msnbc COSMIC LOG

Too much networking?

Posted: Thursday, July 23, 2009 7:27 PM by Alan Boyle

Duane Hoffmann / msnbc.com

Open-source communities may suffer from "an overabundance of connections," an information policy researcher suggests in the journal Science.

Are geeks guilty of groupthink? A network expert argues that less social networking would produce more radical innovation on the Internet.

"An overabundance of connections over which information can travel too cheaply can reduce diversity, foster groupthink, and keep radical ideas from taking hold," Viktor Mayer-Schönberger, director of the Information + Innovation Policy Research Center at the National University of Singapore, writes in this week's issue of the journal Science.

That may be one of the reasons why much of the open-source software currently being produced is rarely altered in anything more than an incremental manner, Mayer-Schönberger says.

"The basic point that I'm trying to make is ... how do we get to the next stage of the Internet, the new-generation Internet, the radical innovation, rather than another dot release on the Firefox browser?" he told me today.

Mayer-Schönberger is focusing on the open-source software community because the perils of groupthink are well-known in the commercial world. Once a complicated piece of software catches hold in the market, there's often a "lock-in effect" that freezes out radical changes in that software.

"Every radical change gives the users the opportunity to make a switch to a potential competitor's product, and as a vendor, you don't like that - except if you're Apple," Mayer-Schönberger said. (Apple can rely on a dedicated following that will be quick to adopt radical innovations, such as the iPhone, he explained.)

Open-source software - like the Firefox Web browser, for example, or the Linux operating system - can be freely modified and redistributed by anyone, which would seem to encourage innovation. But Mayer-Schönberger invokes network theory to contend that the open-source community is so interconnected, using the very tools they helped develop, that fresh ideas don't have as much time to develop before they're assimilated (or disposed of) by fellow programmers.

Companies such as Apple, Google, IBM and Microsoft (which is a partner in the msnbc.com joint venture) get around the groupthink trap by creating incubators for research and innovation, modeled after Lockheed Martin's storied "Skunk Works." The key is to have limited linkages between the idea incubators and the larger enterprise, Mayer-Schönberger said.

The payoffs from the innovations that are allowed to hatch outweigh the costs of maintaining the incubators. Mayer-Schönberger said examples of such payoffs range from the original IBM PC and Apple Macintosh to the atom bomb (which was created through a government-funded incubator known as the Manhattan Project).

When there's no incentive for developing unorthodox products or services, and when the network in charge of creating the main product remains highly interconnected, the lock-in effect is more likely to take hold.

The locked-in Internet ... and Facebook
"The most prominent example is not commercial software, but the Internet, or more precisely the protocols underlying this dominant network infrastructure," Mayer-Schönberger wrote in his Science policy paper. "It is too costly and risky for a commercial competitor to create and market a set of radically improved, but incompatible protocols. This is true for the peer-producing, open-source community as well."

Facebook's rise illustrates the pluses and minuses of open source, he told me. "By opening the API (application programming interface) they created a lot of room for experimentation," he said. "That sealed the fate of MySpace and pushed Facebook forward."

Today, there's an entire software ecosystem that relies on Facebook's open API (including Mafia Wars and "25 Random Things"), and that could lead to the lock-in effect. "Changing the API is much harder now because it has an ecosystem that lives off it," Mayer-Schönberger said. "So it's a double-edged sword, in a way."

There's a rule of thumb that says a network becomes more valuable as it adds more connections, but Mayer-Schönberger said that trend could bog down innovation. "It would be terrible if we reach a basically steady state in the open-source community where we have version 11.5.17 and we change to 11.5.18, and everyone thinks it's a step forward," he said.

The prescription for geek groupthink
So what is to be done? On one level, the National Science Foundation is already fostering a "Skunk Works" Internet by supporting next-generation network development programs known as NetSE and GENI.

Mayer-Schönberger said governments could go further by offering incentives for the creation of smaller, competing development teams. This would reduce "their connectedness to the thinking of the status quo through their social networks," he said. The idea sounds a little bit like an X Prize for information infrastructure.

Another strategy, aimed specifically at the open-source community, would be to break a big project into smaller components, and then let separate teams compete to deal with those components. "That is what propelled Firefox into the forefront," Mayer-Schönberger told me. "When they broke (the browser program) into small modules, then they had competitive open-source projects on all those subcomponents."

Mayer-Schönberger was reluctant to extend his analysis to individual behavior, but he said it might be worth your while to take a look at your own social networking. "Think hard about whether those 1,125 Facebook friends are really friends. Think about how many are hangers-on or chance encounters - and perhaps take them off," he said.

You can build diversity into your own social networking by keeping a division between the various aspects of your life - for example, by using LinkedIn for professional contacts and Facebook for personal contacts.

"We now tend to converge our personal and professional lives, and that's not necessarily a good idea," Mayer-Schönberger said. "Having multiple and slightly overlapping networks is better than having one large converged network."

Reality check
Does Mayer-Schönberger's view make sense? I asked one of the pioneers in network science, Northeastern University's Albert-Laszlo Barabasi, for a reality check. He told me the idea of reducing connectedness to encourage innovation was intriguing and provocative.

"I think many people sympathize with the idea that if you could start from scratch, you would have a much different Internet, and a better one," he said.

However, the reality is that the present-day Internet has so much critical mass that there'd have to be a radical reason for introducing radical change. Maybe a nuclear war. Maybe a totally new communication medium. In any case, it'd have to be something big, Barabasi said.

"If you were to ask me, I think we will be using this Internet for quite a while," Barabasi said. "If there will be a new Internet, it's not going to emerge from a private company, because the Internet is just too big for that. ... It may require serious social engineering to get away from the status quo."

10 years of network theory
In the same issue of Science, Barabasi traces the research that's been conducted into the nature of the Internet, the World Wide Web and other scale-free networks over the past 10 years. (That's how long I've been writing about Barabasi and his colleagues.)

"I think the big thing that has happened is that networks are everywhere now. ... They have pretty much invaded all fields of inquiry," Barabasi told me. Other articles in Science discuss how network theory has been applied to cellular biology, biodiversity, economics and social-ecological systems, how technologies (and diseases) spread, how to fight terrorists and how to get along.

The amazing thing is that researchers are finding strong parallels between the workings of the cell and the workings of the Web. "Because they ended up being so similar to each other, the results that were obtained by studying the World Wide Web could be transferred to the study of the cell," Barabasi said.

One of the questions surrounding the past decade of network theory may well be why scientists didn't notice even earlier how much different types of networks had in common. "In a way, the computer made it possible to get large enough data sets to see these features," Barabasi said. "It was the Internet that allowed biologists to share with each other so we actually had maps."

Today, network theory can be applied to a wide range of questions: How do we avoid geek groupthink? Who will be our future business competitors? How does the brain work? Why does Ashton Kutcher have so many Twitter followers? (I think the last question is the biggest mystery of all.)

"I could not imagine 10 years ago how big networks would become," Barabasi told me. "Which is probably a good thing."

---

Get in on the Cosmic Log groupthink by signing up as my Facebook friend or hooking up on Twitter. If you really want to be friendly, ask me about my upcoming book, "The Case for Pluto." You can pre-order it from Amazon, Barnes & Noble or Borders. Act now before I limit my connectedness!