
      "name": "docmd_static_wiki_slice",
      "step": "E4",
      "success": "lower initial payload than live surface with same navigability"
    },
    {
      "action": "route stream chunks with finite fractional residual memory",
      "name": "bounded_history_scheduler",
      "step": "E5",
      "success": "history-aware route choice beats memoryless route after history bytes"
    }
  ],
  "new_candidate_patterns": [
    {
      "candidate_equation": "route = argmin_r LB(r | phi_signal(chunk), topology_regime, history_state)",
      "first_test": "wiki8 chunk sweep with features: entropy, XML tag density, DCT energy, transient edges, autocorrelation, cosine reuse",
      "id": "N1_signal_shaped_route_compiler",
      "novelty": "combine signal invariant roots with DD route search",
      "promotion_gate": "chosen route beats bz2/zstd baseline after feature/witness bytes",
      "shape": "chunk -> feature vector -> route family -> codec trial -> exact residual",
      "testability": "high",
      "why_it_popped": "signal roots supply cheap morphology; DD supplies admissible route discipline"
    },
    {
      "candidate_equation": "payload_total = html_static + js_core + css_core + selected_plugin_assets + index_external",
      "first_test": "build a small TiddlyWiki/article slice both live and static; compare initial gzip payload and search index cost",
      "id": "N2_runtime_staticization_as_compression_prepass",
      "novelty": "treat docmd-style no-runtime output as a compression prepass for wiki/tiddler publishing",
      "promotion_gate": "same navigation/search affordance with lower initial payload",
      "shape": "tiddlers/articles -> static route pages + external search index + manifest",
      "testability": "high",
      "why_it_popped": "payload shrinks by not shipping dynamic state; maps to gated leaves in DD"
    },
    {
      "candidate_equation": "C = bytes(z) + bytes(model_id) + bytes(residual) + bytes(witness) + bytes(decoder)",
      "first_test": "small structured corpus slice with tokenbook latent IDs and exact residual lane",
      "id": "N3_witness_budgeted_latent_route",
      "novelty": "use generative/flow priors only as proposals with explicit latent/residual byte accounting",
      "promotion_gate": "C < incumbent and decoded hash equals source hash",
      "shape": "latent z proposes transform; exact residual repairs; uncertainty decides hold",
      "testability": "medium",
      "why_it_popped": "generative and invertible priors are useful only when they stop hiding model state"
    },
    {
      "candidate_equation": "h_t = sum_{tau<t, window W} K_alpha(t-tau) * residual_tau; route_t = R(chunk_t, h_t)",
      "first_test": "stream wiki8 chunks; compare memoryless route choice vs bounded-history route choice",
      "id": "N4_fractional_history_route_scheduler",
      "novelty": "bounded-memory scheduler for nonstationary corpus regions",
      "promotion_gate": "history bytes are counted and improve total compressed size",
      "shape": "route choice depends on recent route residuals through a finite fractional kernel",
      "testability": "medium",
      "why_it_popped": "fractional dynamics and cognitive overload both say history changes threshold response"
    },
    {
      "candidate_equation": "regime = classify(invariant_overlap, torsion, round_trip_loss, contradiction)",
      "first_test": "compare tokenbook merges with and without contradiction/torsion holds on docs/tiddlers",
      "id": "N5_topology_regime_guard",
      "novelty": "use beautiful/ugly/horrible semantics as a pre-code safety gate",
      "promotion_gate": "fewer bad merges without losing byte wins",
      "shape": "fold when invariants align, prune when quality loss bounded, isolate when torsion high",
      "testability": "medium",
      "why_it_popped": "semantic compression needs a tear detector before it creates false tokenbooks"
    },
    {
      "candidate_equation": "promote iff positive route beats baseline and matched negative control fails or underperforms",
      "first_test": "for each new transform, include a deliberately bad route with same sidecar budget",
      "id": "N6_physical_signal_probe_feedback",
      "novelty": "borrow CAD/DNA force-probe discipline for compression experiments",
      "promotion_gate": "positive gain survives against negative control",
      "shape": "route hypothesis -> measurable perturbation -> negative control -> receipt",
      "testability": "high",
      "why_it_popped": "the CAD frame made measurement and negative controls explicit; compression routes need the same habit"
    }
  ],
  "primary_read": "Across the local compression, compressed-sensing, signal-root, semantic-topology, and docmd payload notes, the new pattern is not another universal compressor. It is a signal-shaped route compiler: shape the route space before coding, then pay exact residual, witness, decoder, and container bytes after coding.",
  "receipt_hash": "fc06057b20dc2281161e7380a63557171ab4d87ee7a82277fe2e8d74b1446f68",
  "schema": "compression_signal_shaping_synthesis_v1",
  "source_artifacts": [
    {
      "bytes": 3665,
      "path": "6-Documentation/tiddlywiki-local/wiki/tiddlers/PAQ Style Compression Review.tid",
      "sha256": "e36d4ffb329d09a4a42b9a824b92a2ea0008fdbdad0dc636174736a6e18df45f"
    },
    {
      "bytes": 6543,
      "path": "6-Documentation/tiddlywiki-local/wiki/tiddlers/Hutter Equation Metastate Transfold.tid",
      "sha256": "97839ff6f6f60ca827b85f80b5f44fd6a78f0119b5b4976dca11bb8c25ac2d29"
    },
    {
      "bytes": 6474,
      "path": "6-Documentation/tiddlywiki-local/wiki/tiddlers/T16 Candidate Pipeline Equation Prior.tid",
      "sha256": "5e8e2519df1e5b3814f4d234de65b6b49713e2a77eceb13523d9f2d24cb76b94"
    },
    {
      "bytes": 5164,
      "path": "6-Documentation/tiddlywiki-local/wiki/tiddlers/Phi Scaling Response Model Selection.tid",
      "sha256": "993c616c03cf255a2cb9d756511a165f11c37808f2a2067621802d376ef1e447"
    },
    {
      "bytes": 22500,
      "path": "6-Documentation/tiddlywiki-local/wiki/tiddlers/Classical Signal Roots Quantum Translation Program.tid",
      "sha256": "6290c0898b532730b86a0d36c1ac7dece5ecf250588daaf927c4c27f0ad4e59c"
    },
    {
      "bytes": 1545,
      "path": "6-Documentation/tiddlywiki-local/wiki/tiddlers/Semantic Topology Compression Regimes.tid",
      "sha256": "5415310b6d9f9907d26536932724dd2da83beb8196b0f9329488497db213f8c0"
    },
    {
      "bytes": 1916,
      "path": "6-Documentation/tiddlywiki-local/wiki/tiddlers/LLM Compression Architecture Priors.tid",
      "sha256": "42944d5ee4133b6e0aa28833db1fd66e2252a262f448bd35394108b10ba373a9"
    },
    {
      "bytes": 2556,
      "path": "6-Documentation/tiddlywiki-local/wiki/tiddlers/docmd Size Strategy Prior.tid",
      "sha256": "2aa5c194c0e4d55de83012d4566f846f784ff2ba08351888e982579d66573ee1"
    },
    {
      "bytes": 5824,
      "path": "4-Infrastructure/shim/nonlinear_compressed_sensing_structural_prior_receipt.json",
      "sha256": "f1e93d5ede20785bffed4e8b7520f3cdc9050c74c493dba41930d144f69874bc"
    },
    {
      "bytes": 5140,
      "path": "4-Infrastructure/shim/generative_compressed_sensing_prior_receipt.json",
      "sha256": "2eef678ab384e524e3909b310d8e36efa5c4776f699bc4d99ff720a1ee07c34d"
    },
    {
      "bytes": 5462,
      "path": "4-Infrastructure/shim/invertible_generative_inverse_prior_receipt.json",
      "sha256": "da99269c79d7b429524d43a6e69b6e860212c4b34b79421f91faafb88660f27e"
    },
    {
      "bytes": 8441,
      "path": "4-Infrastructure/shim/holographic_fractional_recursive_equation_fold_receipt.json",
      "sha256": "7284397eae4a93679e69a1550c76e322bfb28bc47c10af382fb05d8fd16fd74c"
    },
    {
      "bytes": 17340,
      "path": "4-Infrastructure/shim/signal_equation_invariant_roots_receipt.json",
      "sha256": "d62fe13ad78a481c4c2954bfa06a1e2255a5cd8923fb5f4a6b90d833a7cd1972"
    },
    {
      "bytes": 1948,
      "path": "4-Infrastructure/shim/semantic_topology_compression_regimes_receipt.json",
      "sha256": "e3b09e3090f9c19e5df2819424efd258e0968ba78136ff31c9f25c9d0eacdc3f"
    },
    {
      "bytes": 7224,
      "path": "4-Infrastructure/shim/llm_compression_architecture_prior_receipt.json",
      "sha256": "4e65458e29e17f032fa2b0df0a06d4713a2102ff2a58ef13d96c37dda5691264"
    },
    {
      "bytes": 8845,
      "path": "4-Infrastructure/shim/connectome_protective_cognitive_load_reweighting_receipt.json",
      "sha256": "fc222e265e70f69ee5e6039dd0cee2665a02e549b6718a7d022f9241849b3cf1"
    }
  ],
  "unifying_equations": {
    "exact_cost": "C_total = bytes(payload) + bytes(sidecar) + bytes(residual) + bytes(decoder) + bytes(witness) + bytes(container)",
    "negative_control": "valid_gain iff C(candidate) < C(baseline) and C(candidate) < C(matched_bad_route)",
    "promotion": "promote iff H(decode(r*)) == H(source) and C_total < incumbent and failure_rules == none",
    "route_selection": "r* = argmin_r LB(r | phi_signal(c), semantic_regime(c), history_state)",
    "signal_feature_vector": "phi_signal(c) = [H(c), tag_density(c), DCT_energy(c), transient(c), autocorr(c), cosine_reuse(c)]"
  },
  "what_is_actually_new": [
    "The strongest new move is route-space signal shaping, not a new compressor.",
    "docmd reframes compression as runtime-state omission: do not ship branches you can rebuild.",
    "Signal invariant roots give a concrete feature surface for choosing routes before spending codec time.",
    "Semantic topology supplies a guard against destructive tokenbook merges.",
    "Generative/invertible models should be restricted to proposal charts with explicit residual byte accounting.",
    "Every interesting analogy becomes useful only after it is paired with a negative control and exact decode receipt."
  ]
}