It's sometimes called the Internet's operating system. No, I'm not talking about the Java OS from Sun Microsystems and IBM. Rather, it's Linux, the Unix clone that has suddenly -- or so it seems -- become the programmer's choice to unseat Windows as the operating system of choice on servers and desktops. To date, Linux largely has been the favourite of Internet service providers, some Web sites and a lot of academic institutions. Lately, it's sprouting around some large organisations, such as The Boeing Co., Cisco Systems and Northern Telecom. And it's gaining support from mainstream developers, most notably Informix, Oracle and the Web's first software darling, Netscape Communications. Yet, the idea that Linux could become a serious alternative to Windows still seems absurd, a dream born of desperation. How could any responsible company think about putting an operating system with no unified marketing or support organisation to work in "mission-critical" situations? After all, Apple, Novell and Sun all seem unable to stop Microsoft from dominating the desktop and, eventually, the server. How could a piece of free software, like Linux, ever hope to turn the tide? By changing the underlying economics of operating systems, in time-honoured Net fashion. How did Netscape get its Navigator product to become the dominant browser on the Web almost overnight? By allowing it to be downloaded for free. How did Microsoft take away market share? By giving away its Internet Explorer browser for free, after Netscape began to sell Navigator. Now, the tables could turn again, in the very heart of computing. It's hard to imagine a collaborative effort of professional developers succeeding at giving away an operating system, but that doesn't mean it can't be done. First, the Netscape and Microsoft browser instances show that quality software can be given away -- if there is an ulterior motive for doing so. In Netscape's case, it was to lay the foundation for a larger Net software business that could be built on having its "net-top OS" on every screen. For Microsoft, it was to undo Netscape -- and retrofit its OS for the Net era. The developers who are moulding and enhancing Linux do so not because they want to make money on the operating system. Instead, they want to make money selling information, digital transport and hosting, advice, merchandise or something else on the Net. Or they want to model the atmosphere. Or what have you. Regardless, they want a reliable operating system at the lowest possible cost. Second, usage of Linux will prove whether or not it is indeed more reliable and flexible than Windows -- and Windows NT in particular, whenever the fifth version finally comes out. Because of Microsoft, many people just assume computers crash all the time. But that doesn't actually have to be so. As Linux' namesake and 1998 Net poster boy Linus Torvalds notes: "Computers are really reliable things that do everything you want them to do and nothing else." Finally, there's more to the "free" aspect of Linux than just price. What will also drive usage among information systems (IS) managers is that the code itself is free. If something needs to be fixed, a decent programmer can get inside it, fix it and share the results. With Windows, the engine comes without manuals and the hood is welded shut. Therein lies the back door through which Linux can get to corporate IS departments. What gets Linux in the door may be financial exigency -- there's never enough money to start up a reliable Web site, for instance. But what will keep it around is superior performance, which can be proven by tracking how often Linux servers have to be rebooted vs. Windows NT servers. And if Windows NT proves to be a train wreck, as its continuing delays suggest is possible, then Linux' rise to prominence could prove exquisitely timed: Get Linux onto servers; wait for a decent interface and better office applications to develop; then, move onto the desktop. All the while, keep emphasising that support will come from the thousands of Linux developers -- all who understand its inner workings and who don't work for a single monolithic software company. That's a model that indeed could make life uneasy for Microsoft, which can't ever afford to open Windows to outside developers to tweak or fix. If it ever tried to give Windows away for free, it would have a hard time explaining to an antitrust judge why it used to charge for it. Lest you think that such "open" computing can't possibly win, just look back at the primal lesson of desktop computing of the '80s: Open up your architecture to all comers and win -- or keep it closed, like the Macintosh, and lose.