The Anatomy of a Large-Scale Hypertextual Web Search Engine

This is the original PageRank paper, written when Brin and Page were Stanford graduate students and Google was still called BackRub. The paper describes a system for ranking web pages by treating hyperlinks as citations — a page is important if important pages link to it — and explains the crawling, indexing, and ranking architecture of their prototype search engine. It is a short read, technically accessible, and carries enormous historical weight as the founding document of the company that would restructure the digital economy. The authors are refreshingly candid about the problems of advertising-funded search, noting in a now-famous appendix that advertising incentives are "inherently mixed" with the goal of returning relevant results. For product people, the paper is a lesson in how a single architectural insight — treating the web's link structure as a quality signal — can create a product category. Reading the original is a corrective to every secondhand summary.