In the desktop market, Linux”s progress is more difficult to gauge. There is sharp disagreement about how quickly open-source operating systems and productivity programs are colonizing PCs. IDC estimates that Linux holds about 3 percent of the global desktop PC market and that its share will double by 2008. Red Hat, Novell, Linspire, and others offer desktop Linux packages, and you can now buy Linux desktops and laptops from many computer retailers, including, interestingly, Wal-Mart. The Firefox browser, which runs on both Windows and Linux, already holds more than 5 percent of the world browser market. And then there is OpenOffice. In one of its quixotic attempts to snap at Microsoft”s heels, Sun decided in the late 1990s to purchase and then “open source” a small German competitor to Microsoft Office, just as Linux was starting to destroy Sun”s Unix business. OpenOffice runs on both Windows and Linux and, though presently a tiny player, is increasingly being adopted by individuals and businesses worldwide. Conversely, in the last quarter of calendar 2004, Microsoft”s revenues from Office and related software declined 3 percent relative to the year before, according to Microsoft”s publicly released financial statements.
Of course, Microsoft Office makes use of proprie-tary document formats, and OpenOffice reads them only imperfectly. (For this article, I sent some documents back and forth between the two suites; no data was lost, but formatting often suffered.) And Linux still lags Windows badly in support for the thousands of peripheral devices available for personal computers, in the number of applications that run on it, and in its ability to work with Palms and Blackberries. But for simple things, OpenOffice works, and its compatibility with Microsoft products is improving.
It is not clear that Microsoft can do anything to stop the open-source encroachment onto the desktop. Many of Microsoft”s PC products are now mature. Few users need any additional functionality, and Office exhibits very slow technical progress. Equally importantly, Microsoft has grown heavily dependent upon high prices and forced upgrades for its revenue growth and profitability. But many groups simply cannot afford Microsoft”s prices: students, poor people, educational institutions, and the majority of the developing world (see “South Africa,” April 2005). Microsoft”s products now represent a significant fraction of the total cost of a new desktop personal computer. Not only is Linux free or cheap, but because it is smaller than Windows and runs on many more devices, it can run on very inexpensive hardware.
Sensing a power shift, multinational companies and governmental bodies such as the European Union are beginning to insist that Microsoft provide open interfaces – that is, public descriptions of its software that let other programs interoperate with it. China, in particular, is determined to avoid dependence upon proprietary American software. It is concerned about trade disputes, about building its own software industry, and also about vulnerability to “back doors” that could be used for espionage. This last fear is not entirely irrational. Although there are no publicly known cases of espionage against China involving software, other technologies have been so employed. Five years ago China purchased a new, unused Boeing jet and hired U.S. contractors to refit it in Texas as China”s equivalent of Air Force One. Upon taking possession of the plane, Chinese security officers found that it harbored more than two dozen highly sophisticated, satellite-controlled listening devices, hidden everywhere from the bathrooms to the headboard of the presidential bed.
Geopolitical paranoia, however, is not the principal reason for the success of open source. The most commonly cited explanation is that evolutionary, decentralized, voluntary efforts can yield better results than those ordered by hierarchical management (see “Can Technology Raise Society’s IQ?” p. 80). But while this may be true, there is something even more fundamental at work.