Linus Torvalds
Translation: Axis (Xia Hao)*
1. Linus and Bill
Almost overnight, this name suddenly became as widely known as Bill Gates. But Bill Gates, no matter how many countless millions of dollars he may be worth, will never become Linus Torvalds. This 29-year-old Finn, the creator of the simple yet powerful Linux operating system, has gone beyond the Gates myth. There are also rumors that Gates is an outstanding programmer, but Linus is the real thing. While still in college he completed a genuine operating system. The young Gates called fellow programmers who copied his own mediocre programs “thieves,” while the generous Linus shared his masterpiece with the whole world. Which man, which movement wins, may decide the future of technological development.
Two years ago, if anyone had said that the technology giant and close Microsoft partner HP would be busy making sure the computers it produced could run Linux smoothly, it would have sounded like fantasy. But the development model Linus adopted, inviting the whole world to share and improve his personal original work, sparked a revolution. Skeptics felt that Microsoft, faced with the U.S. Department of Justice lawsuit, was only labeling Linux a serious competitor in order to manufacture an imaginary enemy and wash away the charge of monopoly. But the truth Microsoft let slip may have been more than they intended. At the end of last year, an internal Microsoft white paper called the “Halloween Documents” leaked to the press, and its conclusions might well keep Gates awake at night.
Linus is regarded by many as the best representative of open source software (open source software), and Microsoft’s traditional way of dealing with competitors probably has no effect on him. “Linux and other open source software are increasingly proving credibly that OSS is at least as robust as commercial software, if not beyond it.” So wrote Microsoft engineer Vinod Valloppillil in the white paper: “The ability of OSS to gather and exploit the collective intelligence of thousands on the Internet is simply amazing.”
But what the “Halloween Documents” did not mention is that Linus has already proved himself the smartest and most powerful of the many anti-Gates figures. Oracle CEO Larry Ellison’s NC offensive was only a passing cloud; Sun CEO Scott McNealy, despite his great success with JAVA, seems to have become bogged down in unconditional opposition to Gates; by contrast, Linus has never strayed from his own path. Though he himself would never admit it, you could say he is a smooth and seasoned politician. He avoided the confrontation between pure free-software supporters and profit-seeking new vendors that could have shattered Linux into pieces. He welcomes all commercial Linux versions and commercial Linux software that can improve the software’s reliability and usability, and still spares no effort in supporting the free software radicals who helped open the way with him.
Hard to believe, in this age of IPOs and nonstop high-tech billionaires, Linus does not seem eager to cash out his success. He still drives the Pontiac Grand Am he bought in early 1997 when he left Finland and started his first job in America (also his only job so far—this is Silicon Valley, after all). A year ago, when that little apartment in Santa Clara could no longer hold Linus, his kindergarten-teacher wife, and their two daughters, they rented the first house they looked at—right across the street. All kinds of rumors about his employer, the highly secretive Transmeta, are making a huge stir in Silicon Valley, but so far Linus shows no sign of suddenly making a fortune.
Face to face, the man most people simply call Linus wears glasses, speaks softly, is approachable, proud of his talent and achievements, and very humorous. He used to be famous for eating whatever was in front of him, but now he has a special fondness for sushi. Because of all the programming, his belly sticks out a bit, but neither his voice nor his mischievous smile has lost that Finnish lightness.
Linus seems to be a populist, hoping to break the monopolistic way of thinking about software pricing, and this is reflected in his other tastes as well. For example, he doesn’t buy hardcover books, because he believes hardcover books don’t sell all that well, so they must be overpriced. Likewise, he prefers entertaining Hollywood movies to highbrow European art films. His father works in radio, his mother translates newspaper news, an uncle works for Finnish television, and his grandfather is a newspaper reporter, so he understands the news media, and seems quite pleased with his growing fame.
2. Rebellious, rebellious
Question: Do you consider yourself a traditional hacker in the positive sense, a rebel? I feel simply calling you “a programmer obsessed with programming” is not enough.
Answer: I generally try to avoid using the word hacker. In private conversations with other technical people, I do call myself a hacker. But when dealing with the media I just say “programmer” or something like that, because now hacker often means something else.
Question: Then what about the label “rebel”?
Answer: “Rebel” implies you are fighting for some cause and against something else. I wouldn’t say I’m a rebel, I’d say I’m principled. (ethical instead of rebel) The basic meaning of rebellion is opposition, while I feel I have my own goals—that’s the difference. I have my own principles. I don’t care whether others believe in them too; I think that is entirely a personal choice.
Question: Then how would you describe your principles and goals?
Answer: My basic principle for life is: “Do unto others as you would not have done unto yourself.” This is not just a Christian idea, it is universal. It’s a very simple rule, and in the vast majority of cases it tells you what to do. If you’re hesitating—“What should I do?”—you can ask yourself, “What would I want him or her to do?” Then you know the answer immediately.
Question: Do you believe there is some social or political element in your work or in your technology?
Answer: Not from the beginning. At first it had nothing to do with society; the original motivation was technical and personal interest. There was no politics in it because politics, by nature, is about changing people’s minds, and I’ve always disliked having that happen to me.
But many other “open source” people may not think the same way I do. Many of them are political: some people try to make others accept their views, and I don’t want to do that. I’ll talk about my views too, but only after someone asks me.
Question: There were many chances for you to go into business, or to decide that the Linux operating system should be commercialized, but clearly you didn’t do that, and that surprises me. Why?
Answer: It’s a matter of personal interest. What attracted me to Linux has always been the technology. It wasn’t started to make money, and even when it became possible to make money, that still wasn’t my original intention. Commercialization would change my motivation.
I’m very happy that some people make money from Linux. That adds depth to Linux and brings in new motivations and new factors, things that would be impossible without commercial goals. But when I started Linux, commercialization was not my choice.
Question: Many people would think that was a difficult choice. But you seem perfectly at ease with it. Why do you think that is?
Answer: Maybe in America it’s a difficult choice, but not necessarily in many other parts of the world. America is money-centered, everyone values money, and of course that has its advantages.
But Finland is different. In that cultural setting, success is admired, but other things are valued too, maybe even more. For example, my family places great importance on degrees. They think learning is important; they don’t care that much about money.
3. Linus 2.0
Question: There’s a heavy cloud of mystery around your work at Transmeta. Can you give me any hint about when the truth might come out?
Answer: No. Maybe tomorrow, maybe in 15 years.
Question: Did Transmeta come looking for you, or did you go looking for Transmeta?
Answer: It was a two-way choice. I knew a Swede who worked at Transmeta. He spent a day in Helsinki and met with me, and at that time I said, “At last I can graduate,” because by then I hadn’t had a real break in years. Then he asked whether I wanted to go to Transmeta, and then I talked with his boss. A week later, in the fall of 1996, I was on a plane to California to take a look.
Question: I understand you can’t reveal work-related things, but can you tell me what excites you about this job?
Answer: Partly because by then I had already spent six years on Linux and wanted something else too. I didn’t want to hang myself on a single rope. I didn’t want Linux to be my whole life; I wanted to find something motivating in itself. Also, small companies feel more human. And it’s exciting to get involved in areas where nobody else in the world is involved.
Question: Those things have nothing to do with Linux, right?
Answer: Internally they use Linux. I use it every day. They also welcome the work I do on Linux. Partly because of the PR effect, partly because they can use it internally too.
Question: But your work on Linux is unpaid. Many people wonder how you can spend so much time on something that brings in no money...
Answer: Even for those who can’t imagine that anyone in the world might do things out of interest—and that’s kind of sad, but such people do exist—I can explain it this way: if Linux really goes smoothly for 10 years, then I’ll be able to get whatever I want, like putting money in the bank. That’s not my purpose, but you can explain it that way to people who otherwise can’t understand.
I think of myself as an artist, doing what I want to do. And it’s not as if I have to live miserably, because programmers are treated pretty well.
4. Time is everything
Question: Your views on capitalism are interesting. It seems you also believe in some of the ideas of free software, but you’re not against charging for software and services. In what cases is charging more reasonable, and in what cases is free more reasonable?
Answer: Charging is always reasonable. But when I started, I couldn’t use commercial software, because it was too expensive. One reason Linux exists is that I didn’t want anyone else to end up in that situation again—that’s “Do unto others as you would not have done unto yourself.” But at the same time I don’t think taking money is wrong.
I’m not against money. Money is an interesting concept, but you have to choose what matters most. The good thing about Linux is that if you’re a company willing to pay for 365 days a year, 24 hours a day service, then you’ll have to pay a lot of money—even on Linux that’s expensive—but if you’re a poor student, or in a third-world country, maybe Linux will give you a chance to encounter some new things.
Question: Did the speed of Linux’s development or the direction it took ever surprise you?
Answer: What happened last year had been brewing for some time. In a way, when one company decides to support Linux, it is surprising that so many others then respond. But I’ve never sat at home grinning ear to ear, unable to say anything but “wow”; my reaction is more like, “Hmm, pretty cool, it really happened.”
Question: Can it be said that Linux caught the timing well?
Answer: Linux’s success is a combination of timing, demand, and market opportunity. For example, when Linux was first being developed, people didn’t have CDROMs. A year after Linux appeared, CDROMs were everywhere, so suddenly there was a possibility of commercial Linux distributions with high performance-price value. So yes, the timing was very good. The Internet was also a great opportunity. Linux already existed before the Internet became popular, but once the network infrastructure was in place, Linux definitely benefited from it to one degree or another.
Question: I also feel that, judging from the way things developed on Microsoft’s side, Linux appeared at just the right time.
Answer: For PR reasons, the Justice Department lawsuit became a big anti-Microsoft gathering. Two years ago, nobody questioned Microsoft; everyone thought what Microsoft did was only natural. People worshipped Microsoft: big company, successful, making loads of money, and people thought that was the American dream. One result of the Justice Department lawsuit is that some people who used to like Microsoft have now changed their minds.
5. Open source is democracy
Question: And now they also know there are choices other than Microsoft?
Answer: The fact you mention means that suddenly you begin to notice there are other choices. Before, people weren’t looking for alternatives. Now, even if you don’t switch from Windows to Linux, or make any other big change, you still say: “Looks like what Microsoft says needs to be thought about.” I think that is important.
6. Open vs. closed
Question: How would you compare your open source way of developing Linux with Microsoft’s closed way of developing Windows and Windows NT?
Answer: With Linux, users of the system can influence the direction of development. From the standpoint that users do not have to give up control, that is democracy. Everyone can do anything. Of course, the basic premise is that you must have enough ability, but that’s also a good way to distinguish those who do the work from those who don’t. And even those who don’t modify the software can still make suggestions, do testing, and so on.
I think open source is quite superior as a development model—especially compared with the traditional commercial closed development environment. The essence of the problem is how to motivate people to create.
But there is indeed a lot of work that simply can’t be called interesting. If a job is not interesting, there have to be other incentives. The most obvious one is money. To some extent, it is easier to make money in a closed environment. The closed model relies on not providing others with complete information, thus limiting competition. Limiting competition is beneficial to companies.
Question: Do you have any metaphor to describe the difference between the two models?
Answer: I’d put it this way: the closed model is a publishing house that controls movable type rather than language. Users have the freedom to read and the freedom to write, but the publisher decides what kind of things can be published, distributed, and passed down. You can read, and you can handwrite, but if you want to improve a book, you either have to start over from scratch or ask the publisher.
Question: Then the metaphor for the open source model should be a printer—the printer restricts nothing.
Answer: Right, you can have as many printers as you like, and printers can share fonts.
Question: That’s an interesting metaphor. Why have Linux and the open source model been so successful? Can you tell me what you think?
Answer: One reason Linux has come this far is that it is the weaker side. To some extent, the forces of the market don’t mean that much to Linux developers. People develop Linux because the things they are interested in don’t get enough market share, and because they don’t have some board of directors behind them requiring a financial explanation every quarter.
You change the rules of the game, and in doing so you also change the market. (You also change the market by changing the rules in ways that Microsoft isn't willing to fail on) Open source availability became a selling point. Suddenly there was a new rule of the game. Not everyone cares about that rule, but as long as some people do, they will reject Microsoft without a second thought. It’s just like two armies facing off: you can’t let the enemy choose the battlefield, you have to seize the initiative.
Question: Do you think Linux might force Microsoft or other mainstream vendors to change their strategy or their battlefield?
Answer: To some extent I hope so. Let me explain with a metaphor I think is appropriate: let’s look at the five biggest software companies in America today, and compare them to the five biggest American car manufacturers in the 1950s. Back then all the car companies competed on new features and accessories, putting out new models every year.
Sound familiar? Now every year there’s a new version of Windows. Back then the job of the car companies was to get people to pay attention to each year’s model update. So the 1951 Cadillac and the 1952 Cadillac were very different. Naturally people would pay attention to model year. Why? Because the companies wanted people to care what model they drove, what year it was—the new year’s new model would bring them huge profits. Even though last year’s model still worked perfectly well, the companies still wanted to sell the new one.
Linux and I are like the Japanese car industry back then. Sure, historically there was the oil crisis of the 1970s and other factors. But the essence of the problem was that the American car industry didn’t care about quality, only about new features and yearly new models. Then Japanese cars appeared. At first they had quality, though not necessarily much fame, but gradually users realized: “Hey, Japanese cars are just well built. They may not have tailfins, but really, why should I change cars every year?”
I believe that is an accurate metaphor. If Microsoft can change the way it works the way the American car industry did, I would be very happy. If Microsoft really can change and compete with Linux on quality, that would be my victory.
Question: Microsoft’s Steve Ballmer (the number-two man) recently mentioned opening Windows source code. What do you think of that?
Answer: All my development work is done on the Internet, and the first thing you learn on the Internet is that talk is cheap. (and the first thing you learn on the Internet is that talk is cheap.) Unless I see results, I don’t believe any words, and that applies to everything, including what Ballmer says. Seeing is believing. (Show me the money!)
7. Linux vs. NT
Question: Speaking of Ballmer and his real intentions, do you think Microsoft is genuinely afraid of Linux, or is it just using media hype to deal with the Justice Department lawsuit?
Answer: I believe at first it was a deliberate move by Microsoft, especially last autumn when Linux had just started attracting attention. Nobody inside Microsoft would have taken Linux that seriously, but they saw that it could be used as an argument in court. But they miscalculated: the result was that a lot of people in the media became interested—“What exactly is this Linux?”
Question: Do you think Linux could become more widely used than NT?
Answer: If we are only comparing Linux and NT, then my answer is yes. Windows NT and Linux have roughly similar market share. In the future Linux will surpass NT. When that day comes I won’t be too surprised. The real goal is the desktop system. That won’t happen in one or two years, but maybe in 3, 4, or 5 years.
Question: Do you think within two to three years Linux will challenge Microsoft’s position in the desktop market?
Answer: Three to four years. I hope that by then Linux will be one of the choices for non-technical computer users.
Question: And what effect would that have on the computer industry and on Microsoft?
Answer: It may not go that far. I just think a market where users have no choice is unhealthy.
Question: You use the word unhealthy?
Answer: Yes, unhealthy. You have a very broad and complex market, but one company controls most of it. But in the end, monopolies cannot last.
In emerging markets, one company often holds enormous power. The oil kings of the past did; the auto industry of the past did; the computer industry of the past did too, when people thought IBM was unbeatable. In the end it always develops into a situation where five or six big companies stand side by side, and that is what is stable. So I think the current market is unhealthy.
8. The road over the next 5 years
Question: What do you think about the conflict between free software and commercial forces? Some purists dislike Redhat or any other company selling commercial Linux versions; they think Linux should remain free forever.
Answer: I try to avoid seeing things as only black and white. My view is: with Linux, you really can avoid commercialization. You can download everything you need from the net and not pay a cent. But frankly, I don’t want to download everything anymore. Now, if I have a new machine, I’ll put in a Redhat or Suse CD, install it, and then add the other things I need myself. I don’t have to pay for the CD, but if necessary, I’m willing to pay. They really do provide a real service. That’s choice. If you insist on seeing things as black and white, and think Linux should be completely free, then that too is your choice, but I think it limits you.
Question: It seems you’re not worried that Redhat and companies like that might manipulate Linux?
Answer: That’s because companies like Redhat won’t want to take over kernel development—they understand what that would cost. They will want many kernel developers among their employees, so that if problems arise, their own people can handle them.
But they themselves don’t want to get into kernel development. A decent commercial company will always spend time on market research, figure out what customers are doing, what additional features or value can be brought to users, make sure the product is easy to install, and fix important small holes. The key is the finished product—finishing a product, and marketing and organizational operation. (It's about finishing touches, and it's marketing and logistics.)
Question: What do you think Linux will become in 5 years?
Answer: To me, the most interesting thing has always been its wide variety of uses. I’m interested in the embedded market, because there you get special uses you won’t find elsewhere. From a technical perspective, supercomputers are always sexy, but on the other hand they’re also interesting because they have special requirements you won’t find elsewhere.
I think the most interesting market is the desktop market, because the desktop has no specialization at all. And in turn that means there will never be one perfect solution—people need too many different things. Technically, this is an extremely difficult problem, and that’s why I pay special attention to the desktop. I hope that within 5 years you will see Linux become one choice for desktop systems. Maybe it won’t replace Microsoft, but at least it will become a real choice. When someone goes to a computer store in 2004 to buy his first computer, he will stop and think for a moment whether he wants Linux, Windows, or MacOS.
Question: Linux and the open source model seem to work very well. Are there any other new business models that interest you?
Answer: There is one that interests me but doesn’t convince me: the “Internet phrasing” business model, which consists of putting an E- in front of a company name and then inflating the stock market value by an order of magnitude. I find that interesting, but in the long run it won’t succeed.
Question: Then what do you think will happen?
Answer: I think traditional industries will take the Internet very seriously. A small number of companies out in front will become famous in one stroke and then do well; among the rest, maybe not even one out of ten will have a chance of success. Maybe Yahoo will survive—it has such a great name that it can make money from the name alone—brand recognition is still very important. But they won’t make money through their present business model; maybe they already understand that themselves.
Question: One last question: what do you think will happen to software product prices?
Answer: Back then PC companies and Microsoft entered a market where software was generally expensive by making cheap software. Look at those small but aggressive companies—for example Borland, now called Inprise, which became famous through compilers, or Microsoft, through BASIC—that’s how they entered the market. What is frustrating is that now Microsoft has such a firm grip on the market that a new round of price increases has begun. Inprise caused a sensation back then by being the first to introduce a compiler priced under 100 dollars, but that sensation has been forgotten. We are back to an era of high software prices, because once again big companies can say, “This is how much you ought to pay.”
I hope we will have another sensation like that. The present economic model allows software to be sold at high prices, but I believe that won’t last. This makes some people think I’m a communist, but I’m not—I absolutely believe in competition. I just believe competition will win, and prices will fall again.
我完全同意设想建立DOS组织“DOS联盟” ，也就是说和Wengier、以及“起步”站长莫老师等DOS战友一起来建立这个“DOS联盟”，以发展我国自主OS（操作系统）的高度去完成我们共同的愿望。
------党委书记
StickyI
. This essay still represents one of the most profound criticisms of the new conservative digital hypersterism.