Issue No.08 - August (2011 vol.44)
Published by the IEEE Computer Society
David Alan Grier , George Washington University and a senior member of the IEEE Computer Society
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MC.2011.254
As with all engineering problems, we need to balance the twin factors of stability and control and do so in a way that doesn't damage either of them. This article is also available through David Alan Grier's blog, The Known World (http://www.computer.org/theknownworld).
On closer examination, the car wasn't as distinctive as it appeared. To identify it as a Chicken Bus or, more accurately, a Chicken Sports Car, is a disservice to the Central American mechanics who've learned how to adapt old American school buses to serve as a transportation system that can carry people across the countryside and deliver chickens to the city, goats to the slaughterhouse, or cash crops to market. To advertise their special role in the local transportation system, these vehicles are usually painted with gaudy colors and clever decorations and are called Chicken Buses.
The vehicle that was circling the block had the paint job of a Chicken Bus but the attitude of something very different. It was a low Italian sports car of recent vintage, perhaps a Lamborghini, and it had been decorated in a dark and brooding checkerboard scheme that included arbitrary accents of color and enticing portraits of beautiful women and strong men.
We heard the car long before we saw it. The four of us were slowly walking along the street on our way to dinner when we heard sounds that we interpreted as coming from a badly rusted exhaust system. As we looked for the source of the noise, we saw the Chicken Lamborghini accelerate for about 10 or 15 yards, accompanied by a loud roar. As the car got close to a line of stopped traffic, the driver hit the brakes, which produced a plume of smoke from the rear tires and three or four loud belches from the tailpipe.
We watched the car circle the block three times, each with the same dramatic effects. The third time, the driver was caught at a light and decided to rev his engine, perhaps just to prove that he could.
"Poor guy," said one in our group. "He's got the itch, and he's got it bad."
"At least $250,000 bad," said another, who seemed to know the price of handmade automobiles.
"And nobody's paying attention to him," added a third.
"Certainly not the kind of anybody that he hopes is going to be impressed," returned the first. "Just because you can make a loud noise doesn't mean you have anything to say."
Different Types of Innovations
The Chicken Car quickly faded from the conversation as we sat down to dinner. We had been attending a technical meeting all day and had more interesting things to discuss. In particular, I could report that I had been sitting next to a young engineer who appeared to be using a radical version of a popular tablet computer, which had an operating system unlike anything I'd ever seen. The interface involved spinning globes and twisted ribbons. It had a file system that seemed quite strange to me but was completely obvious to the owner. As the machine responded to commands, the screen would unleash various animated objects that seemed to represent data or applications, or both, or neither.
At a break, I asked the young man about the machine, and he confirmed that it was indeed a recent model tablet that was running an operating system of his own design. Something in my voice must have expressed shock or disbelief, for he immediately described the system's origins. It was based on code developed at a university in southeast Asia, perhaps Kuala Lumpur. He had removed the unnecessary sections, written a new kernel and a few drivers, merged some key modules from an underutilized American operating system, and restructured the entire thing.
"It's lean," he said, "and fast. It has none of the stovepipes that you find on commercial operating systems."
I asked if it could run any of the conventional tablet applications.
"No," he confessed, but he noted that he'd been able to import all the software he needed.
Few of us can assemble a modern operating system, even if we're borrowing parts from other sources. Yet, like the mobile work of art that circled the block as we walked to dinner, this machine seemed to be entirely a personal accomplishment. It demonstrated the depth of his skills, but it had little impact on a bigger community. For the moment, the opportunities for introducing a major new tablet operating system seem to be limited. Those who guide new ideas toward widespread application are looking in other places for innovation.
Both the Chicken Car and the Chicken Tablet, if we can call it that, represent types of innovation. The first is a conventional concept in a rare and expensive container. It speaks to the need of young men to make a mark for themselves and get attention from others. The second is an unusual concept in a more common shell. It may represent a true novelty, but it also may be so personal that it isn't possible to label it as a valid form of innovation.
The nature of innovation is an issue that's of great concern to many of my colleagues. By innovation, they mean the ability to create new goods or services that produce wealth and thereby increase accumulated capital. They're concerned that we aren't generating enough innovation, that it's happening in the wrong part of the world, or that it's failing because of a lack of capital, proper education, moral fiber, or a good, filling breakfast.
Most important, they're concerned that we aren't seeing enough innovation because the wrong people are in control and wrong ideas are holding sway. The "future unfolding right now is very different from past innovation," wrote one critic of current innovation strategies. This writer claimed that the last two decades of the 20th century were a golden age in which open computer architectures and open networks encouraged inventors to develop new products and services. "The future is not one of generative PCs attached to a generative network," he warned. "It is instead one of sterile appliances tethered to a network of control."
Building Black Boxes
Certainly, the technology industry flourished in an environment with open architecture machines that allowed individuals with little capital—and sometimes limited training—to develop innovations. Yet simultaneously, that industry has advanced by an opposite process, one that closes technologies and connects them to existing bodies of knowledge. This process locks ideas into black boxes, which have interior mechanisms that can't be altered. "Science has two faces," wrote philosopher Bruno Latour, "one that knows and the other that does not know yet." Both of these elements are needed for technology to advance.
Almost any technical innovation closes an open system, limiting access to ideas that were once easily modified. A central example within the field of computing technology is E.F. Codd's development of the relational database. In 1970, Codd published a paper that not only established the fundamental ideas for databases but also illustrated the pattern by which we define a black box for technology. In his paper, Codd described a basic set of ideas that became central to database design. "We take for granted the benefits brought to us by relational databases," explains a recent college textbook.
Strictly speaking, the process of building a black box shouldn't be confused with the process of invention, the act of being the first to identify and describe an idea. Over the course of my career, I've received many e-mails from computer scientists who feel that they didn't receive proper credit for ideas they discovered or invented. While I sympathize with these researchers and always try to find an interesting story that values their work, I have to remember that our society honors those who provide the most utility from an idea. The first across the line aren't necessarily those who bring the greatest value.
In building a black box, a researcher is creating a self-contained system that can operate without exposing its inner workings. Hence, most of the ideas that form the black box will already exist. The ideas that provided the basis for relational databases had been discussed for almost a decade before Codd wrote his seminal paper. Indeed, in 1966, workers at RAND created a database system that had most if not all of the features of a relational database.
When he defined the concepts of relational databases, Codd was far from the most distinguished researcher in the field. Most of his writings dealt with problems of timesharing. He'd only been studying databases for a few years. Yet, he was able to extract the basic concepts of data manipulation from its surrounding environment and place it in a framework that many found convenient. The "problems treated here are those of data independence," he wrote. "[This paper] provides a means of describing data with its natural structure only—that is, without superimposing any additional structure for machine representation purposes."
Codd's work required about 10 years to establish itself as a unified technology, as a set of ideas that could be used without dealing with their internal operations. The first recognition came in the academic literature. His paper had 59 citations by 1975 and another 460 by the end of the decade. In parallel with the approval of researchers came interest from industry. By 1980, a half-dozen firms were offering relational database software, including two start-up companies that would have major roles in the software industry.
No matter how quickly ideas are accepted by researchers or form the basis for an industry, they won't survive without a careful and thorough defense. There are always others who think they have a better idea, a more productive plan, a greater stake in an alternative approach. For the next 20 years, Codd defended his ideas as the best ways to create and manage data. In making this defense, Codd was backed by the resources of his employer, IBM, which had made substantial investments in relational databases. With these resources, Codd focused on the technical value of his database. "After the publication of my papers," he explained, "numerous articles began to appear proposing new approaches to database management." He claimed that most of these articles attacked his model with "the false claim that the relational model contains no features for representing the meaning of the data." He would then dismiss these claims and argue that any new contribution could be incorporated into his model.
For the past five or six years, policymakers have talked about the need to create a system of innovation that they describe as generative. In using this term, they're usually referring to the "capacity to produce unanticipated change through unfiltered contributions from broad and varied audiences." They debate various strategies of creating such a system and argue about whether it's best done in open source environments, under the control of private capital, or under the watchful eye of a central government.
While the watchful central government has perhaps been the least effective shepherd of innovation, it has still played a role in the process of closing and opening technology that has spurred new ideas. At key moments, we can declare one technology closed and stable so that researchers can use it as a tool for future development. The forces that close and stabilize a technology can be government regulation, the clout of capital, or the mutual agreement of wary parties who believe that a stable market is the best place for them to develop new ideas. Even in the most congenial market, they'll have to do battle with the army of invisible hands that guides prices to a point that balances supply and demand, evaluates innovation, and leads some to success while it crushes the rest beneath its heavy blow.
My colleagues have also been concerned that the global systems of innovation were failing to inspire truly original ideas because too many technologies were in the hands of too few institutions. After generations of producing wonderful ideas, we were reduced to a situation in which we could do little with those ideas. The owners might let us borrow an idea for a time, give it a clever coat of paint, and drive it around the block in the vain hope that we might encounter someone or some idea of interest. Just because you can combine old ideas doesn't mean that any of those combinations will be novel.
As with all engineering problems, we need to balance the twin factors of stability and control and do so in a way that doesn't damage either of them. We must provide standard concepts and ideas so that anyone can build a new database, a new processor, or a new car. We also need to support a venue that will emphasize the advantages of the new ideas. It may look like a Chicken Bus to a group of critical guys standing on a street corner, but someone may be intrigued to give it a spin and see where it will go.
Selected CS articles and columns are available for free at http://ComputingNow.computer.org.
David Alan Grier is an associate professor of international science and technology policy at George Washington University and a senior member of the IEEE Computer Society. Contact him at firstname.lastname@example.org.