f 1998 and 1999 were the worst years in Microsoft's history, 1991, by contrast, must have been a period when Bill Gates was feeling good. Windows 3.0, launched in May 1990, was a growing success; in the first year alone, 4 million copies had been shipped, a huge number for the time. In May 1991, Microsoft launched Visual Basic, a radically new way of programming that employs visual design methods rather than traditional ones based on editing text files.
Even better, Windows 3.1 was close. Despite the point upgrade, version 3.1 represented a major advance over 3.0 in almost every way. Microsoft claimed that it contained over 1,000 enhancements. Its cool new user interface seduced almost everyone that saw it. When Windows 3.1 shipped in June 1992, it cemented Microsoft's dominance on the desktop. It also created a discontinuity in the software world as companies switched from DOS-based programs to those running under Windows. Microsoft was able to exploit this crossover to wrest leadership in the spreadsheet and word-processor sectors by quickly launching programs such as Excel and Word. Windows 3.1 was not the only major operating system nearing completion in 1991.
Microsoft Windows New Technology, better known as Windows NT, had been started back in 1988 in a drive to create an enterprise-level operating system that would be as widely deployed in companies' back offices as MS-DOS and then Windows were in the front offices. The Windows NT project was headed by Dave Cutler, who had built the operating system VMS for the computer giant Digital, formerly known as DEC. VMS was a rival to Unix, another robust operating system, but was an official corporate product, unlike Unix, which had always been regarded as software for hackers. Windows NT, too, was unashamedly meant as a Unix-killer when, against expectations, Unix-not VMS-emerged as the leading enterprise operating system. NT's chances looked good. By the late 1980s, Unix was highly fragmented, each vendor offering a slightly different version; this meant that application software had to be rewritten many times, and that users were locked into the software of one supplier. Because Unix had also failed to embrace fully the new wave of graphical interfaces, its solutions in this area were crude when compared with the Apple Macintosh or Microsoft Windows. Windows NT, by contrast, was designed to marry the power of a VMS-like operating system with the elegance and usability of Windows 3.1. But plenty was happening outside Microsoft's immediate sphere of interest.
In 1991, Tim Berners-Lee, a British physicist at CERN, the European Centre for Nuclear Research, released for public use a hypertext system that he had been developing over the past two years. The system, which he called the World Wide Web, ran across the Internet, then still small-scale and used mostly by academics. Also in 1991, a new programming language called Java was being drawn up by a team at Sun Microsystems. Born originally as part of an attempt to develop an interactive set-top box for the cable TV industry, Java was later adapted for use on the Internet. An important feature of Java was portability: The same program could run without modification on a wide range of hardware, a novelty at the time. Although the threats these developments represented at the time were negligible compared to their later impact, it is possible that Microsoft was tracking them in 1991. The Internet was well known enough, the Web was in the public domain, and Sun was a competitor whose moves were certainly watched with interest.
But surely it is impossible that, in the same year, Microsoft could have had even the merest whisper of a suspicion that a key element of an equally potent challenge was about to take shape in the bedroom of a second-year computer science student in Helsinki, the capital of Finland. As one of the most northerly capitals in the world, Helsinki is a city of seasonal extremes: dark, cold winters with only a few hours of daylight, and bright summers where the days never seem to end. Geographically, Helinski is close to Russia's St Petersburg, and parts of it are visually similar; culturally, it has closest ties with Sweden, of which for centuries Finland formed a province. When Russia invaded and annexed Finland as part of its empire in 1809, it separated from Sweden. Finland only won full independence in 1917. Two cathedrals form important landmarks on the eastern side of the city center. The Lutheran church possesses an elegant neo-classical exterior, but takes the form within of a huge stone shell almost devoid of ornament; the Russian Orthodox Uspensky Cathedral is a typical concoction of onion domed towers outside, and icons inside. Helsinki is compact; its low buildings and broad streets are mostly laid out on two grid systems that abut at a slight angle. Green spaces abound, and the sight and smell of the sea, which surrounds Helsinki on three sides, is never far away.
Into this ordered but individual world, Linus Benedict Torvalds was born on 28 December 1969. Linus explains the significance of this day in his culture: "It's special in the same way most days are special-many days have some traditional meaning. December 28th is 'menlosa barns dag,' which means roughly 'day of children without defects' although 'menlos' is fairly old Swedish and a lot more poetic than 'without defect.'" The name Linus was an unusual choice: "not unheard of,"Linus explains,"but not exactly common" in Finland either. The name itself has a history that stretches back to the roots of Western civilization. It is mentioned in Homer's Iliad in the original Greek form Linos, where it is associated with a song of mourning. There is also a St. Linus, who is traditionally listed as the second pope after St. Peter. Another famous Linus-though better-known for his last name-is the American inventor Linus Yale.
But rather than any of these, Linus was named in part after the U.S. scientist Linus Pauling, who received two Nobel Prizes: one for chemistry, and one for peace. "I think I was named equally for Linus the peanut-cartoon character," Linus adds, and notes that this makes him "a mix of 50 percent Nobel-prize-winning chemist and blanket-carrying cartoon character." Torvalds is a rare surname: "There are probably something like twenty-five people in the world with that last name," Linus says. "Torvald" is "a rather old-fashioned Nordic name," Linus explains. "You'll find it in Sweden and Norway. The genitive 's' is unusual, an old way of turning a regular name into the name of a farmstead-and from there into a surname." Linus explains that it was his paternal grandfather "who took the name for personal reasons, namely, that he didn't like his family." As a result, he says, "to my knowledge, all Torvalds in the whole world are [his] descendants." In other words, one of the most famous names in computing is "completely made up, and not more than two generations ago." The Torvalds clan is notable not just for its unusually small size: A surprising number of them are journalists. Linus runs down the list: "My dad, Nils Torvalds: reporter for Finnish radio. His brother, my uncle, Jan Torvalds: Finnish TV. My paternal grandfather-the founder of the Torvalds clan, deceased-used to be a newspaper reporter, writer, and poet. My mom: Mikke (Anna) Torvalds: working at the Finnish News Agency. Used to do translations, does news graphics these days. My sister, Sara Torvalds: used to do translations for the Finnish News Agency too, but is moving more towards books and films." The Torvalds family formed part of the Swedish-speaking community in Finland, about 300,000 strong in a total population of 5 million. That their mother tongue has nothing in common linguistically with the Finnish language that surrounds them has doubtless helped them become a very close-knit group. Reflecting this, Swedish speakers themselves call this society within a society "Ankdammen"-the Duck Pond.
Find your favourite book here
One of Linus's friends from Helsinki, future fellow hacker Lars Wirzenius, says, "Almost all Swedish-speaking people know lots of other Swedish-speaking people, who know lots of other Swedish-speaking people, and the end result is that everyone either knows everyone, or knows someone who knows someone." Growing up as a member of the Duck Pond, Linus spoke Swedish at home and with family friends, and only started learning Finnish when he was five. Contact with English, which would become a crucially important factor in his work, came a few years later. It was around the same time that he first encountered a computer. Linus explains that his maternal grandfather was a statistician at Helsinki University; he bought a Commodore Vic-20 microcomputer, "one of the first ones to get on the market, at least in Finland." He adds, "It wasn't exactly what you'd call a number-cruncher today, but it was certainly faster than any calculator." The speed of the processor was just 1 Megahertz (MHz), a thousandth of that of modern PCs. Linus recalls that his grandfather "bought the Vic-20 for his own needs to do math," but soon asked his young grandson to help. "I think my grandfather wanted me to learn, so he made me help him," he explains. "So I started off helping him with the computer, and then I did my own stuff on the side." Linus recalls, "I had the Vic for five years because I couldn't afford to upgrade. I programmed in Basic, maybe the first two years." But Linus soon moved on from this popular beginner's programming language to something more demanding: assembly language. Assembly language commands are easier for the computer to act on but harder for programmers to think in. "I kind of got used to doing stuff myself by reading books about assembly language," he says. "I didn't know about assemblers," programs that make writing assembly language code easier, "so I had to do it by hand; after a few years I essentially had to upgrade because everybody else had better machines." Linus goes on to add another reason for buying a new computer: "I was getting to know the Vic too well." Throughout his computing life, Linus would search for new programming challenges, but it is interesting to see this trait emerge so early. His next machine, and the reasons for that choice, are also highly characteristic of the later Linus. "I was looking at different machines-I didn't want a PC because I really disliked the Z80 [chip] architecture and in the PC, the chip was essentially the same," he says. That is, he decided not to buy a PC because he disliked the design of the Intel chip family that lay at its heart-an unusual way of looking at things. "Back then I was doing just assembly language, and I didn't want to have anything to do with that [particular processor]." Because Linus was writing "low-level" code, which interacted directly with the chip, he was more conscious of the merits and demerits of the various chip families. Most programmers write in "high-level" languages like Basic, which effectively shield them from the details of the hardware.
As Linus himself says, he has always been a "low-level" person. There are probably two reasons for his early interest in this phpect. One was his emerging love of programming at the most fundamental level. The other is more pragmatic: "I had been a performance junkie since forever. When you had to write games for a slow 1 MHz [processor]," Linus explains, "you have to be kind of crazy and tweak cycles." "Tweaking cycles"-getting every last drop of performance out of code-would later mean that Linux was far faster and leaner than comparable programs. In the end, Linus chose an unusual micro for his next machine, the Sinclair QL ("Quantum Leap"). This was a typically quirky product from the British inventor Sir Clive Sinclair. Linus had been content to put up with Sinclair's QL-although it had some fairly obvious shortcomings-for a simple reason. The main thing he wanted, he explained, was "a machine at home that does multitasking." Even though the Sinclair QL was in many ways a toy, it had one very powerful feature, thanks to its choice of chip: it could run several programs simultaneously, just like commercial minicomputers. This area of computing-multitasking-led him to start coding the simple program that would eventually turn into Linux. But this was still some years off from the time when Linus was hacking on his Sinclair QL. First, in the fall of 1988, he entered Helsinki University to study computer science, which by now already looked likely to turn from a passion into a profession. At university, Linus found the same tendency for the two language communities in Finland to keep to themselves. His fellow student in computer science, Lars Wirzenius, comments that "at that time there weren't very many Swedish-speaking computer science students, and those that were, were at least a couple of years older than we were." It was therefore only natural that a pair of Swedish speakers isolated among the Finnish majority of the new intake for computer studies should gravitate towards each other. Wirzenius recalls that the first time he met Linus, "it was one of the introductory lectures for new students.," Wirzenius didn't notice much of his friend that day except that "the end of his big nose tends to twitch up and down when he speaks, and it is fun to look at." Aside from the nose, which, in truth, is not that big, little in Linus's appearance is out of the ordinary: He is of medium height, with brown hair; his blue eyes gaze steadily from behind his glasses. Only the eyebrows, which are remarkably dark and bushy, jar slightly with the otherwise boyish face. In an effort to meet more student members of the Duck Pond, Wirzenius and Linus joined one of the many societies that form an important element of Helsinki University's student scene."
The club that Linus and I joined was called the Spektrum," Wirzenius recalls, "which is the Swedish-speaking club for people who study math, computer science, physics, or chemistry." As far as the teaching at the university was concerned, Wirzenius explains that "there isn't actually a curriculum which says that during the first year you have to take this course and that course. The so-called academic freedom in Finland is interpreted so that students can take any course they want, mostly in whichever order they want. "The nominal study length is five years, but most people study for six or seven years," he continues. Alongside the main courses there were various ancillary subjects. "To get a computer science degree, you have to study a certain amount of math, and also some other subjects. Each week we got some exercises to do, and then we had small groups of people and a teacher who asked each one of us to demonstrate some of the problems we were supposed to have solved. And if you hadn't solved all the problems, that was OK, but you had to solve a certain amount to pass the course. "One of these weeks, Linus-for some reason, I don't remember-hadn't done all his homework. So he just claimed that he had done one of the exercises that he hadn't done-and the teacher asked him to demonstrate his solution. [Linus] walked up to the blackboard... and faced the problem that he claimed he had solved. Linus decides this is a simple problem, draws a couple of diagrams, and waves his hand a lot. It takes a long time for the teacher to understand that yes, this is actually a correct solution." According to Wirzenius, this incident was atypical in one respect: "[Linus] didn't usually try to cheat, because he didn't have to. He knew math well from high school, and had a sort of mathematical brain, so he didn't have to spend much time on homework to get it done." Wirzenius believes there is something characteristic of his friend in the way Linus handled the situation-"the attitude, the arrogance that he displayed-most people would just have acknowledged that they didn't actually have a solution," but Linus hates having to admit that he doesn't know the answer. Wirzenius says that "the arrogance he showed then is still visible" in how Linus handles challenges within the Linux community. "These days, he might claim that a certain kind of approach to a problem is the correct one, even though he hasn't necessarily thought about all the other approaches or solutions, and then the rest of the Internet either agrees with him or tries to convince him otherwise." This approach works because Linus is "careful enough to not do it all the time. Most of the time when he says something, he has been thinking about it," Wirzenius notes, and concludes: "I wouldn't like to play poker against him. He can bluff." Just as important, if after all the bluffing Linus is caught out and shown to be wrong, Wirzenius says, he will accept corrections with good grace, a trait that would prove crucial in managing the community of equally opinionated and able hackers that would later coalesce around him. After the first year's introductory course, there followed a major hiatus in their university studies. "All Finnish men are required to do either military or civil service," Wirzenius says. "Civil service is sort of you get an ordinary job but you don't get paid for it." In 1989, when he and Linus were required to choose, "the shortest time to do military service was eight months [and] all civil service was sixteen months. So I figured that's eight months too much, so let's do the shortest possible. Actually, the totally shortest possible way would be to refuse to do either kind of service, and then it would be six months in jail." Like Wirzenius, Linus, too, decided that he'd rather not spend half a year in prison. But instead of choosing the next shortest option-eight months as a simple army private-he opted for one that lasted eleven months, which was training to become a noncommissioned officer. Of this training, Wirzenius says, "It's sort of useful as an exercise in leadership, but a certain kind of leadership. . . what the army needs is group leaders who can teach their group to act almost on reflex, which was sort of scary. It's the same kind of stuff that all armies teach." He adds, "It's not just a way to learn how to make reflexes, but also a way to keep a group together as a working unit even if they don't like each other, and stuff like that." It is hard to imagine a better description of what would be involved in coordinating the global movement of volunteers that develop Linux. Wirzenius recalls that he and Linus didn't see much of each other during that time even though they were both stationed in the same area of east Finland towards the end of their service; however, they made up for lost time when they were back at the university and had resumed their exploration of the computer world. They began with a course that would have an effect not only on them-Wirzenius said that it felt "sort of like falling in love for the first time"-but on the entire history of computing. In the fall of 1990, they started using the university's recently acquired MicroVAX system, designed by none other than Dave Cutler, at that moment busily at work on Windows NT. Wirzenius and Linus were about to discover Unix.
© 2001 Glyn Moody
Email this article
to this article