[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84160-en":3,"doc-seo-84160-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84160,2336464648746,"Skyler","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Scaling Author Identity Disambiguation to the World of Code: A Methodology","This paper presents a methodology to disambiguate free-text author/committer identities across the World of Code (WoC) collection (V2604, about 107M distinct author strings over about 6B commits) into canonical persons. It extends ALFAA-style fingerprinting and prior multi-million identity resolution by targeting over-merge errors at scale, where bridge identities (bots, role accounts, placeholders, common-name homonyms) fuse clusters. The work reports a full set of twenty experiments, builds a union-graph cut using sampled betweenness centrality plus an edge classifier trained on 2.6M GitHub no-reply-derived labels, and shows strong AUC and cluster-size reduction while improving gold recall and maintaining precision.","Scaling Author Identity Disambiguation to the World of Code: A  \nMethodology  \nAudris Mockus  \nUniversity of Tennessee  \nKnoxville, USA  \narXiv :2607 .06920v 1 [ cs . SE] 8 Jul 2026  \nAbstract  \nWe describe the methodology used to alias the free-text author/committer identities of the entire World of Code (WoC) collection (version V2604, ∼107M distinct author strings over ∼6B commits) into canonical persons, extending the fingerprint-based anti-aliasing of ALFAA [1] and the 38M-id resolution ofFry et al. [3] by an order of magnitude. The central engineering problem at this scale is overmerge rather than missed merges: a small number of bridge identities (bots, role accounts, placeholder emails, multi-author commit fields) transitively weld otherwise-unrelated clusters through the global union step into million-member “mega-clusters.” We report the full experimental record (twenty experiments, including every unsuccessful approach) that led to the deployed design. Node-level gates (information score, project spread, link degree) preserve recall but cannot dissolve the mega; per-value blocklists of bad highquality attributes are recall-safe but cannot break a redundant mesh; the working composition is a betweenness-centrality cut over the exact union graph plus a per-edge classifier trained on 2. 6M labels mined for free from GitHub no-reply ids. The same classifier, filtering the expansion of dormant cross-project shingle groups, joined by GitHub’s own account assertions mined from no-reply ids, then recovers the recall the precision work had foregone. Against humanadjudicated pairs the deployed map’s per-edge model transfers at AUC 0. 99; end to end, the largest cluster falls from 170,431 (and a predecessor’s 3. 0M) to under 7,000 while gold recall rises from 0.44 to 0.70 at increasing precision, and on an independent 21M-alias GitHub ground truth the final map outscores both its predecessorsand the published state ofthe art among global, privacy-preserving resolvers. The record doubles as a catalog of scale lessons: structural cuts do not transfer to edge sets they never saw, gradient boosting shortcut-learns label-construction artifacts that a linear model survives, and recall-only and precision-only benchmarks invert verdicts unless read together.  \n1 Introduction  \nNearly every quantitative claim about open-source software rests on knowing who: developer productivity and turnover, project health and bus factors, the provenance chains behind supply-chain security, the demographics of contribution. Git records authorship asa free-text string, and developers scatter their work across name spellings, work and personal emails, usernames, and, increasingly, deliberate anonymization. Resolving these strings to persons (dealiasing) is therefore foundational infrastructure for mining software repositories, and getting it wrong is worse than not doing it: aresolver that welds strangers together fabricates super-developers, corrupts network analyses, and overrides explicit privacy choices.  \nThis paper documents, as a complete experimental record, the construction of the author identity map for World of Code (WoC) [4]  \nversion V2604: ∼107M distinct author ids over ∼6B commits, an order of magnitude beyond the largest published resolutions [3] . At this scale the dominant failure mode inverts from the literature’s usual concern. Classical de-aliasing maximizes recall against missed merges; in a global transitive union over 108 identities, the binding constraint is over-merge: a handful of bridge identities (placeholder emails, role accounts, template author strings, commonname homonyms) weld unrelated clusters into mega-components that absorb millions of people. Our predecessor production map carried a 3. 0M-member cluster; the first ungated union of V2604 produced a 170,431-member one. Most of this paper is the systematic dissolution of that cluster without sacrificing the legitimate merges around it, followed by the recovery, once ","cbCaiusF4NLeFl60","https://ap.wps.com/l/cbCaiusF4NLeFl60","pdf",767812,1,19,"English","en",105,"# Abstract\n# Introduction\n## Problem and motivation\n## Contributions\n# Methodology and experiments\n## Over-merge diagnosis\n## Labeling from GitHub no-reply ids\n## Composed design: union-graph cut and edge pruning\n## External evaluation","[{\"question\":\"What is the main challenge when disambiguating author identities at WoC scale?\",\"answer\":\"Over-merge dominates: a small set of bridge identities (e.g., bots, role accounts, placeholders, homonym defaults) can transitively fuse otherwise unrelated clusters into very large mega-clusters.\"},{\"question\":\"How does the approach obtain training labels without manual annotation?\",\"answer\":\"It mines GitHub no-reply ids embedded in raw commit strings to generate 2.6M labeled identity pairs at zero annotation cost, then trains an edge classifier to transfer gate-signal features to human-adjudicated labels.\"},{\"question\":\"What does the final deployed design do to avoid mega-clusters while keeping legitimate merges?\",\"answer\":\"It reconstructs the exact union graph, applies a sampled-betweenness-centrality cut over key nodes, prunes residual homonym fragments at the edge level, and expands dormant cross-project shingle groups using classifier filtering plus GitHub-asserted same-account edges.\"}]",1784193522,48,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"scaling-author-identity-disambiguation-to-the-world-of-code-a-methodology","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/scaling-author-identity-disambiguation-to-the-world-of-code-a-methodology/84160/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the main challenge when disambiguating author identities at WoC scale?","Question",{"text":75,"@type":76},"Over-merge dominates: a small set of bridge identities (e.g., bots, role accounts, placeholders, homonym defaults) can transitively fuse otherwise unrelated clusters into very large mega-clusters.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the approach obtain training labels without manual annotation?",{"text":80,"@type":76},"It mines GitHub no-reply ids embedded in raw commit strings to generate 2.6M labeled identity pairs at zero annotation cost, then trains an edge classifier to transfer gate-signal features to human-adjudicated labels.",{"name":82,"@type":73,"acceptedAnswer":83},"What does the final deployed design do to avoid mega-clusters while keeping legitimate merges?",{"text":84,"@type":76},"It reconstructs the exact union graph, applies a sampled-betweenness-centrality cut over key nodes, prunes residual homonym fragments at the edge level, and expands dormant cross-project shingle groups using classifier filtering plus GitHub-asserted same-account edges.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]