In an ACM QUEUE article "Evolution or Revolution?", Mache Creeger complains: "While there are some genuine revolutions in the technology industry,... the vast majority of new products are evolutionary, not revolutionary." Creeger looks back on an earlier time and says that "three decades ago, I believed the world would have changed dramatically by now. Pure research in computer science was introducing new concepts at breathtaking speeds. ...[but] It seems to me that the pace of innovation has actually slowed." He attributes the slowing of innovation to the fact that "the vast majority of intellectual property became privately held," there has been a "decline of pure research in computer technology," and to "the massive accrual of business-method and algorithm patents."
Creeger clearly doesn't understand evolution very well... At least not as evolution is understood today. His complaint that the apparent "slowing" of innovation indicates that something is wrong and his contrasting of "Evolution" with "Revolution" indicate belief in the now old-school uniformitarian idea of Phyletic gradualism which argues that "evolution occurs at a fairly constant rate." (i.e. If innovation (evolution) was once moving at "breathtaking speeds," then something is "wrong" if that speed is now reduced.) However, Creeger's argument holds within it evidence for the now broadly accepted theory of punctuated equilibrium which explains that evolution often proceeds sporadically. (i.e. Evolution is expected to be both gradual and "revolutionary.") With some evident sadness, Creeger looks back on the days of his youth and suggests that they were a veritable "Cambrian Explosion" of innovation while bemoaning today's relatively slower, more-Permian rate of innovation.
Modern evolutionary theory teaches us:
- Evolution proceeds via both small and large innovations (mutations).
- Over long periods of time, there are vastly more small innovations than large ones.
- Over long periods of time, the distribution of innovation "size" conforms to a power-law distribution.
- Over short periods of time, evolution may be seen to be either gradual (dominated by many small innovations) or revolutionary (dominated by major innovations).
- Major innovations tend to spur bursts of minor innovations.
- Over time, the accumulation of a large number of minor innovations lays the foundation for later major innovations.
Given the above, and my own 30+ years experience in this business, it is clearly the case that the period of "breathtaking" innovation that Creeger experienced back in the 70's and 80's was simply an expected case of the occasional "Cambrian" period in computer science -- particularly as it related to distributed systems. It was a period in which a small number of major, but profound, enabling innovations impacted our industry and has been followed by avalanches of less significant but nonetheless wonderful innovations that continue today. This is the way of evolution -- punctuated equilibria -- gradualism interwoven with revolutions.
Thirty years ago, we were just beginning to bring the first networks on-line and we were just beginning to apply computers to every-day tasks -- rather than the tasks of scientists and engineers. We had the luxury and fun of being first to have a whole set of big ideas... Those of us who were working then could look forward to a vast "green field" of applications that had simply not been visible to those who came before us. The result was that it was quite easy for us, over a cup of coffee with friends, to come up with ideas more original and more innovative than could be generated today by hundreds of the industry's best thinkers after months of concerted work. But, it is important to know that today's thinkers are no less smart and no less innovative than were the folk working "back in the day." The difference is that today we're all still focused on working through the implications of the last revolution. In time we'll exhaust the realm of easily achieved secondary innovations and we'll then be ready to move on to more revolutionary "Cambrian" times again. It is always like this. It always has been and it always will be.
Creeger lists a number of what he calls "genuine revolutions." His list includes: "cellphones, GPS (global positioning system), quantum computing, encryption, and global access to content." It's an interesting list but I think these are actually mostly minor innovations. The real revolutions are rooted in ideas and "memes," not technologies or products. Products and technologies are like the "individuals" or "species" of our business. Ideas are the genesis of whole new phyla and classes. For instance, the revolutions that impacted our industry at the beginning of the current period were triggered by the spread of ideas like "communications between machines," "non-scientific computing," "content on demand," "ubiquitous connectivity." These are the root, innovative ideas that defined whole new realms of computing and set off the avalanches of secondary innovations like global access to data through TCP/IP and packet-switching, Office Automation and email, iTunes/Tivo/PodCasts/VOD, cell phones and the Web. Of course, each of these secondary innovations set off its own avalanche of innovation. The process of innovation is fractal.
I will spare you a long dissertation quoting the math-heavy and theoretical writings of Stuart Kauffman, Per Bak, Stephen J. Gould, and the work of the Santa Fe Institute... For a less theoretic and more easily approached discussion of evolution and innovation, one might read the works on "innovation" (and thus evolution) authored by Clayton Christensen of the Harvard Business School. In his popular books ("The Innovator's Dilemma" and "The Innovator's Solution") Christensen's discussion of the distinction between "sustaining innovation" and "disruptive innovation" closely mirrors what we know of evolution. Sustaining innovation is characterized by improvements to an existing product or process while disruptive innovation creates new product categories and new markets. Sustaining innovation is typically a continuous and frequent occurrence in competitive markets while disruptive innovations are vastly more rare. It is intriguing that Christensen focuses a great deal on the idea that "disruptive innovations" are often "not as good" when measured against the the products that already exist in the markets into which they enter. Disruptive Innovations attack evolutionary niches not already addressed by the market and thus can be best at what they do without being "better" than what was already there. Only after a great deal of sustaining innovation do disruptive innovations get built out to the point where they are so fully and rigidly defined that they themselves become the target of future disruptive innovations... It takes time to work through enough innovation to motivate and enable the next cycle of disruption. It takes time to move between two Cambrian eras.
In closing, I should probably say that I agree with Creeger that the current state of the patent system is an absolute mess. It seems possible today to get a patent on even the most obvious of processes and the result is a real hampering of innovation. However, I disagree with his conclusion that the failings of the patent system are a primary cause of the apparent reduction in the pace of innovation in our industry. Most patents are focused on the secondary or sustaining innovations that are derivative of the great ideas or memes that trigger truly "breath-taking" periods of innovation. Those ideas and memes are not, however, typically patented or even patentable since the truly revolutionary ideas usually spread long before being reduced to method and process.
bob wyman
