{"id":"W2090432523","doi":"10.1007/s10664-008-9076-6","title":"“Cloning considered harmful” considered harmful: patterns of cloning in software","year":2008,"lang":"en","type":"article","venue":"Empirical Software Engineering","topic":"Software Engineering Research","field":"Computer Science","cited_by":363,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Cloning (programming); Source code; Computer science; Software engineering; Code (set theory); Software system; Software; clone (Java method); Maintainability; Web application; Codebase; Programming language; World Wide Web; Biology; Genetics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.00063403,0.0005862429,0.0008906313,0.0008212666,0.0001637641,0.0001073126,0.001304384,0.0003110822,0.00009247373],"category_scores_gemma":[0.01122249,0.0006485433,0.0002502568,0.00146781,0.0001048461,0.0007002025,0.0008272546,0.00112495,0.00005015781],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003241661,"about_ca_system_score_gemma":0.0003391605,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001731418,"about_ca_topic_score_gemma":0.00001634555,"domain_scores_codex":[0.995464,0.0001149783,0.001037949,0.001016343,0.001036025,0.001330708],"domain_scores_gemma":[0.9931839,0.004786635,0.0001704528,0.001111533,0.00028943,0.0004580265],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0000134123,0.0001138282,0.9472035,0.000254443,0.00006843892,0.0009116988,0.001610781,0.04721961,0.000600313,0.00008519979,0.0008399313,0.001078797],"study_design_scores_gemma":[0.002237117,0.0002040457,0.9482715,0.0007903987,0.00001633014,0.0006366497,0.00004592921,0.02974329,0.01474748,0.0001725155,0.001720953,0.001413768],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4914618,0.0004283489,0.5060543,0.0001819808,0.0004281427,0.0002550912,0.00001286842,0.00117063,0.000006899595],"genre_scores_gemma":[0.8596228,0.00003718934,0.1398272,0.0001534572,0.0001124139,0.00007262189,0.00001269367,0.00009975988,0.00006179701],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.3681611,"threshold_uncertainty_score":0.9995966,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.044662327963924,"score_gpt":0.2815639292607685,"score_spread":0.2369016012968445,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}