Who will build our digital future?

As far as revolutions go, the opening salvo was muffled. But for those within earshot, the reverberations were far- reaching. Last month, China - the largest single potential market for almost anything - selected an upstart computer-operating system called Linux for installation on 1 million computers next year. Ultimately, the country plans to install similar systems on 100 million to 200 million machines.

But the deal represents much more than a software deal - or China's declaration of independence from software giant Microsoft. Analysts say it marks a significant victory for an emerging way of building things. Open and highly dispersed networks of motivated people are organizing around galvanizing ideas, often offering results of their work for free.

Such collaborative networks have long been part of human experience, from scientific research to terrorism. But as the approach moves into the commercial realm, especially the software business, it's challenging fundamental notions about who owns ideas and how best to foster innovation.

"Whether it's the rise of a global civil society, economic globalization, or the war against terrorism, all of these things are extremely information-dependent," says John Arquilla, professor of defense analysis at the Naval Postgraduate School in Monterey, Calif. "The software issue offers us a whole new way of looking at the world."

Known as "open source" in the software world, the concept is spreading to other arenas. At the end of September, for instance, the Massachusetts Institute of Technology in Cambridge, Mass., announced that it had reached its initial goal of posting course materials for 500 of its classes on the Web. Eventually, the school plans to post online material for virtually all its 2,100 formal courses. The material can be used freely by anyone and altered to meet local needs, as long as MIT is credited as the source for the material and no one charges for it.

Its name, "Open Courseware," is a direct nod to open-source software as its model. "The emergence of Linux was like a global barn-raising," bringing free, high-quality software to countries and institutions that otherwise might get left behind in the global information economy, says Steven Lerman, director of MIT's Center for Educational Computing Initiatives. "If we're successful, we'd like to see the same effect" in higher education.

Similar networks have been built around the human genome project and its descendants; the offering by artists of free online music; and a new research-journal project called the Public Library of Science. Even Al Qaeda has incorporated the approach to build its loosely knit network of terrorist cells.

Its most visible manifestation, however, remains software. Although the open-source approach had been around for decades, it took off in 1991, when Helsinki University student Linus Torvalds took a freely available, stripped-down version of UNIX software and modified it for a PC.

He posted the code; others began to use it, found and fixed bugs, added features, and "Linux" began to spread. Today, Linux has moved into a distant-but-solid second place behind Microsoft for software that runs network computers called "servers" in corporations, banks, and government offices worldwide. During the third quarter of 2003, the number of servers shipping with Microsoft's software grew by some 21 percent over the third quarter of '02. The number of servers shipped with Linux grew 51 percent, according to IDC, an analysis firm in Framingham, Mass.

An increasing number of countries, particularly in the developing world, are turning to Linux and open-source software. China's deal may help accelerate that, analysts say. In addition, China, Japan, and South Korea are reported to be joining forces on a new open-source software project that would focus on everything from new applications to a new operating system.

Even in the US, federal, state, and local governments are opening their IT departments to give open-source a closer look. In Washington, the Pentagon this year issued guidelines for the use of free and open-source software after a Mitre Corp. study revealed how broadly the programs were used in the Defense Department. Even some of open-source's most ardent supporters agree that the computer solutions it offers in some cases may not be as highly developed as those offered by proprietary programs. Thus users should be free to select the programs that best meet their needs.

Particularly in the developing world, Linux and other free and open-source software have economic and political attractions, says Dan Kusnetzky, vice president for systems software research at IDC.

Politically, a shift to open-source can be a digital declaration of independence in an era when the United States and its software industry are not universally trusted.

"There have been rumors for years that US intelligence agencies have persuaded all North American suppliers to put 'back doors' into software" to allow government cyberspies to keep tabs on sensitive information abroad, Mr. Kusnetzky says. "Vendors claim it's not true, but those reassurances do not stop the professional pessimists" who head foreign intelligence services. Even if a US software company offers to show its program's proprietary code to foreign officials, as Microsoft has to China, he adds, it's still hard to know if the code you saw is actually the code a company ships.

On the economic side, the software is essentially free. Installing a proprietary operating system - never mind the rest of the software it would need - on 10,000 computers could set a company or government back $15 million or more. With open-source, such as Linux, it may cost as little as the price of a download or the nominal cost of a distributor's "package," perhaps around $300.

Ironically, this on-the-cheap approach could substantially reduce software piracy in developing countries, some analysts say, because users can now get high- quality software free for the downloading.

Over the long term, he continues, countries recognize that their economic progress depends on their mastery of information technologies. Open-source software, with its global, free-wheeling, "stone soup" development approach, allows these countries to potentially shave 10 to 15 years off the time it would take them to nurture their own IT sectors at home. The reason: They have a world of expertise freely available.

Surveys indicate that the majority of programmers contributing to open-source software are professionals who donate their expertise to enhance their reputations in the community, to have the satisfaction of producing a set of code that fits immediately into an application, and to clean up others' code in a peer-review system that can be ruthless in its appraisals.

And the community is responsive. Kusnetzky cites an example in which a firm was developing a sophisticated software program. Its brightest developers found a problem they couldn't solve. The code was posted on the Web with pleas for help, and in an hour, an elegant solution came back. "The company didn't know the person, didn't know where he lived," he says.

Yet if the currency for development is information and ideas in the 21st century, where is the balancing point between the rapid innovation and responsiveness that collaborative projects can trigger and the longstanding legal structure designed to protect intellectual property rights?

That is a key question that must be addressed, says James Lewis of the Center for Strategic and International Studies in Washington. This year, Mr. Lewis joined nearly 60 economists, lawyers, and technology experts in asking the 180-nation World Intellectual Property Organization to hold a meeting next year to examine whether that balance exists. Although the idea received a warm reception in Geneva, the US Patent and Trademark Office and lobbyists knocked on the WIPO's door. The meeting was canceled.

Yet the issues aren't going away. "The fear is that tight intellectual property protections will stifle innovation. That's clearly not 100 percent true," Lewis says. "But we need to find out where the friction occurs" and if need be make adjustments.

Yet the very nature of the information age and networking could render much of the question obsolete. In an era of networks, it will become almost impossible to distinguish the parenthood of particular ideas, Mr. Arquilla says.

"You're seeing the first whispers of a gale on the horizon," adds Paul Saffo of the Institute for the Future.

Linux software runs centralized computers known as servers. Analysts project that global shipments for the operating system will more than double by 2005, pushing it to the No. 2 spot behind Windows in the server market. Here are the number of server units sold or expected to sell, by operating system:

Operating system 2002 2005

Windows 3.169 million 3.992 million

Linux 425,000 1.018 million

UNIX 484,000 387,000

Other 532,000 349,000

Total: 4.610 million 5.746 million

Source: Gartner Dataquest (September 2003)