Albert-László Barabási. Linked.
Albert-László Barabási's Linked provides an overview of the scientific research into the nature of networks, beginning, for the most part, with the random network theories of Erdős and Rényi, formulated in 1959, and extending to Barabási's research as recently as 2003. Barabási finds that most of the complex networks that have been studied, including the internet, aspects of the metabolic and regulatory functions of cells, aspects of language, social webs, and the networks of Hollywood actors share a generic, scale-free topology. While the various quantities found in most naturally occurring phenomenon follow a bell curve, which would, for example, yield a characteristic scale or average of node connectivity, with a rapidly decaying curve preventing nodes with a degree of conectivity that deviates signifantly from this scale, the quantities describing the connectedness of nodes in a large network instead conform to power laws. Histograms characterizing power law distributions display a continuously decreasing curve that, in the case of networks, imply a large majority of nodes with small, relatively similar degrees of connectivity with a small number of nodes with far more connections. For example, while Barabási's measurements of a sample of 203 million Webpages indicated that 90 percent of those pages had fewer than 11 links pointing to them, 3 pages were linked to by more than a million others. Such nodes are known as hubs and help to hold the network together.
These findings lead me to question whether the apparent ubiquity of scale-free networks is merely a consequence of a functional superiority that nevertheless leaves the door open for actualizations of other kinds of topologies, or whether the apparent laws of large networks do, in fact, rule out large, socially relevant actualizations of Deleuze and Guattari's rhizomatic networks.