THE SERENDIPITY MACHINE

A look at the unexpected consequences of the computer revolution.

Green (Information Technology/Monash Univ., Australia) sees the key impact of the growth of computerization as being an increase not just in the quantity but in the complexity of information available. Comparing the capabilities of early computers and those currently in wide use, he estimates a 100,000-fold increase in data from the mainframes of 20 years ago to the desktop model he uses today. Even more radical growth is evident on the Internet, which 15 years ago was used almost entirely by academics and the defense establishment. More significantly, the ability to combine sets of data has increased, allowing the discovery of hitherto unexpected relationships and also creating a new degree of complexity. Computer design has developed tools to help users handle this complexity; one such tool, for example, is modularity, the breaking-down of complex tasks into smaller subroutines, not unlike filing systems that sort data first by broad categories, then by narrower ones. A similar principle allows passing messages over large networks in an economical number of steps. The linear arrangement of data characteristic of the book is no longer necessary for computers, which can easily find connections between randomly sorted data. An example is data mining, or the use of computers to discover relationships in a vast quantity of data—say, analyzing computerized checkout records in grocery stores to find individual customers’ buying patterns. Similarly, biotechnology has bloomed as computers have made it possible to analyze genetic data, with results both positive (as in new therapies) and ethically nightmarish (as in cloning). Likewise, the global village created by the Internet creates opportunities both for increased understanding and for increased crime, from Nigerian e-mail scams to massive terrorist attacks. While Green is clearly a cybernetic booster, he doesn’t duck hard questions.

Sometimes stiff, but should reward thoughtful readers.