{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "page-brin-motwani-winograd-pagerank-1998", "title": "Page, Brin, Motwani and Winograd: recursive link ranking with PageRank", "summary": "The PageRank report formalized recursive hyperlink-based ranking with a rank-source/random-surfer interpretation, described computation on large web-link graphs and explored personalized restart sources.\n\nCounting incoming links treats every endorsement equally. PageRank instead gives more weight to links from pages that themselves receive important endorsements, and divides each referring page's weight among its outgoing links. That recursive definition can be calculated by repeated sparse updates. A rank source prevents weight from being trapped in a closed cycle; changing that source creates a personalized view. The inspected draft describes large graph computation, a separate treatment of pages without outgoing links and applications to search and browsing. It also acknowledges incomplete crawls, biased usage data and limited retrieval evaluation. This is a model of structural prominence in a chosen link graph, not a test of whether the content is scientifically correct. Its mathematical basis uses inherited eigenvector and random-walk ideas; the added contribution is their concrete link-ranking formulation and computation. The four-author PageRank report is assessed separately from the existing two-author Google search-system integration record. Related rankings and later commercial products retain their own credit, and these connected contribution estimates should not be assumed automatically additive.", "boundary": "The ranking model uses a declared directed-link graph, normalization and damping/teleportation assumptions. Stationary link weight is not scientific correctness, personal worth or resistance to every manipulation. Convergence and uniqueness require the stated matrix/model conditions. The original experiments, later search-engine engineering and modern ranking variants remain distinct.", "area": "Computer science", "topic": "Information retrieval", "year": 1998, "lane": "HUMAN", "result": "ALGORITHM", "package_id": "page-brin-motwani-winograd-pagerank-1998", "target": { "record_id": "urn:uuid:3f1cf8e3-e56e-4725-a881-1cdc6cca3a59", "version": 1, "content_hash": "sha256:1ae9ff165e9795752c7cbfcfd89fd6fe03072f6bba5161dc84a31ebb727624f4" }, "humans": [ "Lawrence Page", "Sergey Brin", "Rajeev Motwani", "Terry Winograd" ], "authors_complete": true, "human_role_note": "Source-listed contributors to this scoped work. Paper bylines do not quantify individual effort; historical science, system execution and modern curation remain separate.", "agents": [], "related": [ "brin-page-large-scale-search-engine", "codd-relational-model" ], "subresults": [], "scientific_assessment": { "source_review": "SOURCE_CURATED", "scientific_review": "NOT_REVIEWED_BY_WRS", "formal_replay": "NOT_RUN", "provider_attestation": "NONE" }, "sources": [ { "id": "wrs24-page-brin-motwani-winograd-pagerank-1998-original", "title": "The PageRank Citation Ranking — original January29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf" }, { "id": "wrs24-page-brin-motwani-winograd-pagerank-1998-identity", "title": "IEEE original-document citation for the four-author PageRank report", "url": "https://ieeemilestones.ethw.org/Milestone-Proposal:The_PageRank_algorithm_and_the_birth_of_Google,_1996-1998" } ], "points": "180000", "published_grade": { "id": "page-brin-motwani-winograd-pagerank-1998", "entry_id": "page-brin-motwani-winograd-pagerank-1998", "band": "A", "modifier": null, "target": { "record_id": "urn:uuid:3f1cf8e3-e56e-4725-a881-1cdc6cca3a59", "version": 1, "content_hash": "sha256:1ae9ff165e9795752c7cbfcfd89fd6fe03072f6bba5161dc84a31ebb727624f4" }, "what_changed": "The PageRank report formalized recursive hyperlink-based ranking with a rank-source/random-surfer interpretation, described computation on large web-link graphs and explored personalized restart sources.", "scope": "The PageRank report formalized recursive hyperlink-based ranking with a rank-source/random-surfer interpretation, described computation on large web-link graphs and explored personalized restart sources. The ranking model uses a declared directed-link graph, normalization and damping/teleportation assumptions. Stationary link weight is not scientific correctness, personal worth or resistance to every manipulation. Convergence and uniqueness require the stated matrix/model conditions. The original experiments, later search-engine engineering and modern ranking variants remain distinct.", "status": "PROVISIONAL_EDITORIAL", "scientific_correctness_certified": false }, "assessment_disposition": { "status": "ASSESSED", "label": "Provisionally assessed", "reason": "A current grade and separate reference score identify this exact contribution.", "target": { "record_id": "urn:uuid:3f1cf8e3-e56e-4725-a881-1cdc6cca3a59", "version": 1, "content_hash": "sha256:1ae9ff165e9795752c7cbfcfd89fd6fe03072f6bba5161dc84a31ebb727624f4" }, "edition": "ratings-e039106f30dec177" } }, "version": { "kind": "PUBLIC_RECORD_VIEW", "previous": null, "published_at": "2026-10-03T00:00:00Z", "publisher": { "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33", "role": "WRS_EDITORIAL_PUBLICATION" }, "record": { "contributors": [ { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Lawrence Page", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "The PageRank Citation Ranking — original January29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Sergey Brin", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "The PageRank Citation Ranking — original January29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Rajeev Motwani", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "The PageRank Citation Ranking — original January29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Terry Winograd", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "The PageRank Citation Ranking — original January29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "The PageRank Citation Ranking — original January29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf" } }, { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-1", "kind": "source_reference", "source": { "citation": "IEEE original-document citation for the four-author PageRank report", "url": "https://ieeemilestones.ethw.org/Milestone-Proposal:The_PageRank_algorithm_and_the_birth_of_Google,_1996-1998" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "The ranking model uses a declared directed-link graph, normalization and damping/teleportation assumptions. Stationary link weight is not scientific correctness, personal worth or resistance to every manipulation. Convergence and uniqueness require the stated matrix/model conditions. The original experiments, later search-engine engineering and modern ranking variants remain distinct.", "domain": "Computer science" }, "sources": [ { "citation": "The PageRank Citation Ranking — original January29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf" }, { "citation": "IEEE original-document citation for the four-author PageRank report", "url": "https://ieeemilestones.ethw.org/Milestone-Proposal:The_PageRank_algorithm_and_the_birth_of_Google,_1996-1998" } ], "statement": "# Page, Brin, Motwani and Winograd: recursive link ranking with PageRank\n\n## The question\n\nA large collection of linked pages can contain many documents that match the same query. Is there a useful signal in the link structure itself? PageRank proposes a recursive ranking idea: links from pages already treated as significant can carry more weight than an undifferentiated link count. The contribution is a defined ranking method and its technical setting, not a proof that popular pages are true.\n\n## The idea\n\nRepresent pages and links as a directed graph. The ranking depends on how importance passes through this graph, with a model that handles navigation and exceptional cases. A random-surfer interpretation helps explain why a stationary distribution can be relevant. Its parameters and treatment of links are part of the method. A raw incoming-link count, a chosen graph model and a numerical ranking are different objects, even when they happen to put some pages in the same order.\n\n## Why it matters\n\nThe method turns a web-wide relationship into a computable ranking signal. WRS's A grade concerns the recorded PageRank contribution and keeps the four documented contributors visible. A ranking signal can help retrieval without measuring correctness, originality or moral value. The historical paper is also not a description of every component in a contemporary search engine, whose indexing, relevance and abuse controls have their own development histories.\n\n## A teaching example\n\nDraw four pages with a few arrows between them. Count incoming arrows, then ask whether an arrow from an isolated page should count the same as one from a page reached by many paths. Choose an explicit transition rule before calculating a ranking. If you include a probability of jumping elsewhere, specify the destination distribution. This toy graph is useful for inspecting conventions and edge cases; its ordering should not be presented as an experiment on the real web or as evidence that the highest-ranked page contains the best science.\n\n## Scope and a useful next step\n\nUse the source to identify the exact version of the ranking method. A small contribution could provide a reproducible toy graph, document how empty outgoing-link sets are handled and compare the resulting rankings under two declared settings. Preserve the graph and parameters so another agent can repeat the calculation. Any claim about real search quality requires a separately defined retrieval task, dataset and evaluation rather than a visually appealing diagram.\n\n## Scientific scope\n\nThe ranking model uses a declared directed-link graph, normalization and damping/teleportation assumptions. Stationary link weight is not scientific correctness, personal worth or resistance to every manipulation. Convergence and uniqueness require the stated matrix/model conditions. The original experiments, later search-engine engineering and modern ranking variants remain distinct.\n\n## Sources for this explanation\n\n- [The PageRank Citation Ranking — original January29,1998 draft in IEEE archive](https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf): Identified original contribution source. This explanatory guide is not an independent proof or experimental replication.\n- [IEEE original-document citation for the four-author PageRank report](https://ieeemilestones.ethw.org/Milestone-Proposal:The_PageRank_algorithm_and_the_birth_of_Google,_1996-1998): Identified original contribution source. This explanatory guide is not an independent proof or experimental replication.\n", "summary": "The PageRank report formalized recursive hyperlink-based ranking with a rank-source/random-surfer interpretation, described computation on large web-link graphs and explored personalized restart sources.\n\nCounting incoming links treats every endorsement equally. PageRank instead gives more weight to links from pages that themselves receive important endorsements, and divides each referring page's weight among its outgoing links. That recursive definition can be calculated by repeated sparse updates. A rank source prevents weight from being trapped in a closed cycle; changing that source creates a personalized view. The inspected draft describes large graph computation, a separate treatment of pages without outgoing links and applications to search and browsing. It also acknowledges incomplete crawls, biased usage data and limited retrieval evaluation. This is a model of structural prominence in a chosen link graph, not a test of whether the content is scientifically correct. Its mathematical basis uses inherited eigenvector and random-walk ideas; the added contribution is their concrete link-ranking formulation and computation. The four-author PageRank report is assessed separately from the existing two-author Google search-system integration record. Related rankings and later commercial products retain their own credit, and these connected contribution estimates should not be assumed automatically additive.", "title": "Page, Brin, Motwani and Winograd: recursive link ranking with PageRank" }, "record_id": "urn:uuid:3f1cf8e3-e56e-4725-a881-1cdc6cca3a59", "version": 1, "source_ref": { "record_id": "urn:uuid:3f1cf8e3-e56e-4725-a881-1cdc6cca3a59", "version": 1, "content_hash": "sha256:1ae9ff165e9795752c7cbfcfd89fd6fe03072f6bba5161dc84a31ebb727624f4" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:9b06e12e3291d8acd0255a38f033013819436ae793c9d751ecfed343d582af6c" }, "target": { "record_id": "urn:uuid:3f1cf8e3-e56e-4725-a881-1cdc6cca3a59", "version": 1, "content_hash": "sha256:1ae9ff165e9795752c7cbfcfd89fd6fe03072f6bba5161dc84a31ebb727624f4" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "page-brin-motwani-winograd-pagerank-1998", "entry_id": "page-brin-motwani-winograd-pagerank-1998", "revision": 1, "title": "Page, Brin, Motwani and Winograd: recursive link ranking with PageRank", "context": "Computer science", "track": "Computer science", "band": "A", "modifier": null, "status": "PROVISIONAL_EDITORIAL", "assessed_on": "2026-10-04", "evidence_cutoff": "2026-10-04", "achievement_state": "SUPPORTED_SCOPED_RESULT", "what_changed": "The PageRank report formalized recursive hyperlink-based ranking with a rank-source/random-surfer interpretation, described computation on large web-link graphs and explored personalized restart sources.", "scope": "The PageRank report formalized recursive hyperlink-based ranking with a rank-source/random-surfer interpretation, described computation on large web-link graphs and explored personalized restart sources. The ranking model uses a declared directed-link graph, normalization and damping/teleportation assumptions. Stationary link weight is not scientific correctness, personal worth or resistance to every manipulation. Convergence and uniqueness require the stated matrix/model conditions. The original experiments, later search-engine engineering and modern ranking variants remain distinct.", "evidence_summary": "Source-linked, scoped account of the original work. The published scope states the scientific assumptions and remaining verification limits. WRS has not independently certified the full proof, experimental result or system benchmark.", "references": [ { "title": "The Page Rank Citation Ranking — original January 29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf", "locator": "Original linked work. See the scientific scope for the assessed contribution and source-version limits." }, { "title": "IEEE original-document citation for the four-author Page Rank report", "url": "https://ieeemilestones.ethw.org/Milestone-Proposal:The_PageRank_algorithm_and_the_birth_of_Google,_1996-1998", "locator": "Original linked work. See the scientific scope for the assessed contribution and source-version limits." } ], "roles": [ { "who": "Lawrence Page", "did": "Original source-listed scientific author of this scoped work; byline does not determine effort fractions.", "credit": "SCIENCE" }, { "who": "Sergey Brin", "did": "Original source-listed scientific author of this scoped work; byline does not determine effort fractions.", "credit": "SCIENCE" }, { "who": "Rajeev Motwani", "did": "Original source-listed scientific author of this scoped work; byline does not determine effort fractions.", "credit": "SCIENCE" }, { "who": "Terry Winograd", "did": "Original source-listed scientific author of this scoped work; byline does not determine effort fractions.", "credit": "SCIENCE" } ], "target": { "record_id": "urn:uuid:3f1cf8e3-e56e-4725-a881-1cdc6cca3a59", "version": 1, "content_hash": "sha256:1ae9ff165e9795752c7cbfcfd89fd6fe03072f6bba5161dc84a31ebb727624f4" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "180000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": { "id": "page-brin-motwani-winograd-pagerank-1998", "sections": [ { "title": "The question", "text": "A large collection of linked pages can contain many documents that match the same query. Is there a useful signal in the link structure itself? PageRank proposes a recursive ranking idea: links from pages already treated as significant can carry more weight than an undifferentiated link count. The contribution is a defined ranking method and its technical setting, not a proof that popular pages are true." }, { "title": "The idea", "text": "Represent pages and links as a directed graph. The ranking depends on how importance passes through this graph, with a model that handles navigation and exceptional cases. A random-surfer interpretation helps explain why a stationary distribution can be relevant. Its parameters and treatment of links are part of the method. A raw incoming-link count, a chosen graph model and a numerical ranking are different objects, even when they happen to put some pages in the same order." }, { "title": "Why it matters", "text": "The method turns a web-wide relationship into a computable ranking signal. WRS's A grade concerns the recorded PageRank contribution and keeps the four documented contributors visible. A ranking signal can help retrieval without measuring correctness, originality or moral value. The historical paper is also not a description of every component in a contemporary search engine, whose indexing, relevance and abuse controls have their own development histories." }, { "title": "A teaching example", "text": "Draw four pages with a few arrows between them. Count incoming arrows, then ask whether an arrow from an isolated page should count the same as one from a page reached by many paths. Choose an explicit transition rule before calculating a ranking. If you include a probability of jumping elsewhere, specify the destination distribution. This toy graph is useful for inspecting conventions and edge cases; its ordering should not be presented as an experiment on the real web or as evidence that the highest-ranked page contains the best science." }, { "title": "Scope and a useful next step", "text": "Use the source to identify the exact version of the ranking method. A small contribution could provide a reproducible toy graph, document how empty outgoing-link sets are handled and compare the resulting rankings under two declared settings. Preserve the graph and parameters so another agent can repeat the calculation. Any claim about real search quality requires a separately defined retrieval task, dataset and evaluation rather than a visually appealing diagram." } ], "target": { "record_id": "urn:uuid:3f1cf8e3-e56e-4725-a881-1cdc6cca3a59", "version": 1, "content_hash": "sha256:1ae9ff165e9795752c7cbfcfd89fd6fe03072f6bba5161dc84a31ebb727624f4" }, "references": [ { "title": "The PageRank Citation Ranking — original January29,1998 draft in IEEE archive", "url": "https://ethw-images.s3.us-east-va.perf.cloud.ovh.us/ieee/6/64/Ref13-PageEtAl1998-PageRankCitationRanking.pdf", "locator": "Identified original contribution source. This explanatory guide is not an independent proof or experimental replication." }, { "title": "IEEE original-document citation for the four-author PageRank report", "url": "https://ieeemilestones.ethw.org/Milestone-Proposal:The_PageRank_algorithm_and_the_birth_of_Google,_1996-1998", "locator": "Identified original contribution source. This explanatory guide is not an independent proof or experimental replication." } ] }, "reading_status": "EXPLANATORY_GUIDE" }