The WWW and the Internet are outstanding counterexamples to this. Over the past 25 years, the Internet has grown at an amazing rate; no one knows exactly how large it is or how much traffic it really carries. The Internet was not programmed according to specifications derived from requirements; it evolved. The Internet Engineering Task Force (IETF) works from a distinctly nonlinear motto: "rough consensus, running code, (6)" which is a concise statement of the fitness function for survival of new ideas on the Internet. The Internet clearly conforms to Holland's description of complex systems: intrinsically dynamic, far from a global optimum, and continually adapting to new circumstances.
How is it that the Internet continues to thrive and spontaneously produce new emergent systems such as the World Wide Web and Mosaic? The Internet is surely one of the most massive intellectual adventures of our century, but how can this massive, complex system survive without central authority and control? What accounts for this vitality?
The Internet cannot be explained in terms of linear concepts. The whole is greater than the sum of its parts. One cannot study the TCP/IP standard and understand the Internet's behavior any more than one can study an ant and understand the ant colony's behavior.
The Internet must be understood in nonlinear, evolutionary terms. It is an organic process: successful ideas are replicated; unsuccessful ones die away. One model might be to consider the Internet to be one continuous usability test. Modifications which are usable (measured by the fact that they are used) propogate and prosper on the net. These changes, in the model of Jonas Salk, are "morphic." (7) Changes which absorb more resources than they produce are "entropic" and soon disappear from the Internet landscape.
