said. Not only was the Web a high-profile example of a vast, elaborate network, it made many other networks accessible for research. Web crawlers and search engines made it possible to map out the links of the Web itself, of course, and the Web also made it possible to catalog other large networks and store them for easy access (the database of movie actors being one prime example). Data on the metabolic reactions in nematode worms or the gene interactions in fruit flies could similarly be collected and transmitted.

“Big databases became available, and researchers could get their hands on them,” Strogatz observed. “People started to think about things as networks.” Before that, he said, even actual networks weren’t usually viewed in network terms—the electric power network was known as a grid, and you were just as likely to hear the term telephone “system” as telephone network. “We didn’t think of them so much as networks,” Strogatz said. “I don’t think we had the visceral sensation of moving through a network from link to link.”

With the Web it was different. It was almost impossible to think of it as a whole. You had to browse, link by link. And the Web touched all realms of science, linking specialists of all sorts with network ideas. “In many different branches of science,” Strogatz observed, “the kind of thinking that we call network thinking started to take hold.”

Still, the revolution in network math did not begin until after the Watts-Strogatz paper appeared in 1998. They showed how to make a model of a “small-world” network, in which it takes only a few steps on average to get from any one node of the network to any other. Their model produced some surprises that led to a flurry of media coverage and the subsequent network mania. But Strogatz thinks some of those surprises have been misrepresented as being responsible for network math’s revival. Some experts would say, for example, that the Watts-Strogatz paper’s major impact stemmed from identifying the small-world nature of some particular real-world networks. Others have suggested that “clustering” of links (small groups of nodes connected more than randomness would suggest) was the key discovery. “This is to me the bogus view of



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement