{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp12_substrate_denial72_20260805","cell_id":"additive_measure_space_design__computer_science","arm":"CONSTRAINED_HIGH","candidate_id":"additive_measure_space_design__computer_science__CONSTRAINED_HIGH","decision":"PROPOSAL","abstention_reason":null,"proposal":{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp12_substrate_denial72_20260805","cell_id":"additive_measure_space_design__computer_science","arm":"CONSTRAINED_HIGH","candidate_id":"additive_measure_space_design__computer_science__CONSTRAINED_HIGH","proposal_index":1,"version":0,"title":"Passive Phase-Change Calorimeter Cartridges for Additive Thermal-Energy Attribution in Modular Compute Nodes","problem":"During a bounded concurrent hardware benchmark, computer-systems engineers may need to attribute dissipated thermal energy among a CPU package, accelerator, memory bank, power-conversion stage, and chassis residual. Whole-node electrical measurements cannot provide that partition, while estimates formed by subtracting or combining overlapping component readings can change with the chosen boundaries, exceed the node total, or produce negative residuals.","actors":["Computer-systems test engineer","Laboratory thermal-safety owner","Modular compute-node hardware","CPU, accelerator, memory, and power-conversion modules","Passive phase-change calorimeter cartridges","Residual chassis calorimeter"],"observable_state":"For the same run-plus-cooldown interval, component-attribution methods using different boundaries yield incompatible subtotals, a component sum larger than the independently bounded whole, or an unexplained residual. The physical target state is the nonnegative melted fraction in each cartridge, with every heat-producing region assigned to exactly one cartridge or to the residual chassis calorimeter.","consequence":"Without a partition-consistent thermal-energy measure, an engineer cannot tell which physical module or unassigned chassis path breached a thermal budget, so heat-sink, packaging, or power-delivery changes may be directed at the wrong subsystem.","affected_objective":"Obtain auditable, module-resolved thermal-energy measurements for a short concurrent compute-node test while preserving a coherent total across disjoint physical heat domains.","intervention":"Build an instrumented replica node in which each declared, nonoverlapping heat domain is thermally coupled to a standardized sealed phase-change-material cartridge and insulated from sibling domains. A separate cartridge captures heat from the declared residual chassis domain. Each cartridge has a transparent, graduated melt-volume column calibrated in joules using a traceable resistive heater. The node is observed through a bounded run and cooldown to a common terminal temperature. Melted PCM volume provides a passive, nonnegative integral of heat delivered to each cartridge; readings from disjoint cartridges use the same joule scale and can be added directly. Empty positions, below-resolution changes, saturated cartridges, and boundary leakage are marked distinctly rather than converted to ordinary zero. No computation is required to create, retain, or read the physical measurement.","structural_mapping":[{"archetype_element":"Measurable Universe Scope","domain_realization":"All heat transferred from the instrumented node's declared physical domains into calorimeter cartridges during one specified run-plus-cooldown interval."},{"archetype_element":"Measurable Subset Family","domain_realization":"The empty domain, each nonoverlapping module domain, the residual chassis domain, and unions of those domains; overlapping logical labels such as 'memory-intensive work' are not measurable physical subsets in this instrument."},{"archetype_element":"Empty-Set Zero Rule","domain_realization":"An unoccupied, thermally isolated reference position should produce no melt beyond the calibrated background allowance."},{"archetype_element":"Nonnegative Size Assignment Rule","domain_realization":"A cartridge reports zero or positive joules from its melted PCM volume; it cannot report negative heat."},{"archetype_element":"Disjoint Additivity Rule","domain_realization":"Because every heat-producing region belongs to one thermally isolated cartridge domain, the thermal-energy measure of a union is the sum of the cartridges belonging to that union."},{"archetype_element":"Normalization or Scale Anchor","domain_realization":"Identical joule graduations are established by applying known electrical energy to each cartridge with a calibration heater under the same starting and terminal conditions."},{"archetype_element":"Null-Set and Negligibility Policy","domain_realization":"Below-resolution melt, an empty cartridge position, saturation, and suspected leakage receive separate physical flags and are not silently treated as measured zero."},{"archetype_element":"Partition Consistency Register","domain_realization":"A calibration plate can heat several isolated dummy modules separately or as a declared union; the summed cartridge readings are compared with the union's known heater input."},{"archetype_element":"Integration and Downstream Use Contract","domain_realization":"The readings represent captured dissipated heat for the declared physical partition and interval, not logical-operation counts, electrical input energy, or unrestricted production behavior."}],"mechanism_mapping":[{"mechanism_slug":"measure_space_specification","role":"The cartridge layout physically instantiates the universe, mutually exclusive module domains, residual domain, and allowed unions before a test begins.","counterfactual_removal":"Without exclusive domains and a residual domain, heat crossing an undeclared boundary could be omitted or attributed twice, so the component readings would not constitute one additive measure."},{"mechanism_slug":"normalization_constant_calibration","role":"Known resistive-heater inputs establish a common conversion from melt volume to joules for every cartridge.","counterfactual_removal":"Without a common physical calibration, melt fractions from cartridges with different PCM masses or geometries would be incomparable and could not be validly summed."},{"mechanism_slug":"finite_or_countable_additivity_test","role":"Dummy heater trials compare energy delivered separately to disjoint cartridges with energy delivered across their union.","counterfactual_removal":"The device could still produce readings, but leakage, cross-heating, or nonlinear saturation could remain undetected and defeat the claimed additive mapping."},{"mechanism_slug":"partition_sum_table","role":"Graduated cartridges expose each disjoint term and permit direct arithmetic recomposition of any declared physical union.","counterfactual_removal":"Module readings would remain individually observable, but there would be no disciplined check that an alternative valid grouping preserves the same total."},{"mechanism_slug":"null_set_policy_register","role":"Reference cartridges and visible overrange or below-resolution markers distinguish absence, negligible response, and invalid measurement.","counterfactual_removal":"A failed, saturated, or insensitive cartridge could be mistaken for a genuine zero-heat domain."},{"mechanism_slug":"measure_invariance_review","role":"Equivalent dummy-heat inputs are tested under alternative allowed module groupings and cartridge positions to determine whether the measured total is stable.","counterfactual_removal":"Position-dependent losses or boundary choices could masquerade as differences in module heat."}],"causal_chain":["Electrical activity in each physical compute module becomes heat during the test and subsequent cooldown.","Thermal insulation and dedicated conductive paths direct each declared domain's heat primarily into its assigned phase-change cartridge rather than into sibling domains.","Absorbed heat melts a corresponding amount of calibrated phase-change material, physically integrating heat over time without sampling or software inference.","The graduated melt column retains a visible, nonnegative reading in common joule units.","Because module domains are disjoint and use the same calibrated unit, their readings add to the reading for any declared union, subject to measured leakage and resolution bounds.","The explicit residual chassis cartridge prevents unassigned heat from being forced into a named component estimate.","Engineers can compare the resulting physical module measures with thermal budgets and localize which physical heat domain warrants further hardware investigation."],"baseline":"Use a whole-node plug power analyzer or motherboard telemetry, then estimate module contributions from vendor counters, isolated reruns, or subtraction. This baseline may be useful for electrical input but does not directly create a disjoint physical measure of heat from concurrently active modules.","nearest_rivals":["Per-rail shunt or power-analyzer instrumentation, which directly measures electrical input on separable rails but cannot cleanly assign shared conversion losses or energy crossing rail boundaries.","Separate liquid-cooling loops with flow and inlet/outlet temperature calorimetry, which can continuously attribute removed heat but require pumps, flow calibration, and multiple synchronized measurements.","Whole-node calorimetry combined with one-module-at-a-time runs, which supplies a total but changes concurrency and interaction conditions during attribution.","Embedded on-chip or board energy counters, which preserve normal packaging but depend on electronic sensing, coverage assumptions, and potentially proprietary computation."],"remaining_contrastive_claim":"For short, controlled run-plus-cooldown tests, passive phase-change cartridges can provide a directly visible thermal-energy measure over predeclared disjoint hardware domains, including an explicit residual, without synchronized electronic sensing or algorithmic attribution. The claim is limited to captured heat within demonstrated leakage, saturation, and resolution bounds.","authority_safety":{"decision_authority":"The laboratory thermal-safety owner decides whether calibration or powered-node testing may proceed; the systems engineer may only configure measurements within the approved thermal and electrical envelope.","authorized_first_step":"Construct and calibrate two small cartridges on non-computing resistive-heater mock modules at low voltage, then test separate heating, simultaneous heating, and a zero-input reference.","excluded_actions":["Installing unvalidated cartridges on production or safety-critical computers","Operating a processor or accelerator without its required thermal protection","Using flammable or chemically incompatible phase-change material without containment review","Exceeding cartridge energy capacity or pressure rating","Treating a saturated, leaking, or below-resolution cartridge as quantitative evidence","Inferring software-level energy attribution from physical module readings"],"halt_rollback":"Stop heating if containment deforms, pressure or temperature approaches its rated limit, the reference cartridge changes unexpectedly, or cross-domain leakage exceeds the preset bench criterion. De-energize the mock load, allow passive cooling behind a barrier, quarantine any leaking cartridge, and revert to the original approved heat sink before operating compute hardware."},"negative_tests":{"strongest_counterevidence":"A calibrated dummy module delivers the same known heat under two allowed partitions, but the summed cartridge readings differ beyond the combined calibration uncertainty, or substantial heat appears in an unheated sibling cartridge.","problem_falsifier":"Direct per-module electrical or thermal measurements on the target node already form mutually exclusive domains, include conversion and residual losses, and recompose consistently across alternative valid partitions; then the stated attribution problem is absent for that node.","intervention_falsifier":"Across repeated heater trials within the intended range, melt-volume readings are not monotone with delivered heat, cannot be placed on a common joule scale, saturate before the test completes, or lose enough heat across cartridge boundaries that disjoint sums fail the declared acceptance bound.","risks":["Cartridges may alter module temperatures and therefore change the heat-generating behavior being measured.","Heat may escape through sockets, cables, radiation, or airflow and violate the declared physical partition.","Cross-heating between cartridges may double-attribute or misattribute energy.","PCM hysteresis, supercooling, imperfect mixing, or sensible-heat contributions may make melt volume nonlinear.","Cartridge saturation can truncate high-energy readings.","Containment failure can expose hot or chemically incompatible material.","A physical-module measure may be incorrectly presented as attribution to software functions or users."]},"next_evidence_step":"On a benchtop fixture, use two thermally isolated resistive heaters and one residual plate with three cartridges. Run a preregistered matrix of zero-input, each-heater-only, simultaneous-heater, swapped-position, and near-capacity trials at three known energy inputs. Record melt-column readings manually, estimate repeatability and cross-domain leakage, and reject the concept before any computer installation unless readings are monotone, unsaturated, and the disjoint sum agrees with known combined heater energy within a preset engineering tolerance.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"No other experiment proposals were inspected under runtime isolation, so cross-proposal diversity is not asserted. Internally, this candidate is anchored in passive latent-heat instrumentation and physical thermal partitioning rather than a software, policy, reporting, or incentive mechanism.","revision_record":{"parent_version":null,"progress_targets_addressed":["Initial construction of a concrete computer-systems problem with a preserved additive-measure mapping","Compliance with the constrained physical and measurement substrate","Explicit residual domain, null handling, rivals, falsifiers, and bounded bench evidence step"],"conceptual_changes":["Initial version; defines measured size as captured thermal energy over disjoint physical hardware domains."],"operational_changes":["Initial version; limits first testing to low-voltage resistive-heater mocks and excludes deployment on computing hardware."],"evidence_changes":["Initial version; prior art remains unsearched and all proposed validation evidence is prospective."],"claim_changes":["Initial version; claim is limited to short controlled tests and demonstrated leakage, resolution, and saturation bounds."]},"substrate_contract":{"primary_allowed_process":"MEASUREMENT_INSTRUMENTATION","counterfactual_independence":"If all software, algorithmic inference, databases, dashboards, reporting systems, incentives, authorization rules, and procedural enforcement are removed, heat entering a cartridge still melts a calibrated amount of phase-change material and leaves a visible physical reading. Disjoint cartridges still retain independently additive quantities. A person can read the graduated columns directly; computation is not the source of attribution.","forbidden_channel_audit":"The intervention contains no model, software control loop, database, recommender, information-routing mechanism, or required downstream analytics. Safety authority and test protocol constrain hazardous use but do not produce the measurement. Manual labeling and arithmetic communicate the result but are not its causal source. The essential chain is physical heat transfer, latent-heat absorption, retained melt volume, and direct instrumental observation."}}}