Steal This Program
Richard Stallman doesn't do Windows. Instead, the maverick computer scientist created his own operating system -- and he wants to share it with you. Stallman, one of the hackers who made the MIT artificial intelligence lab famous in the Sixties and Seventies, has dedicated his computing career to a seemingly quixotic mission: making software free. Well, almost. At software trade shows, Stallman can often be found exchanging his wares for cash. True, in the sea of a glitzy, high-tech promotional displays, the booth for Stallman's Free Software Foundation (FSF) looks more like a Cambridge yard sale-a white plastic table stacked with manuals and books, T-shirts dangling over the edge. But like the Lotus booth across the way, the piles of CD-ROMs have price tags: For individuals, the price for a Stallman disc is sixty dollars. So what's the deal? "I develop free software," explains Stallman, a pale, Tolstoyan character with a mass of black hair and a long beard. "I do not necessarily distribute it for free. Free software is a matter of freedom, not price. Think of free speech, not free beer." What this means is you can take one of Stallman's programs, scour its source code, figure out how it works, and change the code to better meet your needs. You can copy it and share it with friends, even sell it if you like. In a National Public Radio-like spirit, Stallman believes that those who can shell out sixty bucks for the CD-ROM should do so, and those who can't-well, he'd like them to have a chance to use the software, too. So far, with the aid of numerous benefactors, Stallman has been able to set up and maintain a small army of full-time programmers. And thousands of volunteers have joined his crusade via the Internet. Stallman's quirky philosophy is, to put it mildly, not the way the computer industry does business. Companies like Microsoft sell proprietary software: You can use it, but you certainly don't have access to the code, and you can't tinker with its insides. And if you copy it and share it with friends-just hope that Bill Gates's lawyers aren't peering over your shoulder.For the renegade Stallman, this difference is not a mere question of property law; it's a full-fledged moral issue. "I don't have the right to stop you from sharing with a friend," he says, "just in the hope that I'll make more money by stopping you. That's evil." The notion really gets Stallman worked up: "If there were a Satan," he continues, "Satan would be going around trying to tempt you to promise not to share with your neighbor."Stallman's stand against corporate computer culture has not gone unnoticed: He has been touted as "one of the greatest programmers alive" by HotWired magazine and was named a MacArthur Foundation Fellow in 1992. Of course, for most of the folks out in Silicon Valley, Stallman's share-and-share-alike mentality simply does not compute. Fumes Nathan Myhrvold, Microsoft's chief technical officer: "Richard has this quaint notion that software should be free. I say quaint because he doesn't think hardware should be free. Well, why not? The notion that software doesn't deserve as much protection as a Mies van der Rohe or some Herman Miller chair...goddamn it! If someone creates an interesting piece of software that someone sits in front of all day, I think he should at least be able to charge as much for it as the fucking chair he's sitting on." It's easy to snipe at Stallman's arguments. For one thing, his egalitarian impulses are clearly those of a member of the programming elite: He believes all computer users ought to be equal, but it seems that when he talks about these users he has only the most serious computer aficionados in mind-the type of folks who dream in UNIX code. Yet ironically, Stallman's hacker nostalgia has begun to look to some like forward thinking. His key argument-that electronic information is destined to be widely and cheaply copied-is one that even Microsoft and Lotus must now grapple with, particularly in this age of Internet downloads. After all, how can you make money from products that can be so easily bootlegged? As Bernie Greenberg, a Boston programmer and former Stallman colleague, puts it, "Stallman understands that we're leaving the age of a barter-based economy-where you exchange value X for something of an equal value-for an economy where information is of value. And there's no cost to copy information. So, the whole paradigm of 'you give me your wife, I'll give you my horse' has to be reexamined from the top down."Although Stallman's hacker principles may strike some as off-the-wall, they are a natural outgrowth of the heady atmosphere of MIT's artificial intelligence lab in the Sixties and Seventies. Stallman became a part of that pioneering group in 1971, while still an undergraduate at Harvard. Even as he was earning a bachelor's in physics, he spent his nights at MIT, working on operating-system development and programming in computer languages like Lisp.Stallman not only loved the intellectual activity but also felt at home in the community of programmers who worked together, shared information, and forbade the use of passwords and security systems. Among their many credos, according to Steven Levy's classic book, Hackers, were: "Information Should Be Free," "Access to Computers Should be Unlimited," and "Mistrust Authority." The astonishing successes of the artificial intelligence lab, however, also ensured its doom. By 1983, most of the hackers had been lured away by private companies. Only Stallman, one of the most gifted, refused to leave the place he referred to as "home." ("I was heartbroken about it," he recalls. "What I loved most in all the world was gone.") While his friends were off launching software start-ups, Stallman continued to work for what he believed in-equal access to software programming code. Stallman continued at MIT as a staff programmer until 1984, when he resigned out of concern that the university might dictate how his software was distributed. MIT let him keep his old office, though, and a year later Stallman launched the FSF. Its motto: "Stamp out software hoarding!" With the advent of the Internet, Stallman was almost able to re-create those idyllic days at the artificial intelligence lab. "We created a new hacker community," he says, "not geographically localized, but a community that essentially everyone is welcome to join if they're capable." Now, people from all over the world write, test, and modify new tools and utilities; the pool of collaborators ranges from academics to armchair enthusiasts. E-mail lists devoted to FSF programs collect complaints from users; each list is read by a "maintainer" charged with fixing these bugs-largely by rustling up volunteers.Stallman's first big success was the creation of EMACS, a flexible text editor that allows programmers to easily enter and streamline computer code. EMACS soon became the software developer's equivalent of a Swiss Army knife-a basic tool kit you couldn't do without. Of course, it wasn't perfect. Through the Internet, EMACS enthusiasts began contributing improvements to the source code. The result of this collective computing: Over the past ten years, EMACS has blossomed into a large and varied system that's used for everything from reading e-mail to keeping a calendar. In a nod to its seemingly endless uses, the EMACS logo is an overflowing kitchen sink. Stallman estimates that thousands of programmers have tweaked the original EMACS code; 160 of the largest contributors receive acknowledgment in the most recent manual.Stallman's vision, of course, reached far beyond EMACS. He proceeded with plans to write an entire operating system-the collection of basic programs that organizes standard computer functions like file storage. After laying this foundation, Stallman and his colleagues would then create application programs, with the ultimate goal of allowing people to run their computers entirely with free software. In what some companies viewed as an aggressive move, Stallman decided to model his operating system on the popular proprietary system UNIX-because, he announced in an on-line manifesto, "compatibility makes it easy for UNIX users to switch" to free software. (In this spirit, Stallman cheekily christened his project GNU, for "GNU's Not UNIX!") For several years, the FSF programming community worked non-stop on GNU. By 1992 all but one part of the operating system had been assembled; what was missing was the "kernel"-the code linking the parts together. Fortunately, help came from the University of Helsinki's Linus Torvalds, who released his own free kernel, called Linux. The FSF quickly made Linux compatible with GNU, and Stallman's dream was effectively realized: Computer users no longer had to purchase any proprietary software to get their machines up and running. (Ever the perfectionist, Stallman will release his own kernel this fall.)Today, Stallman's software empire continues to expand. GNU software tools are now used in such far-flung places as Wall Street trading floors, universities, and semiconductor factories. The GNU compiler, which distills programming languages like C++ into raw machine code, is considered by many to be the best on the market, particularly because it can be used on a variety of different systems. Stallman single-handedly wrote the compiler, which has since been updated and enhanced by thousands of others. Professor Gerald Sussman, of MIT's electrical engineering department, who hired Stallman as a staff programmer in the Seventies, describes Stallman's devotion as "a certain kind of sainthood." But Stallman sees his achievements in more earthly terms. "We've shown that the free software movement can deliver a broad spectrum of good quality software," he says. "We have delivered what people said a decade ago would be impossible without a specific business structure behind it." Now that hundreds of thousands of people make use of free software, what's next? A web browser, for one. Stallman also plans to develop word processing software. It's a tantalizing notion, one that might bring Stallman's ideas more directly into the mainstream. Instead of waiting the years it takes for Microsoft to release (and charge for) a new update of its popular Word program, users of Stallman's software could e-mail suggestions directly to the FSF Web site. Scientists could modify the program to deal with complex mathematical notations, while humanists might help develop the ultimate poststructuralist thesaurus. These specialized versions would each be available for downloading on the Internet. All that you'd be missing would be the shrink wrap and the styrofoam.Stallman maintains that not only is free software good for society, but it is also more reliable than proprietary software. After all, thousands of programmers think better than one (or even a hundred). Recent academic studies bear him out. Computer scientist Barton P. Miller and his colleagues at the University of Wisconsin tested the utilities of seven commercial UNIX systems as well as GNU in 1995. The tests involved subjecting various utilities to random input streams to see how often they crashed or got hung up. The researchers found that the commercial UNIX utilities had a failure rate that ranged from 15 percent to a whopping 43 percent. In contrast, the failure rate for GNU was only 6 percent. Miller concluded that "the reliability of the basic utilities from GNU...was noticeably better than those of the commercial systems." Microsoft's Myhrvold denies that free software is more reliable. Tellingly, however, Linux-based GNU systems have been adopted by hundreds of companies, such as hospitals, for whom reliability is crucial. The Roger Maris Cancer Center in Fargo, North Dakota, uses free software to track patients' medical histories, coordinate drug therapies, and gather data on patient demographics. Stockbrokers at Merrill Lynch share information on an in-house computer network running free software. Even NASA has become a free-software convert: The agency uses Linux-based GNU as the foundation for a new supercomputer that tracks the trajectories of space "junk," such as dead satellites. But just as free software seems to be hitting its stride, the FSF has been forced to lay off staff. Whereas the group once paid fifteen full-time employees, it now has seven. Hewlett Packard, which gave grants of $100,000 in 1989 and $75,000 in 1990 cut back on giving during the recession, and never resumed it. Stallman says he was happy to pocket a total of $50,000 for the FSF during a recent trip to Japan, not to mention an additional $8,000 for himself. (Most of the FSF's funds, Stallman reports, come from small donations, seminars, and payment for CD-ROMs.) Despite these occasional difficulties, Stallman appears unfazed. "We're here to say," he says confidently.Although Stallman envisions himself as a galactic star battler confronting the great money-making, freedom-stealing forces of the world, it may be that the two sides of this battle can coexist. Says Microsoft's Myhrvold, "Long before Richard started this, there was free software, and long after Richard is gone, there will be free software-free in his meaning. Richard Stallman can be the Pied Piper of free software and if the world's software designers decide to become the world's impoverished artists-then fine. Great. I think it's unlikely, and my prediction is that ten years from now it'll be about the same hunk of the market that it is today, which is about the same as it was ten years ago. Which is to say, it's there, it's vital, it's important to some people, but it's not some tidal wave that's about to crush the rest of the business. It's a footnote." Myhrvold may be right. Then again, the Internet-the fastest growing area of the computer market-has in some ways mimicked Stallman's vision. Stanton McCandlish, program director of the Electronic Frontier Foundation, says, "The free-software model has been instrumental in the development of the Internet." Because the first Net browsers (like Mosaic) were created at universities and given away for free, it is well-nigh impossible to sell one these days. Microsoft's Internet Explorer, for example, may be downloaded for free on the Web; you only buy it if you want a full manual. (Of course, no matter how much you pay, you'll never get Microsoft's source code.)With software so easy to download, companies may be forced to generate revenue from support services rather than from the sale of diskettes. In fact, at least two companies have made millions by providing customer support for businesses that run Stallman's free software: Cygnus Solutions, of Sunnyvale, California, and Ada Core Technologies, in New York City. Stallman fully supports such ventures. He says he's "not against people making a buck"-as long as his source code remains open to all.as for the various schemes companies have come up with to protect source code, Stallman sees them as woefully shortsighted. In his view, a software program and a chair just aren't the same thing: It's a terrible mistake to lump together the concepts of physical and intellectual property. "Our intuitions about physical property come from the fact that a given physical object can only be in one place," he explains. "So for example, you don't want your car to be stolen because, if it were stolen, you couldn't drive it. But if someone snuck up to your car with a car copier when you weren't there, and plugged the copier into the car and out the other side of the copier, another car started to expand and eventually it was full sized and he disconnected the car copier and got into his new car and drove away and the next day you came out and your car was there just the same as it had ever been, you wouldn't feel that your car had been stolen, because you'd still have it."Stallman may be a great kind of car salesman of the information age. But Microsoft's Myhrvold gives the car metaphor a different spin. The computer revolution, he argues, resembles the rise of the automobile in one crucial respect: The less a user needs to know about the machinery, the more popular it becomes."If you had a car in the year 1900," he says, "you were an enthusiast and you went out there and you cranked it and you changed your own tires and there weren't any roads and you carried your own gas and it was a big deal. It wasn't an efficient means of mass transportation. The very thing that made computing popular was moving away from the notion of everyone being a hacker."Stallman, of course, bristles at the idea of deciding what the public does or doesn't want. "To say that all those other people don't want to know what's inside a program mystifies software," he declares. "It's saying that if you're not one of the priests of software then we're sure you won't understand. That's just not the right way to treat people."