Key points are not available for this paper at this time.
This has been a computer scientists revolution; but we all share in its results. Only a few years ago, the World Wide Web was known just to a small research community; it is hard to remember that the voluminous content we see on the WWW, expanding by an estimated million pages each day, has grown up around us in so short a time. Researchers now release their results to the Web before they appear in print; corporations list their URLs alongside their toll-free numbers; news media and entertainment companies vie for the attention of a browsing audience. The Web has become the most visible manifestation of a new medium: a global, populist hypertext. The speed with which this medium has emerged is a testament to the universality of the computational models on which it is built; much of the software and network infrastructure supporting the Web was developed long before there was a Web to support. In much the same way, when we investigate the structural properties of the WWW, we may make use of well-studied models of discrete mathematics-- the combinatorial and algebraic properties of graphs. The Web can be naturally modeled as a directed graph, consisting of a set of abstract nodes (the pages) joined by directional edges (the hyperlinks). Hyperlinks encode a considerable amount of latent information about the the underlying collection of pages; thus, the structure of this directed graph can provide us with significant insight into its content. Within this framework, we can search for signs of meaningful graph-theoretic structure; we can ask: What are the recurring patterns of linkage that occur across the Web as a whole?
Jon Kleinberg (Wed,) studied this question.