First we must confront the question of what happened during the late 1990s. Viewed from 2003, such an exercise is undoubtedly premature, and must be regarded as somewhat speculative. No doubt a clearer view will emerge as we gain more perspective on the period. But at least I will offer one approach to understanding what went on.
I interpret the Internet boom of the late 1990s as an instance of what one might call ``combinatorial innovation.''
Every now and then a technology, or set of technologies, comes along that offers a rich set of components that can be combined and recombined to create new products. The arrival of these components then sets off a technology boom as innovators work through the possibilities.
This is, of course, an old idea in economic history. Schumpeter [1934], p. 66 refers to ``new combinations of productive means.'' More recently, Weitzman [1998] used the term ``recombinant growth.'' Gilfillan [1935], Usher [1954], Kauffman [1995] and many others describe variations on essentially the same idea.
The attempts to develop interchangeable parts during the early nineteenth century is a good example of a technology revolution driven by combinatorial innovation.3 The standardization of design (at least in principle) of gears, pullies, chains, cams, and other mechanical devices led to the development of the so-called ``American system of manufacture'' which started in the weapons manufacturing plants of New England but eventually led to a thriving industry in domestic appliances.
A century later the development of the gasoline engine led to another wave of combinatorial innovation as it was incorporated into a variety of devices from motorcycles to automobiles to airplanes.
As Schumpeter points out in several of his writings (e.g., Shumpeter [2000]), combinatorial innovation is one of the important reasons why inventions appear in waves, or ``clusters,'' as he calls them.
... as soon as the various kinds of social resistance to something that is fundamentally new and untried have been overcome, it is much easier not only to do the same thing again but also to do similar things in different directions, so that a first success will always produce a cluster. (p 142)
Schumpeter emphasizes a ``demand-side'' explanation of cluster of innovation; one might also consider a complementary ``supply-side'' explanation: since innovators are, in many cases, working with the same components, it is not surprising to see simultaneous innovation, with several innovators coming up with essentially the same invention at almost the same time. There are many well-known examples, including the electric light, the airplane, the automobile, and the telephone.
A third explanation for waves of innovation involves the development of complements. When automobiles were first being sold, where did the paved roads and gasoline engines come from? The answer: the roads were initially the result of the prior decade's bicycle boom, and gasoline was often available at the general store to fuel stationary engines used on farms. These complementary products (and others, such as pneumatic tires) were enough to get the nascent technology going; and once the growth in the automobile industry took off it stimulated further demand for roads, gasoline, oil, and other complementary products. This is an example of an ``indirect network effect,'' which I will examine further in section 10.
The steam engine and the electrical engine also ignited rapid periods of combinatorial innovation. In the middle of the twentieth century, the integrated circuit had a huge impact on the electronics industry. Moore's law has driven the development of ever-more-powerful microelectronic devices, revolutionizing both the communications and the computer industry.
The routers that laid the groundwork for the Internet, the servers that dished up information, and the computers that individuals used to access this information were all enabled by the microprocessor.
But all of these technological revolutions took years, or even decades to work themselves out. As Hounshell [1984] documents, interchangeable parts took over a century to become truly reliable. Gasoline engines took decades to develop. The microelectronics industry took 30 years to reach its current position.
But the Internet revolution took only a few years. Why was it so rapid compared to the others? One hypothesis is that the Internet revolution was minor compared to the great technological developments of the past. (See, for example, Gordon [2000].) This may yet prove to be true-it's hard to tell at this point.
But another explanation is that the component parts of the Internet revolution were quite different from the mechanical or electrical devices that drove previous periods of combinatorial growth.
The components of the Internet revolution were not physical devices as all. Instead they were ``just bits.'' They were ideas, standards specifications, protocols, programming langu