Public Methods, Replication, And Review Surface

This is where CCT's search grammar becomes executable. Formal constraints, prospective simulation questions, frozen comparisons, resource fronts, and physical exposure paths become objects another reader can inspect or rerun.

The selected release gives CCT a compact working memory. It records what was asked, which assumptions were active, what the comparator knew, how resources were counted, what outcome changed the route, and which question follows next. This lets formal work constrain experiments, simulation select higher-value exposures, and physical results return to a common language.

This release is part of CCT's Foundation: reusable formal, simulation, decision, resource, and exposure machinery. It feeds the Frontier by making the next theorem, campaign, bench exposure, access record, or composition question easier to specify, while requirements from the Guiding Horizon can be translated backward into the same executable form.

Download CCT Selected Public Replication Release V1.1

The archive is approximately 3.4 MB. Its SHA-256 is:

6edb487b9cd89d08bb38be0e3492a390946077bfef36d8c67043b17b8544473a

Inside the archive, README.md is the operator quickstart, CLAIMS_EVIDENCE_MAP.md connects each result to its artifact and next exposure, SOURCE_SET.json freezes the allowlist, and verify_release.py checks the package end to end.

What The Surface Makes Possible

Turn theory into active search constraints

CCT's theorem stack does more than sit beside the experimental program. It reveals which patterns are easy to manufacture, which conclusions are underdetermined, which resources must be counted, which causal paths must be conditioned, and which stronger claim requires a new proof object.

The package pairs bounded formal results with executable companions. The analytical proof establishes the result. The companion exercises its declared scope, assumptions, counterexamples, route priorities, and machine-readable certificate.

The compact proof spine contains five reader-facing theorem families:

  • Multiwell Anti-Uniqueness (OP0a);
  • Finite-Window Measurement Envelopes (OP1);
  • Causal Command Attribution (OP3);
  • Passive-Wave Boundary Bounds (BT7b);
  • Multi-Resource Programmability Fronts (Vector OP4).

Beyond the compact V1.1 download, the working replication tree now carries a multi-stage observer-conditioned formal sequence. Resource-indexed quotients, compositional record flows, typed physical-scale stability, and affine parameter flow make channel-relative relevance, order dependence, physical scale typing, fixed-point attraction, and coordinate-conjugacy candidates executable. Nonlinear, bifurcation, and stochastic successors add certified inner basins, branch change, multiple attractors, retention and exit semantics, clock-sensitive continuous-time evolution, and finite-sample kernel and generator uncertainty through independent complete-jump records and independent bounded episodes. The analytical notes state the bounded results; dependency-free companions exercise their assumptions, counterexamples, route priorities, and exact finite certificates.

The first source-linked measured-record application now follows that sequence. GW150914 Observer-Scale Flow Preparation V0 translates retained real-data-derived amplitude records across six doubled integration windows. All three calibration bands recover the ordinary fixed-frequency amplitude translation; a fit-withheld final edge on the same finite records retains it in two bands and breaks it in one. The object makes a useful boundary executable: a measured observer/estimator scale flow can expose where a description holds without becoming a physical-system flow or fixed-point claim.

Explore Selected Formal Results and Open Questions

Test prospective regime discovery

Four simulation chains test distinct links in prospective regime discovery: execution quality, coordination order, opportunity mapping, and limited-probe experiment selection.

  • Structured space-time excitation improved a fixed-wave process frontier at equal incident energy, with destructive controls for phase, timing, and model alignment.
  • A frozen coordination-order grammar predicted useful handles and transferred to an external excitable-system model.
  • A common task-aligned intervention contrast prospectively separated positive, boundary, and adverse temporal-coordination regions across two standard-physics model families.
  • A structural representation changed prospective experiment selection under matched information, policy access, acquisition logic, and probe budgets, reducing regret-curve area against a tuned raw-schedule search.

The package carries the frozen configurations, source runners, result tables, reports, and artifact verifiers for all four chains.

A completed follow-on benchmark now extends the limited-probe chain across 36 new cases in three model systems. It treats the search representation and candidate menu as empirical parts of the method, mapping distinct validity regimes across systems and showing that menu construction can change the search ranking. That paper-stage benchmark is summarized in Mapping the Validity of Structured Search; its manuscript and standalone replay package are candidates for a later selected replication release rather than contents of the current release.

Make architecture and resources first-class

Two narrower exhibits show how architecture can create useful handles:

  • a distributed schedule that writes retained geometry and fails under route-destroying controls;
  • frequency-position-mode co-design that exposes a controllable deformation mode absent from a lower-handle architecture.

Multi-Resource Programmability Fronts (Vector OP4) then make the comparison front-valued. Energy remains central, while latency, calibration, synchronization, memory, reliability, setup, recovery, and support can change which strategy is genuinely preferable.

Beyond the compact V1.1 download, the working public-replication tree now carries additional access, formulation, attribution, and transfer contracts being prepared for a later selected release.

Complete-Arrangement Bottleneck Attribution V0 now makes the reason a limit moved inspectable. Its route surface separates physical possibility, observer/estimator legibility, coupling/control accessibility, representation/menu searchability, and complete-task usefulness. The first source-linked application attributes the aggregate and LLG limited-probe changes to the searchability class while preserving unresolved FHN and component-isolation routes. This tells the next campaign whether to change theory, observation, access architecture, search formulation, or the task/resource boundary.

Candidate-Menu Generator Contract V0 makes the intervention menu itself inspectable before scarce probes are spent. It records whether proposed actions are physically distinct, source-traceable, observable, actuable, resource-bounded, and accompanied by incumbents and controls. Its source translations reconstruct a 32-policy equal-drive menu and show that two disjoint 64-policy menus can reverse the aggregate method comparison. A frozen steel-surrogate campaign now carries this question prospectively across two menu generators and four search representations; implementation and scoring are complete, with one independently custodied fresh execution next. Its hash-bound atlas and bottleneck companion is ready to classify that retained result as a searchability gain, narrowed validity map, equivalence, or changed search route.

Cross-Campaign Inheritance V0 extends the comparison across time. It gives two workflows the same frozen Campaign A evidence in different predeclared representations, then compares their consequential Campaign B decisions under matched observations, actions, outcomes, resources, and custody. This makes cumulative scientific memory testable: whether a CCT access-and-validity record improves later experiment selection, control, calibrated deferral, or earlier closure relative to the strongest ordinary summary.

Primitive Transfer and Scale Contract V0 makes the next cumulative question executable: whether the same bounded operation survives a declared change in object, environment, location, nuisance state, duration, scale, implementation, or model family. Its source applications record one unresolved model-family boundary and one fixed-wave target contract frozen with zero pre-outcome reserved evaluations. After the separate engine opened that target block once, the transfer evaluator reverified the retained chain without plant re-execution. All 8/8 source-qualified candidate operations and 5/5 source-qualified incumbent operations survived with zero target adaptation. This supports model transfer under the frozen nuisance change; physical transfer and true scale change remain separate next exposures.

General Resource-Bounded Physical-Access Envelope V0 now gives the stack a common finite set object for observational, control, and intentionally verified programmability envelopes under error, reliability, horizon, safety, and eight resource limits. Resource-Indexed Access Equivalence V0 compares complete arrangements under that grammar. Ordered, network/concurrent, and dynamic-feedback composition successors then test interface fit, multiway demand, common-mode reliability, causal delay, and loop stability; their source-linked photonic application identifies the exact joint records still needed for a physical composition decision.

The fixed-wave atlas and coupling profile have now passed their shared canonical model exposure. Under the frozen 640-observation contract, the structured candidate retained finite-block authority on 9/10 functions versus 7/10 for the calibration-selected incumbents. Two functions were candidate-only, none was incumbent-only, and the atlas produced zero certificate violations and no domain escape. The full raw receipt chain remains externally retained; the working replication tree carries its exact hashes and compact result summary.

Carry results toward physical exposure

Three public-safe exposure cards translate the current search into physical questions:

  • measurement regime and observer mode;
  • field geometry and stable control basins;
  • route-state topology, retention, and reset.

They freeze the comparison class, measurements, branch outcomes, and next public artifact while leaving build-specific implementation in the controlled execution layer.

Translate physical primitives into Tau-X questions

The release includes one mission-ledger template and one synthetic translation example. These ask what an earned primitive would change in timing, sensing, coordination, correction, recovery, reliability, infrastructure, and the complete resource front.

This makes the mission horizon testable from both directions. Tau-X can send a hard requirement backward into theory, simulation, or CCT Labs; an earned primitive can travel forward into a more capable mission architecture.

Explore Tau-X

Choose A First Route

If you want to inspect... Start with... What it demonstrates
Formal constraints The five notes and companions in open-theorem-working-notes/ and cct-public-replication/ How bounded assumptions, counterexamples, and route consequences become executable.
Observer-conditioned equivalence, scale, basins, and stochastic flow The quotient, compositional-flow, typed-scale, affine/nonlinear flow, bifurcation, stochastic-basin, continuous-time, and finite-sample coverage objects, plus GW150914 observer-scale preparation in the working replication tree How observer contracts change exact model partitions, whether declared maps compose, when robust distinctions and finite invariant candidates survive, how branch and path structure changes, and how measured estimator-scale behavior is kept distinct from physical-system flow.
Structured physical information Fixed-wave pilot and confirmation How phase, timing, and architecture affect a process frontier under equal incident energy.
Coordination order Coordination-order challenge and FHN transfer How an interaction-order grammar can predict and transfer a useful control sequence.
Cross-domain opportunity mapping Phase-0R map and Phase-1 discrimination How one prospective quantity separates positive, boundary, and adverse regions.
Limited-probe discovery Structural search and tuned raw-schedule GP How intervention structure can change which experiment is selected and reduce search burden.
Structural-search validity Prospective three-system benchmark summary How the usefulness of a search representation changes across system families and candidate menus.
Candidate-menu formulation Generator-contract route suite and source reconstructions How intervention menus expose distinct physical handles, controls, access constraints, and construction-sensitive validity boundaries.
Architecture leads Retained geometry and frequency-position-mode co-design How route structure and independent physical handles can change reachability.
Physical exposure exposure-cards/ How the selected evidence becomes a measurement, basin, or retained-state question.
Mission translation tau-x/ How an earned local primitive would enter a mission ledger.

Representative Reruns

After extracting the selected release:

python -m pip install -r requirements.txt
python verify_release.py

The runbook also provides individual commands for every formal capsule, simulation chain, and architecture exhibit. Campaign reruns begin by creating a fresh mutable workspace, preserving the frozen selected release and its hashes.

The current working source tree also has a manifest-driven aggregate verifier. Its latest run selects 83 checks: 80 pass and three confirm intentional dependency behavior. These post-V1.1 objects are candidates for the next selected release.

From Executable Object To Retained Physical Access

The public surface helps a candidate move through a connected sequence:

  1. State the question. Define the observable, estimator, intervention, target, and comparison class.
  2. Apply formal constraints. Identify what can already be proved, ruled out, or narrowed.
  3. Map the candidate regime. Use simulation and synthetic cases to expose operating regions, confounders, and failure boundaries.
  4. Freeze the decision. Record what each workflow knows, what it selects, and what result changes the route.
  5. Expose it physically. Carry the selected question into CCT Labs or an appropriate external facility.
  6. Return the result. Update the theorem target, estimator, branch map, resource front, or mission architecture.
  7. Retain what survives. Record a repeatable operation with its validity domain, failure modes, and complete resource boundary as a candidate physical-access primitive.
  8. Test composition. Ask whether several retained operations remain jointly possible and useful when their interfaces, environments, timing, controls, and resources interact.

This is how public replication supports discovery rather than arriving only after it. The same objects that let another reader inspect a result also help decide which result is worth pursuing. The current selected release supports this sequence through physical-exposure planning; newer observer-equivalence, access-record, coupling-profile, composition, and frozen model-exposure objects extend the working method beyond V1.1 and are candidates for a later selected release.

Public Reference And Controlled Execution

CCT uses two coupled execution layers.

The public reference layer exposes selected definitions, theorem companions, code, schemas, frozen comparisons, result tables, public-safe outputs, and reusable exposure structures. It gives outside readers and collaborators a shared interface for inspecting and extending the program.

The controlled execution layer carries build-specific materials, geometries, operating windows, calibration values, search maps, hardware constants, private notebooks, and partner-specific implementation choices. It preserves the continuity and experimental advantage needed to turn public questions into effective physical runs.

The layers meet through public-safe manifests, preregistrations, result summaries, and promotion records. That connection keeps the public method usable while allowing physical execution to remain cumulative.

Where Exact Status Lives

Different pages answer different questions:

  • What CCT Has Built and Opened presents selected capabilities, results, and next-stage possibilities.
  • Selected Formal Results and Open Questions presents the compact proof spine and active theorem frontier.
  • The selected release README.md provides the operator quickstart.
  • CLAIMS_EVIDENCE_MAP.md ties each public result to its source and next exposure.
  • SOURCE_SET.json and CHECKSUMS.sha256 record exact release contents and integrity.
  • Artifact-level READMEs and result files own the detailed interpretation of individual runs.

The selected release is the runnable front door. The Research Library preserves the wider conceptual and technical context.

See What CCT Has Built and Opened · Explore CCT Labs · Enter the Research Library