{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp09_archetype_breadth150_20260804","research_id":"eoa_inverse_innovation_exp09_light_prior_art_20260804","cell_id":"additive_measure_space_design__computer_science","search_lanes":{"direct_problem_and_intervention":{"queries":["regression test selection overlapping coverage union changed code coverage greedy set cover","change based test selection coverage changed entities union test cases CI","CI test selection changed code additional coverage runtime budget research paper"],"source_ids":["SRC1","SRC2"],"no_result_note":null},"synonyms_and_historical_terms":{"queries":["test suite prioritization additional coverage total coverage algorithm statement coverage overlap","test selection time budget maximize coverage overlap redundant test cases set cover primary research","test suite minimization coverage requirements each test covers set greedy set cover paper"],"source_ids":["SRC1"],"no_result_note":null},"products_practices_and_standards":{"queries":["coverage.py combined data union coverage multiple test runs documentation","JaCoCo execution data merge different class version class id warning","JaCoCo coverage counters missed covered instructions merged execution data OR"],"source_ids":["SRC3","SRC4"],"no_result_note":null},"component_combination":{"queries":["regression test selection maximum coverage problem time constrained tests overlap coverage","CI test selection changed code additional coverage runtime budget research paper","code coverage stable unique identity mismatch uninstrumented units"],"source_ids":["SRC1","SRC2","SRC3","SRC4"],"no_result_note":"No retained source described the complete combination of a build-scoped change-impact universe, coverage-signature partition audit, explicit unresolved-unit queue, and selector-binding contract; this phrase-level miss is not evidence of novelty."}},"sources":[{"source_id":"SRC1","title":"Regression Test Case Prioritization by Code Combinations Coverage","publisher":"arXiv; authors' version of Information and Software Technology research","url":"https://arxiv.org/abs/2007.00370","source_type":"PRIMARY_RESEARCH","claims_supported":["Code-coverage-guided regression-test prioritization is established research.","Total and additional coverage strategies are popular comparators, and coverage is represented by binary membership of code units per test.","Regression suites can be costly, motivating early or selective execution."]},{"source_id":"SRC2","title":"Efficient Incremental Code Coverage Analysis for Regression Test Suites","publisher":"IEEE/ACM International Conference on Automated Software Engineering; arXiv","url":"https://arxiv.org/abs/2410.21798","source_type":"PRIMARY_RESEARCH","claims_supported":["Code coverage is widely used in continuous integration and can impose substantial execution cost.","Executing only regression-test-selected tests can produce incorrect coverage results.","Correctly integrating change-based test selection with coverage updates is nontrivial and can be checked against full-suite coverage."]},{"source_id":"SRC3","title":"Combining data files: coverage combine — Coverage.py documentation","publisher":"Coverage.py","url":"https://coverage.readthedocs.io/en/latest/commands/cmd_combine.html","source_type":"FIRST_PARTY_PRODUCT","claims_supported":["A mature coverage tool combines separate run data into a single file representing their union rather than summing covered-unit totals.","Coverage.py warns when an input data file cannot be read.","Path remapping is required when equivalent source files appear under different locations."]},{"source_id":"SRC4","title":"JaCoCo — Class Ids","publisher":"Mountainminds GmbH & Co. KG and JaCoCo contributors","url":"https://www.jacoco.org/jacoco/trunk/doc/classids.html","source_type":"FIRST_PARTY_PRODUCT","claims_supported":["JaCoCo uses class identifiers to relate execution observations to analyzed classes and distinguish class versions.","A runtime-versus-analysis class mismatch can cause an executed class to be reported as zero-percent covered.","JaCoCo exposes mismatch evidence in its Sessions report, while compiler, configuration, and runtime transformation differences can destabilize identity."]}],"problem_evidence":{"status":"PARTLY_SUPPORTED","finding":"The literature and product documentation make the underlying problem visible: coverage-based prioritization distinguishes total from additional coverage, union aggregation is implemented explicitly, selected-test execution can yield incorrect coverage, and identity mismatches can turn observed execution into reported zero coverage. The retained sources do not establish how frequently CI selectors literally sum per-test covered-unit counts or that ordinary arithmetic sums vary with ordering; the ordering/grouping defect therefore remains only partly evidenced.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"closest_prior_art":[{"name":"Additional code-coverage regression-test prioritization","source_ids":["SRC1"],"overlap":"Uses per-test code-unit membership and prioritizes according to additional coverage, directly discounting units already covered by previously selected tests and addressing redundant coverage.","remaining_difference":"The retained work does not specify the proposal's reproducible change-impact universe, explicit unresolved-unit queue, coverage-signature partition register, or versioned measure contract."},{"name":"Union aggregation in Coverage.py","source_ids":["SRC3"],"overlap":"Combines multiple coverage observations into the union of executed units, providing the proposal's central partition-invariant aggregation behavior.","remaining_difference":"It is a reporting-data operation, not a time-budgeted change-based test selector, and it does not provide the proposed boundary queue or partition-recomposition audit."},{"name":"JaCoCo versioned class identity and mismatch reporting","source_ids":["SRC4"],"overlap":"Provides stable build-artifact identity, separates class versions, and exposes cases where execution data cannot be related to analyzed code.","remaining_difference":"A mismatch is still reported as zero-percent coverage in the coverage result rather than being governed by the proposed explicit nonmeasurable-unit queue and downstream selection contract."},{"name":"iJaCoCo incremental coverage integrated with regression-test selection","source_ids":["SRC2"],"overlap":"Integrates change-aware regression-test selection with coverage maintenance and treats equality with full-suite coverage as a correctness condition.","remaining_difference":"It minimizes tests needed to update coverage data; it does not optimize distinct change-impact union coverage under a recorded time budget or audit coverage-signature partitions and unresolved units."}],"prior_art_disposition":"ADJACENT_PRIOR_ART","contrastive_claim_remaining":"For one frozen build scope, adding an explicit executable-unit universe, a separately reported unresolved-unit queue, coverage-signature partition recomposition, and a versioned contract requiring the selector to consume union cardinality will detect or prevent overlap- and identity-driven inconsistencies that ordinary total-coverage ranking or coverage-file union alone leaves unaudited.","contrastive_claim_falsifier":"The claim is falsified if an existing selector using ordinary additional coverage and current coverage-tool identity checks already reproduces the same universe, unresolved-status accounting, partition invariants, and rankings, or if shadow replay finds no overlap- or identity-driven inconsistency and the added contract changes no affected suite comparison.","gates":{"adequate_source_search":{"status":"PASS","rationale":"Four search lanes covered the proposal directly, older total/additional and set-cover terminology, product practices, and component combinations. Exactly four opened sources from at least three publisher groups were retained, including two primary studies and two first-party sources.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"supported_problem":{"status":"PASS","rationale":"The problem is partly supported: established additional-coverage methods and union-merging practices confirm overlap is material, while CI coverage correctness and class-identity failures confirm scope and missing-evidence hazards. Prevalence of the proposal's exact naive summation baseline was not established.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"distinct_testable_claim":{"status":"PASS","rationale":"Although union coverage and additional prioritization are established, the residual claim about the combined universe, boundary queue, partition audit, and selector contract is operationally distinguishable and has an explicit falsifier.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"bounded_next_test":{"status":"PASS","rationale":"An offline replay of 20 frozen pull-request builds can compare baseline and union-measure rankings, verify recomposition, enumerate unresolved identities, and observe whether overlap corrections alter selections without changing production CI.","source_ids":["SRC1","SRC2","SRC4"]},"no_obvious_safety_or_authority_stop":{"status":"PASS","rationale":"The proposed first step is a reversible shadow analysis under CI-owner authority, preserves existing release decisions, and includes halt conditions for unstable identities, unreproducible universes, failed recomposition, and silent coercion of unknowns to zero.","source_ids":["SRC2","SRC4"]}},"screen_survival":true,"world_novelty_boundary":"This bounded four-source screen establishes only that the proposal is researchable and adjacent to established additional-coverage prioritization, union aggregation, incremental coverage, and artifact-identity practices. It cannot establish world novelty, patentability, market size, expert acceptance, implementation feasibility across ecosystems, or realized operational value."}