Jump directly to the Content
Jump directly to the content

Alan Jacobs

Life Among the Cyber-Amish

Computer Control, Part 2

This is the concluding article in a three-part series.

Article 1: Computer Control

Article 3: The Virtues of Resistance

The one unique benefit that was creating this enthusiasm was not that this stuff was better or faster or cheaper, although many will argue that it was all three of those things. The one unique benefit that the customer gets for the first time is control over the technology he's being asked to invest in.
—Bob Young, CEO of Red Hat Linux, as quoted in Glyn Moody's Rebel Code

Sometime around the turn of this year, IBM started running a series of television commercials featuring—for reasons not clear to me or, I suspect, to many other viewers—retired professional basketball players enacting brief allegorical dramas about business computing. George Gervin, a.k.a. the Iceman, is the coach of this team, and some of his players include Firewall (played convincingly by the now portly and indeed rather wall-ish Bill Laimbeer, once a Detroit Piston) and Middleware (Xavier McDaniel, formerly the intimidating X-Man, and somewhat thicker in the middle himself). But for me the most interesting member of this team is "acted" by Detlef Schrempf, one of the more recently retired of the bunch—he was playing in the NBA as recently as last season. As Schrempf's name perhaps indicates, he's no boy from the 'hood, but rather a Germ an import, one of the first European players to excel in the NBA. Schrempf, the only player in the commercials who gets to dunk—perhaps because he's the only one who still can—plays "Linux," and the commercials identify Linux as the team's new star. This preeminence is not unrelated to Linux's willingness to work for "peanuts," simply because "he loves the game."

The casting of Shrempf was a witty move by the makers of the commercials, because Linux (the real Linux) is a computer operating system that originated in Europe—specifically, in Finland. In the last decade it has risen from a fragmentary set of routines whose existence only a few dozen hackers knew of, and still fewer cared about, to an increasingly widely used and widely discussed alternative to Microsoft's Empire of Windows. And indeed, it still works for peanuts.

(Throughout this essay I will use the term "hackers" to refer to highly skilled people who mess around in the digital innards of computers because they love to, whether they get paid for it or not—that's how they refer to themselves. People who illegally "hack" into systems where they don't belong constitute a subset of this larger group.)

The story needs to be told in some detail, because the rise of Linux has serious implications for the future of computer technology—and, more to the point, the future of human beings' use of computer technology. The story goes beyond Linux itself; it is the story of "open-source software," and the open-source movement will, in the coming years, have a lot to say about the extent to which we use computer technology or it uses us.

Linux, or rather the first attempt at producing the "kernel"—the most elementary and universally necessary routines—of what is now Linux, was written in the latter part of 1991 by a student at the University of Helsinki named Linus Torvalds. Torvalds started to write the code in order to test a possible alternative for another OS called Minix, which had been created by a programmer named Andrew Tannenbaum as a teaching tool: along with an accompanying book, Minix helped computer science students understand how an OS works. For this purpose Minix seems to have been excellent, but students who learned Minix often experimented with it to see if they could improve it and make it more functional for a wider variety of purposes. Linus Torvalds was one such student.

But, before we proceed, let's pause to deal with a question: what's with all the X's? Linux and Minix, yes, but investigators of the history of computing will also encounter Ultrix, AIX, HP-UX, and a number of other OS's that don't announce their parentage quite so openly. And indeed it is a matter of parentage, for all these systems derive from one Ur-father: Unix.

Unix was written in 1969, primarily by Ken Thompson, a programmer who worked for AT&T's Bell Labs. Thompson and his colleague Dennis Ritchie had been working on an ambitious multi-company project when the whole thing fell apart; but they thought that some of the work they had done on the project could be rescued and adapted for relatively small computers. Thus Unix was born.

And very soon Unix began to produce offspring, because the code which Thompson had written was made available to a wide range of programmers and software engineers in a wide range of universities, government agencies, and businesses. Improving Unix became a collective endeavor, and the OS got much better very quickly because of the many minds working on it. This experience would later become a key, perhaps the key, principle of the open source software movement: as one of that movements's key theorists, Eric Raymond, wrote, "Given enough eyeballs, all bugs are shallow"—that is, even the most intractable software problem will eventually yield if enough people devote themselves to solving it.

But soon the Unix world became fragmented: versions of Unix proliferated, each tailored to the needs of a particular company or adapted for a particular set of hardware. Unix ceased to be "open," as many of these companies refused to share their versions; a spirit of proprietary ownership replaced the collective, cooperative spirit of Unix's early days. Indeed, in the 1970s a conviction began to emerge that the "intellectual property" embodied in a piece of software—its code, the instructions it gives to a piece of hardware, written by persons just as novels are written by persons—had an economic value that the first generations of computer professionals either had not noticed or had not cared about.

Such neglect may seem odd to us, knowing as we do the vast fortunes that have been made in software, but in those days computers were used by very few, chiefly in universities and the government. Programmers paid by those institutions could not make more money by keeping the code they write secret, so they freely shared it with others. Only with the great expansion of business computing and the advent of the PC did it become evident that any software that enabled non-experts actually to use a computer—a rather forbidding machine, after all—pretty much determined the value of the machine itself. One of the first computer professionals to see this point was a fellow named Bill Gates: in 1976 he wrote an "Open Letter to Hobbyists" to protest unlicensed use of some software (a version of the programming language Basic for the MITS Altair, the first personal computer) written by him and his partners in the young company then known as Micro-Soft.

Fragmentation and proprietorship; these were the enemies for many ambitious programmers in the 1980s. When Andrew Tannenbaum wrote Minix, he employed the basic principles that underlay Unix; the ancestry of his OS was very clear to all knowledgeable observers. But he couldn't use any preexisting Unix code without being in violation of copyright; he had to write Minix from scratch. Linus Torvalds was in the same position when he started coding Linux. (But for all practical purposes both Minix and Linux are versions of Unix; they use very similar commands, and anyone with Unix experience is perfectly at home in Minix, almost so in Linux.) Both programmers made their work available to the community of hackers, seeking affirmation where possible, correction where necessary. But at that point their paths diverged, and the divergence explains why Linus Torvalds and his OS Linux are increasingly famous and influential, while Andrew Tannenbaum is little known outside the circle of computer professionals.

Minix, as I have said, was written for a particular purpose: as a teaching tool. Therefore, when the many hackers who used Minix reported bugs and offered fixes for them, or suggested better ways for certain subroutines to run, or argued for the need to extend Minix so that users could run more kinds of applications on it—do more things with it—Tannenbaum decided which suggestions to follow on the basis of his conception of what Minix could and should be, given the function for which he had designed it.

But Torvalds had no such constraints. His sole interest was making Linux a better OS—and a complete, fully functional one, not just a "toy"—so, at first anyway, he eagerly incorporated almost every useful suggestion. For the hackers who sought to participate in the project, this was heady stuff: they were being invited to participate in the making, not of some minor utility or limited application, but of the very operating system itself—the defining structures of a computing environment.

In Rebel Code, Glyn Moody tells the story of Matt Welsh—later the coauthor of Running Linux, but when he discovered the OS a student at Cornell—who installed Linux and almost immediately "got the fever. It was all about doing things yourself, adding new features, and constantly learning more about how computers and operating systems work." (The work of many of these hackers can still be found among the myriad applications and utilities that come with the versions of Linux available today.) Soon there was an entire virtual community, comprising people from all over the world, investing thousands of hours in the development and improvement of Linux.

As a result, Torvalds started getting far more input than he could handle, but he kept plugging away as best he could, and Linux grew and grew—not only in size, as it incorporated more functions, but also in cleanness and stability. Linux devotees could be heard to brag that they didn't even know how to make their OS crash—even as versions of Windows became ever more "crufty" (that is, in hackerese, infested with extraneous, outdated, and badly written code) and prone to every sort of crash or freeze imaginable. Moreover, early in its development, Linux started incorporating utilities from the BSD (Berkeley Standard Distribution) of Unix—one of the less rigorously proprietary of the Unix variants—and imported applications produced by the GNU Project. The most comprehensive and ambitious single organization in the open-source movement, GNU is directed by mit's Richard Stallman, a brilliant and obsessive programmer who hates with a prophetic passion the proprietary model of software engineering, and in that cause founded the Free Software Foundation.

(For years now Stallman has had to deal with misunderstandings of his use of the term "freedom." He does not mean that software should never be sold, but rather that its source code should be made available to anyone who wants it for improvement and modification. To do otherwise is, in Stallman's view, "immoral," because it violates certain freedoms he thinks vital: thus, the Free Software Foundation. "Free speech, not free beer," as Stallman helpfully encapsulates the distinction."1)

Other programmers, individual and collective, pitched in, and soon Linux boasted an impressive array of applications written or adapted for it. Companies began putting together "distributions" of Linux—the OS and key applications bundled together, to save people the work of assembling a whole system out of the bits and pieces that were floating around the Internet—and programmers started buying them.

Fans of Linux, many of whom were also contributors to its cause, started dreaming big dreams: could it be possible that this almost homemade OS would eventually rival Windows itself? A ridiculous notion, akin to imagining that a group of Amish farmers could displace Archer Daniels Midland as a kingpin of American agribusiness—because, in a way, the open source people are the Cyber-Amish. Like the Amish, they aren't Luddites (they don't repudiate technology altogether), but they advocate a certain small-is-beautiful approach to technology that keeps power in the hands of individuals, informal networks, and communities of practitioners. It would seem that, by definition, such folks couldn't be players on the national technological stage.

But certainly the quality was there. Linux was both stable and powerful—that is, it (and its applications) rarely crashed, and it could run quickly and efficiently on machines with far less memory and processor power than Windows or the Mac OS demanded. For many programmers and systems administrators it was obvious that Linux offered stability and functionality that they couldn't find elsewhere. The stability of Linux was particularly desirable for people in charge of maintaining websites, and as the Internet became a more pervasive part of business culture, Linux—running the Apache web server software—soon became the system of choice for web administrators, first outstripping IBM's server technology and then Windows nt. (Once IBM realized that they couldn't provide service to compare with Apache servers running on Linux, they decided to buy into the Linux alternative—thus the television commercials noted at the beginning of this essay.) Maybe Amish agribusiness isn't such an oxymoron after all.

But something was missing. At this point everything Linux could do it did from the "command line"—that is, the user gave the computer instructions only by typing commands like mkdir test2 or chmod go-w test2. But ordinary users, even those who may have started on computers back in the command-line days of DOS or MS-DOS, had long since grown accustomed to a screen dominated by images and icons, and they depended on navigating that screen with a mouse or some other pointing device. If it were ever going to make a splash in the world of desktop computing, Linux needed a GUI, a Graphical User Interface.

But it wasn't until the late 1990s that it got a workable one—in fact, two: KDE and GNOME (the latter a product of Stallman's GNU Project). In 1996, when the German computer scientist Matthias Ettrich announced plans to develop KDE, he referred to it as "a GUI for endusers"—"endusers" being people like me, who don't know how to do much, if anything, from the command line, and are therefore wholly locked into the pointing-and-clicking world of the graphical desktop. With the arrival of KDE and GNOME, the Linux world claimed to be ready for me. But was I ready for it?

Setting up Corel Linux is a snap, but you should be at your mental and physical best.
—Corel Linux for Dummies, p. 16

The story of my life with Linux is a comic tale of misadventure, gross error, the occasional ephemeral triumph, and lots of cursing. At least, I'm trying to see it as a comic tale—though to this point there have been more tears than laughs. But, since I am writing these words on a computer running Linux, the experiment hasn't been a total failure. (Yet.) Linux is now a regular part of my computing experience; though how central it will be, and how long I will stick with it, are questions still unanswered.

I explained in my previous essay on this topic that in recent years I have become increasingly frustrated with my lack of control over the work I do on my computer: control in two senses of the word. First, I don't know much about how my computer actually works: how it carries out my commands, how it does what I want it to do—or (and when this happens my frustration becomes acute) how it does what it wants to do. Second, the dominance of two companies—Microsoft and AOL Time-Warner—over the world of American computing has made it increasingly difficult for computer users to exercise meaningful choice over the way they interact with their computers. I also explained that reading Neal Stephenson's little book In the Beginning Was the Command Line—with its distinction between the technologically knowledgeable Morlocks and the technologically inept Eloi—was a catalytic moment for me.

But (and I did not explain this in the earlier essay) I think I can remember the very moment I decided I had to extricate myself from the condition of the typical computer user. In a weak moment I had acquired Microsoft Office 2001 for my Macintosh—not for everyday use, but in order to share files more easily with colleagues and friends in the Windows world. But when I tried to work with Microsoft Word I was maddened by its insistence on divining what I wanted to do—maddened because the stupid program kept guessing wrong: its skills of divination seemed uncannily inept, and I spent a lot of time working my way through the Preferences screens, trying to disable the more assertive of the program's default settings.

This was not difficult, but it was tiresome, and since the more I used the application the more often it would do unexpectedly annoying things, after a while I found myself just allowing Word to make its own decisions: "No," I thought, "that's not how I want to format that paragraph, but it's too much trouble to fix it, so I'll leave it as is." At that point I realized that I was becoming the kind of user Microsoft wanted me to be: passive, accepting, ready to accommodate myself to whatever Word chose to do.

Thinking about this experience, I am put in mind of something I read in Glyn Moody's Rebel Code: a comment by Bob Young, the president of Red Hat Linux, about the kinds of people he meets at gatherings of Linux users: "I'm going, ok, let's see, my target market is rocket scientists at NASA or it's blue-haired art students in Toronto." The first group, the "rocket scientists," he had presumably expected to find: they were amateur or professional hackers, do-it-yourselfers—the lineal descendants of the "hobbyists" who, a quarter-century ago, had built their own Altair and Heathkit computers. But surely that other group was a surprise to Young. Who would have expected that significant numbers of people would gravitate to Linux not because they were hard-core techno-geeks, but because they wanted to protest against the hegemony of corporate America as embodied in Microsoft and Intel? Well (I thought), if the blue-haired art students can do it, so can I.

So my course was set. But once I decided I was going to try Linux, I needed two things: software and hardware. The software was easy enough: having studied cnet.com to find out which distribution of Linux was right for me, and having decided that (despite its popularity) Red Hat would probably lead me pretty quickly into problems I couldn't solve, I went down to my local Borders and bought Corel Linux for Dummies, which included Corel Linux on a cd.

And hardware wasn't too much of a problem, either. Since I had no intention of messing with my Mac (even though several distributions of Linux run on Macs), I called Wheaton College's Computing Services department to see if I could borrow something—surely they had an old PC lying around that no one was using. And it could be pretty "old," at least by the rapid evolutionary standards of today's computers: after all, don't Linux aficionados point with pride to its ability to run on machines that the Windows world deems thoroughly outdated? Even a Pentium I processor would do, my Dummies book told me.

I'm not sure what the people in Computing Services thought of my strange request, though the news that an English professor wanted a machine on which to run Linux could scarcely have filled them with joy. But they were, as always, polite and accommodating, and soon Ted Myhre showed up in my office with a cart containing a grubby, dusty old Gateway computer and monitor.

I've been using nothing but a laptop for several years now, and I felt as though someone had dropped a rhinoceros on my desk. The thing must have weighed fifty pounds; I managed to move it to a small second desk, which creaked under the weight, and when I pressed the machine's On button it shuddered and heaved like a beast under anesthesia. I inserted the Corel Linux cd and restarted, but it wouldn't boot. I checked the bios (the pre-OS utilities on most PCs that control some of the computer's most basic settings) and found that this machine could boot either from the hard drive or the floppy drive, but not a cd. So I found another PC, inserted the cd, and copied the necessary files to a floppy disk. Then I returned to the behemoth, made sure that it was set to boot from the floppy drive, inserted both the cd and the floppy, and restarted. Soon enough I was given the option of wiping the hard disk and installing Linux, which I chose to do, and by the time I went out to get a Coke and came back I had Linux up and running.

Now, I felt pretty good about this. Clearly the reviews I had read on cnet.com that promised that Corel Linux (unlike several other distributions) would easily install were correct; plus, I had been able to solve a problem by myself, though to be sure it was a minor one. As I looked at the screen, however, I wasn't sure exactly what to do next. I opened the File Manager and fooled around with that for a while. I copied some alternative "themes" from the cd so I could try out different looks for the desktop, but still, it wasn't nearly as attractive as the Mac desktop, and I had that dull, sober feeling I always have when I sit down at a machine running Windows.

Beyond these matters of mere aesthetics—and I was determined to think of them as "mere"—the substantive problem, I realized, was that I couldn't connect to the Internet, and so couldn't download the many cool and powerful Linux applications I had read were available for free. I played a little Solitaire, drew a little with a drawing application, and wrote a few paragraphs with the Text Editor—pretty much the sort of thing I did when I bought my first Mac in the spring of 1985 and spent quite a while exploring the fun offered by MacWrite and MacPaint. But this time the exploration got boring pretty quickly. I turned back to my PowerBook and answered some email, the beast snoring heavily behind me all the while. I could almost feel warm, moist breath on the back of my neck as I typed.

While at my Mac, I sent an email to Computing Services asking whether I could be hooked up to the campus network, but this plea was not met with the same snappy approval as my request for an old, unwanted PC. Besides, Behemoth had no Ethernet port to connect the machine to our local network; and a prominent magic-markered command on its back pointed to the phone jack and commanded DO NOT USE. But if I wanted to do more than draw a few pictures and type plain-text files, I needed some software; so a few days later I went to the computer store to scope out the Linux shelf.

The first thing anyone notices about Linux software is how much cheaper it is than similar stuff for Windows or Mac. The various distributions of Linux tend to cost around thirty bucks, and contain not only the OS, but also dozens of applications. Most contain StarOffice (the Sun Microsystems alternative to Microsoft Office), image editing software (including the GIMP, the GNU Project's free alternative to Adobe Photoshop, which many professionals actually prefer), any number of web browsers and email applications, games and other goodies—as well as a raft of applications and utilities widely used in the Unix world. I did some quick, informal comparisons and figured that the average thirty-buck Linux distribution contained software whose Windows-compatible equivalents would cost a couple of thousand dollars. But which distribution of all this stuff did I choose? On CNET.com I had learned that Linux Mandrake gets high marks for ease of installation and use; I grabbed a copy and started for home.

Driving home, one thought occupied me: Behemoth must go. Every time I had walked into the office, since its delivery, I had winced: there it hulked, dominating the desk, blocking the light from the window, leaving me no room to work. I realized, when I considered it, that I hadn't been more persistent in asking for an additional connection to Wheaton's network because I feared that Computing Services would comply—thus leaving me stuck with that horrible machine. I've been working with my PowerBook for too long to be comfortable with computers that size. I've really enjoyed the spaciousness that my laptop provides my desk, plus the opportunity to sit back and put my legs up with the computer in my lap. So I just couldn't bear to use Behemoth anymore. I unplugged it and stacked the pieces in a corner. (I was embarrassed to ask Computing Services to retrieve it—I thought I'd wait a few more weeks, to give the impression I had taken their gift seriously.) But since I wasn't ready to abandon my Linux experiment, I was faced with a simple problem: how to get a laptop that would run Linux.

I won't bore you with the negotiations I conducted with my wife; they were subtle, protracted, and—from my point of view, anyway—ultimately successful. So I ordered a bottom-of-the-line IBM ThinkPad from an online computer store, and kept my fingers crossed that my new Linux Mandrake would run on it.

It wouldn't. I had no luck booting from the CD, but I had been there before with Behemoth, so I tried the boot floppy which, I was pleased to see, had come in the package: BOOT FAILED, the screen said. So I went to another computer and made my own boot floppy: BOOT FAILED. OK. I could handle this. Back to Corel Linux—it had worked with Behemoth, right?—so I tried that, again from a boot floppy, because CD installation clearly wasn't happening on this machine either. This time things seemed to go better, because it started installing—though in command-line mode, no images—and indeed got all the way to asking for my username before something went astray. No matter what I did, I couldn't get past this point.

Now the cursing began in earnest. But I was not ready to give up. Several "compact" distributions of Linux promising ease of installation were available for free download on the Internet. I tried three; none worked. I borrowed a more recent version of Linux Mandrake from a friend. Nope. Finally, in desperation, I went back to the computer store to try one more highly rated distribution, from the German company SuSE (pronounced "Sousa," like the composer). After all, what was thirty more bucks, given the amounts of money and time I had already invested? I brought SuSE Linux home, inserted the CD, and started the computer. It booted up and installed without a hitch. Go figure. But I wasn't asking questions; now I had a Linux laptop stuffed with applications for every purpose. I was primed to rock and roll.

It's vital to refrain from getting frustrated with the system. Nothing is earned by taking an axe—or worse, a powerful electromagnet—to your Linux system in a fit of anger.

—Running Linux, p. 43

Veterans of Linux no doubt smiled, with some mixture of condescension and sympathy, as they read my last paragraph. They know that my troubles were only beginning; it would not be long before I had burned into my brain the comment I found on a Linux website: "Nothing in Linux works the first time."

But initially the sailing was smooth. I perused the dozen or so word-processing applications and text editors newly loaded on my Linux box, found one I liked, and started drafting the first section of this essay. Once I got a thousand words or so, I figured I had better back up my file—but how? The ThinkPad had no internal floppy drive, couldn't write to a CD, and wasn't connected to the Internet. And while I owned both a floppy drive and a Zip drive, my new OS didn't seem to recognize either of them. I was able to find on the Net various "howto" documents from the many websites created by Linux aficionados to assist "newbies" like me, and I pored carefully over the pages of Running Linux (an otherwise admirable resource), but I couldn't seem to make much progress with any of my problems—perhaps as a result of slight variations in the Linux distributions. Eventually I called my colleague Jeff Beaird, an experienced and skilled Linux user, to help me get connected to the Wheaton College network (this time with the blessing and assistance of Computing Services), so at least I could email the document to myself. Jeff did this with relative ease—though even he ran into enough unexpected problems to convince me that I could never in ten years have gotten the machine properly configured all by myself. Still, I told him to leave the floppy and Zip drives to me; I was determined to do something on my own.

Well, I still haven't gotten those problems solved. My greatest achievement so far came as a result of my shutting down the computer improperly: the next time

I turned the computer on, the startup process was interrupted by this message: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY I had no idea what this meant, but I turned to my Mac and found a "howto" document on one of the Linux sites that took me step-by-step through the procedure. Soon I had the computer running normally again, and had lost no data. And I was very, very proud of myself.

But plainly I have a long way to go, and a lot of work to do; and it's not every day that I feel that it's worth it. Since I started pursuing this project, it has become increasingly obvious that the lines between the open-source world and corporate America are being blurred. IBM's relentlessly televised touting of its alliance with Linux is the most conspicuous example, but Apple's introduction of Macintosh OS X is more intriguing and perhaps more significant. Since I installed OS X on my PowerBook, I have been running what is essentially a version of Unix (the aforementioned bsd) with a visually dynamic interface superimposed on it.

Indeed, the chess program that comes with OS X is actually GNU Chess, which means that it is available to users under the terms of the famous GNU Public License. This fascinating document mandates that Apple (and anybody else distributing the application) make the application's source code—the actual instructions the application sends to the computer; which collectively constitute the "intellectual property" that Bill Gates and others have insisted on the need to preserve—available on request, and denies Apple, and anybody else, the right to prevent free access to that source code. (Because this is such a strange use of copyright—not to ensure the owner's control over his of her intellectual property, but to prevent anybody else from exercising such control—Stallman calls it "copyleft.")

And, strangest of all to anyone familiar with the whole Macintosh enterprise, OS X includes a "terminal"—a screen that responds only to typed commands—so that many traditional Unix applications, like the text editors Emacs and vi (pronounced "vee-eye"), can be run from the command line. Apple has even made the source code for the kernel of OS X available, and has invited the open-source community to collaborate with them in its development.

It all seems too good to be true—and indeed, say many in the open-source world, it is just that. To anyone who remembers or even has read about the Sixties, it's a very familiar debate: the old, intractable question of "co-optation." Are the Big Corporate Brothers co-opting the revolutionary work of the open-source hacker community for their own insidious and thoroughly proprietary purposes—just as university presidents once invited student protesters into university governance and thereby domesticated the malcontents? (After all, Apple hasn't given up corporate rule over OS X's source code.) Are the glorious rebel hackers—with the notable exception of Stallman, who won't cooperate with anyone—simply being bought off?

Perhaps; but as the literary theorist Gerald Graff once asked, in reference to the Sixties version of the co-optation debate, isn't "being co-opted" merely a derogatory description of success? Graff suspects that many would-be revolutionaries prefer to remain marginal; they seem to fear that if their ideas gained acceptance it would only prove that those ideas weren't revolutionary after all, just as Groucho Marx didn't want to belong to any club that would accept him as a member. If Graff's suspicions are right, then for many in the open source community, the developments I have been tracing here—the embrace of Linux by IBM and Unix by Apple, the creation of slick gui's for Linux, the interest of Eloi like me in the whole open-source endeavor—are clear signs that it's time to move on to something else, something less accessible to the uninitiated.

But whatever they choose to do, it seems clear to me that distributions of Linux are going to get easier and easier to install and configure; soon it will be possible even for people less comfortable with computers than me to run this rebel OS. Indeed, one of the most recent distributions of Linux, from a company called Lycoris, claims to be as easy to use as Windows or Mac, a claim that some reviewers have heartily endorsed. Still, I'm not sure how I feel about the increasing prominence and influence and accessibility of Linux.

It's certainly not a wholly bad thing. Given that there are so many distributions of Linux out there and no possibility of one taking over, anyone running Linux can escape the domination of the Big Two of the computer industry, Microsft and AOL Time Warner. And that's a very big deal, as far as I and the blue-haired art students are concerned. But the do-it-yourself aspect of the Linux experience will continue to be eroded.

Neal Stephenson has described the Linux world as "a bunch of RVs, yurts, teepees, and geodesic domes set up in a field and organized by consensus," so that the residents can work together to build state-of-the-art machines—a Sixties collectivist revision of my Cyber-Amish metaphor. Even if they continue to build machines of the same quality, something will be lost when the yurts and teepees are replaced by geometrical rows of prefabricated houses. Indeed, for the real hackers, the transformation of the field into a subdivision was essentially complete when the distributions started coming out: you can hear their prophetic denunciations and demands for a return to purity on the various websites that purport to teach you how to build "Linux from scratch."

There's no doubt, then, that Linux has been "co-opted" at least in this sense (i.e., it has become successful), and as a result the other aspect of control with which I have been concerned—control over one's use of a computer, over one's interactions with it—is being diminished. It's getting harder to tell the difference between Linux and Mac OS X; in both cases the command line lurks there in the background for anyone who happens to notice it, but soon there may be Linux users who (like almost all Mac users) never do notice it. And the Mac interface is far prettier.

It may be, then, that for people like me these are the days, the Golden Age of Linux: using Linux today doesn't require impossible-to-master technical skills, but it is demanding enough to be a source of instruction and literacy. I have definitely become far more computer literate as a result of my Linux experiment—and will continue to learn, I trust—but at this point I'm wonderng what that literacy and my newfound skills are worth. I've paid a considerable price, more in time and energy than in money, to reach this point: have I spent my resources well? Well, that's a matter that bears more reflection than I have time to give it right now. Because if I don't find my /etc/rc.config file—or, perhaps, my /etc/rc.config.d/hotplug.rc.config file—I'm never going to get this Zip drive working.

Alan Jacobs is professor of English at Wheaton College. He is the author most recently of A Theology of Reading: The Hermeneutics of Love (Westview Press).

1. The full range of Stallman's projects may be found at www.gnu.org. There you will also find that Stallman offers a history of open-source software somewhat at odds with the one I have told here, one which explains how his own attempts to create a "Unix-like operating system" called GNU were fused with the evolving Linux kernel to create an OS that should (says Stallman) be called "GNU/Linux" rather than merely "Linux." There is no doubt that the GNU Project was absolutely necessary to the success of Linux, but whether the OS should receive the bifurcated name that Stallman gives it remains in dispute. (GNU, by the way, is a "recursive acronym" that stands for "GNU's Not Unix"—a heritage, like Minix and Linux, of the need to avoid infringing on the copyrights of the proprietary versions of Unix.) Stallman's side of the GNU/Linux story is told in detail in Free As in Freedom.

Sources Used in the Writing of This Essay:

Jesse Feiler, Mac OS X: The Complete Reference (Osborne/McGraw-Hill, 2001).

Bill Gates, "Open Letter to Hobbyists" (1976): blinkenlights.com/classiccmp/gateswhine.html

Gerald Graff, "Co-Optation," in The New Historicism, ed. H. Aram Veeser (Routledge, 1989).

Stephen E. Harris, with Erwin Zijleman, Corel Linux for Dummies (IDG Books, 2000).

Glyn Moody, Rebel Code: Inside Linux and the Open Source Revolution (Perseus, 2001).

Eric Raymond, "The Cathedral and the Bazaar" (1998): tuxedo.org/~esr/writings/cathedral-bazaar/.

Matt Welsh, Matthias Kalle Dalheimer, and Lar Kaufman, Running Linux, 3rd ed. (O'Reilly, 1999).

Sam Williams, Free as in Freedom: Richard Stallman's Crusade for Free Software (O'Reilly, 2002).

Most ReadMost Shared