Printable page E-mail this

THE words must have shocked the global community of volunteer software developers known as the open-source movement. Volker Wiegand, president of the American subsidiary of SuSE, a leading Linux vendor, recently declared on the website Linuxgram that Linux as a business wasn’t working out. His company had laid off two-thirds of its staff in early February. The free operating system was a “fallen angel” and a victim of irrational expectations, he said.

The statement, in essence, repeated what a Microsoft executive had argued a few days earlier. “There isn’t much value in free,” according to Doug Miller, group product manager for competitive strategies. He predicted that many Linux businesses would falter before the end of 2001, and that this, together with technical shortcomings, would “call into question whether Linux should be used at all” (instead of his own company’s ubiquitous Windows, that is).

The Linux hype has undeniably crested. SuSE is not the only company that has had trouble building a viable business based on the free operating system. VA Linux, for example, a start-up which, in late 1999, enjoyed the most successful flotation ever—with shares gaining almost 700% on their first day of trading—recently said it would cut a quarter of its staff and take nine months longer than planned to achieve profitability. Only the market leader, Red Hat, is doing better than expected.

But this should not be read as a sign of the imminent demise of open-source software via Linux, its standard-bearer. Most people in the software industry believe that open-source is here to stay. Steve Ballmer, Microsoft’s chief executive, recently called Linux “threat number one”. Steven Milunovich, a leading analyst with Merrill Lynch, an investment bank, argues that open-source is a “disruptive technology” that could topple such industry heavyweights as Microsoft and Sun.

In fact, the open-source movement is less about “world domination”, which hackers often joke about, and more about an industry which, thanks to the Internet, is learning that there is value in deep co-operation as well as in hard competition. “Much more than a cause, the open-source movement is an effect of the Internet,” says Tim O’Reilly, head of an eponymous firm that publishes computer books, and a leading open-source thinker.

Open-source is often described as the software industry come full circle. Indeed, in the early days of computing, programs came bundled with the hardware and complete with the source code (the set of computer instructions which are then translated into binary code, the form of software that computers can understand and act on). Pioneers needed to tweak their programs, and were happy to share the improvements they made.

It was only in the 1970s, as computing spread, that firms such as Microsoft started to withhold the source code, thereby making software proprietary and turning it into a big business. Firms can sell a program without revealing the instructions that underlie it, just as Coca-Cola can market its soft drinks without giving away its secret recipe (though there have been plenty of attempts at reverse engineering, in software as well as soft drinks).

Many early hackers were horrified by the decision to withhold the source code. Proprietary software was “spiritually wasteful”, they said, because it discouraged co-operation. One of them, Richard Stallman, founded the Free Software Foundation in 1983 and developed the concept of “copyleft” (as opposed to copyright), which he codified in a licence that now comes with most open-source software. It states that developers can do whatever they want with the programs, even sell their own versions, as long as they make the source code available.

Although this licence, called the General Public License (GPL), has never been enforced, it has done much to keep open-source software from splintering into competing commercial versions, open-source advocates say. The GPL, in effect, removes the incentive to turn a program into a proprietary product, they argue, because the licence is “viral”: all changes to the source code automatically become part of the software commons. That is why James Allchin, who is in charge of operating systems at Microsoft, recently called the licence an “intellectual-property destroyer”.

But open-source projects are best understood as a new generation of standards bodies, much like the Internet Engineering Task Force (IETF). The traditional methods for setting technical standards have turned out to be rather inefficient. Governments and industry committees are slow and often get it wrong. Monopolies are faster and do a better job, but they tend to put the brake on innovation in order to keep their dominant position.

The Internet has opened up a third way. It allows engineers in different corners of the world to collaborate at almost zero cost. Since communication is so easy, there is no need for official leaders or a big bureaucracy to keep things running. And the decision-making process is more transparent: discussions and documents, for instance, are easy to get at and to search. What counts is the quality of the argument, not the power of special interests.

The Internet pioneers were the first to create online communities for their work. They modelled their groups on organisations they had grown up in: universities and engineer corps. That is the main reason why the IETF turned out to be an interesting cross between a scientific community and a guild. It is run by elders, and its main mantra is: “We reject kings, presidents and voting. We believe in rough consensus and running code.”

Most open-source projects are organised in much the same way. Their members are motivated mainly by fame rather than fortune; it is considered a coup to write a “patch” that passes the peer review of fellow developers and gets incorporated in the next release. And most open-source projects are governed by a “benevolent dictator”, an individual with exceptional programming, organisational and communications skills—such as Linus Torvalds, the creator of Linux.

The power of these leaders, however, is not absolute. If developers are unhappy with them, they can always “fork”—take the source code and start a new development branch. To Brian Behlendorf, one of the leading developers of the Apache web server, another successful open-source project, and now chief technology officer at CollabNet, this keeps the benevolent dictators from becoming real ones.

Yet cheap communications alone are not usually enough to get open-source communities off the ground, nor to make them as successful as Apache or Linux. They need a starting point and a framework. For Apache, these were provided by a program developed by the National Centre for Supercomputing Applications (NCSA) and by the open standards of the Internet. Linux had the original, very simple version written by Mr Torvalds as a starting point and Unix, a programme created in the 1970s, as a blueprint.

To be successful, open-source software must also be designed in a modular way so that groups of programmers can work independently on different components. Most of the work on Linux, for example, is done not by a large horde of hackers all working on the same code, but by small groups of a dozen or so developers each of which concentrates on one small part of the program.

Lastly, an open-source alternative needs a market. Linux had a ready-made one: hardware makers wanted a free alternative to Windows so as to increase their margins and gain some independence from Microsoft. The recent antitrust case against the software giant gave them a chance to support another operating system without fear of immediate retaliation. From Compaq to Dell, from Hewlett-Packard to IBM, they have all jumped on the Linux bandwagon, offering computers with the program pre-installed and investing in Linux companies.

Ironically, it is IBM, the hackers’ original arch-enemy, that has gambled the most on Linux so far. Big Blue recently announced that it will spend $1 billion on Linux this year because it really wants the program to become a computing standard. It is doing this because it sees itself mainly as a provider of e-business solutions, and because a standard operating system would make it much easier to integrate them, explains Irving Wladawsky-Berger, IBM’s man in charge of its Linux operations.

Hence IBM’s efforts to make all its important programs run on Linux and transfer the operating system itself to all kinds of computers, from tiny hand-held ones to mainframes. Using special “partitioning” software, these workhorses of computing can be turned into a virtual collection of thousands of Linux servers. IBM is also trying to help the Linux community beef up the program for heavy-duty corporate computing, where it is still a rarity.

Some people like to dismiss Linux as nothing more than a happy accident, but the program looks more like a textbook example of an emerging pattern. If a piece of software is well understood, to the point of becoming a commodity; if it can be built in a modular way; if there is a critical mass of users who are also software developers; if there are talented project managers among them; and if there is a demand for a program that is not controlled by a single vendor—then the chances are it will be suitable for an open-source project.

Admittedly, that is a lot of ifs. But Linux is by no means the only successful open-source project. Take Sendmail, which is used to route two-thirds of the world’s e-mail and has spawned a start-up of the same name. It makes a living by selling a commercial product, but it also helps others to develop a free version of the program. The community of developers creates a basic product platform that Sendmail can polish and enhance to make money.

And there are many smaller, less well-known open-source projects that could gain momentum. In October, Sun Microsystems launched OpenOffice, a free alternative to Microsoft’s Office suite. Forrester Research, a high-tech consultancy, forecasts that by 2004 companies will be spending 20% less on software licences than they do now because they will be using these kinds of open-source programs.

However, there are two areas of the software industry that open-source will have trouble penetrating. One is enterprise software that relates to a company’s core activities. Companies will hesitate to bet their business on free software, at least until it is backed by heavyweights such as IBM, as Linux is now. The operating system will be the test case of whether open-source software can move into heavy-duty corporate computing. Linux’s most recent release is a big improvement, but it still has some way to go.

The other difficult territory is truly innovative programs. This kind of software seems to be better designed and built by a tightly knit group of developers within a single company. Open-source is extremely good at optimising existing programs, says Ray Ozzie, chief executive of Groove Networks and creator of the popular Lotus Notes. But he cannot imagine his firm’s software, which allows small groups of people to collaborate online, being developed in a decentralised way. He and four colleagues operated “like one brain” for three years to solve all the engineering problems, he says.

Still, nobody can predict what kind of “mob software” may turn up. Richard Gabriel, a distinguished engineer at Sun, describes this as “a kind of semi-chaotic, self-organising behaviour in which numerous small acts of repair can lead to quickly built, complex and massive creations”. Open-source is certainly a mass phenomenon, with tens of thousands of volunteer programmers across the world already taking part, and more joining in all the time, particularly in countries such as China and India. SourceForge, a website for developers, now hosts more than 18,000 open-source projects that keep 145,000 programmers busy.

Meanwhile, the corporate world is warming up to open-source in unexpected ways. Firms are showing increasing interest in using this development methodology for software they create in-house or with a limited number of partners. For example, Dresdner Kleinwort Wasserstein, an investment bank, in January released the source code of an application it developed to link different types of computer systems, and has hired CollabNet to build a developer community around it.

Even Microsoft is learning to love open source. The firm recently announced that it will share the source code of the latest versions of Windows with 1,000 of its best customers—but only on condition that they do not modify the program.

The official software industry, for its part, is already busy building the infrastructure for the next wave of computing: web services.