But what about the pack of photographers, the satellite trucks from CNN and ABC, the audience members calling their editors or their friends with live updates at every break in the action? Would they have come clamoring to Saratoga if not to contemplate Torvalds in the flesh? Perhaps not -- even with the lure of valet parking and a fancy lunch.
And yet there certainly was plenty of real meat to chew on at this particular chip demo. After years of hard work and an estimated $100 million or so, Transmeta had cooked up two tiny microprocessor* chips, dubbed "Crusoe," that may well usher in a new era of ubiquitous, low-cost, mobile computing. On display at the left edge of the stage was an array of gadgets that any self-respecting early adopter would have a difficult time not slavering over, including a "Web slate" designed to be an ultra-portable interface to the Net and a laptop with battery staying power three or four times the current norm.
So not only was this a gathering that demanded attendance, but it also unabashedly encouraged an exuberant display of high-tech fetishism. The case for mobile computing did not need to be made to this wannabe cyborg audience. Hardly a visitor came near who was not equipped with at least one cellphone, personal digital assistant, digital camera and/or wireless modem-equipped laptop.
I was no exception. Splayed out across my lap, as I sat in the small theater where Transmeta execs, grinning from ear to ear, declaimed upon their unique "code morphing"* software and the astonishingly low power consumption of their chips, lay my own cherished gadget, a brand-spanking-new Sony Vaio laptop computer of which I was inordinately proud.
It wasn't just the sleek, burnished design or the feather-like weight that pleased me about my laptop. My laptop made me happy because, in microcosm, it exemplified some of the changes sweeping through the software industry that were personified, on a much larger scale, by Transmeta's products and Torvalds' code. When I bought the machine, it came installed with Windows 98. But with surprisingly little trouble, I was able to transform it into a "dual-boot" system: Depending upon my whim, I could choose which operating system the computer loaded, or "booted up,"* first -- in this case, Windows 98 or Red Hat Linux 6.1.
Dual-booting is not for everyone. It's a geeky thing. Most people don't need two operating systems on their computers. But for me, it represented the possibility of choice in a dangerously monopolistic environment. The vast majority of computer users accept, without much demurral, that if they purchase a personal computer it will most likely come pre-installed with Windows (let's put aside, for the moment, the question of the Macintosh minority). Microsoft's control over that opening screen gives the company great power, as has been demonstrated in the Microsoft antitrust trial. By taking control of the boot-up sequence, I was rejecting Microsoft's claim to preeminence on my computer and reducing it to just another contender. It was up to me if I wanted the slick ease-of-use of Windows or the powerful flexibility of a Linux-based operating system -- if I wanted to be comforted by supposedly idiot-proof proprietary software that held my hand, or teased with the uncertainties of the free-software way of life.
Transmeta was also pursuing a dual-boot strategy. The company was placing its bets, or chips, as it were, on two operating systems: One chip was designed to work with Windows, the other one with Linux-based operating systems. To demonstrate this strategy graphically, Torvalds and another Transmeta employee, Dave Taylor, trotted out to battle each other in a networked bout of the exquisitely violent first-person shooter* video game Quake III. Taylor fought from a Crusoe-powered computer running Windows, while Torvalds wielded his weapons on a Crusoe-powered computer running Linux.
The showdown launched a photographic frenzy, as the assembled corps of camera-toting journalists surged toward the stage. Valley veterans must have been shaking their heads. Transmeta, a proud aspirant to the glorious chip-making heritage of the valley, was showing off chips that it believed could change the world, just as, much earlier, Intel's microprocessor chip had set the stage for the personal computer explosion. But the real excitement of the day was the sight of a jeans-wearing, sandal-clad young man attempting, without any success at all, to avoid being blown to bits by his opponent, who was sporting leather pants and tails.
Torvalds died, early and often, to the amused dismay of his fans. But it wasn't really his fault, nor, as he was quick to claim, could his pathetic showing be blamed on defects in Linux. Taylor, his opponent, was one of the original authors of Quake; heaven only knew how many hours he had logged hunting down foes in garish labyrinths that he had helped to create.
A 10-minute break followed Torvalds' unseemly demise. As I typed in some notes, a ponytailed man sitting beside me gave me some friendly grief for using a Microsoft product, Word, to write about Linus Torvalds. A fair criticism, I conceded. But I was planning to file my copy directly from the Villa Montalvo, I told him, and I wasn't going to take any chances with my still feeble Linux mastery when operating under a tight deadline in competition with every other technology reporter in Northern California.
As the break wore on, it occurred to me that this was a golden opportunity to ask for advice. My amiable critic was a reasonably well-known Linux advocate for one of the more high-profile Linux companies. I had been having a slight problem with the boot-up sequence on my laptop; maybe he could help.
The problem was a silly little thing. After powering on, the laptop presented me with a "boot prompt": If I typed in "dos," Windows 98 started up; if "linux," then Linux. But if I didn't type anything, the machine defaulted to Linux in just three seconds. I wanted it to wait longer. Theoretically, a small change to a simple configuration file should have solved the problem.
Theoretically. Life with Linux is one long learning curve, and for some reason I couldn't get it to work. But my Linux guru friend beside me smiled with the confidence of a veteran power user. "I can fix that," he assured me. "Want me to take a look?"
Sure, I said, and handed him my machine. His fingers flew across the keyboard, making changes faster than I could process. Then he handed it back to me, still running Linux. I had to reboot to start Windows so I could continue taking notes. But my computer wouldn't let me boot back into Windows. Something was wrong.
For a moment I suspected sabotage. I complained. My companion was embarrassed. His fingers flew again. And suddenly the machine would not boot, period. No Windows, no Linux, no nothing. A classic example of how a little freedom can be a dangerous thing. Microsoft Windows attempts, not always successfully, to hide its inner workings from you, the better to prevent you from amputating your own head; in Linux, self-mutilation is a snap.
I had to drive back to San Francisco to write my story. As I fought my way through a Pacific storm pummeling the coastal mountain range, I mulled over how best to start the piece. For a moment I even contemplated mentioning my boot-up misadventures.
Booting up, is, of course, a great place to start. The term is derived from the word bootstrap -- as in, to pull yourself up by your own bootstraps. The boot-up sequence is the first set of orders a computer receives upon awakening. Those orders initiate other orders -- other programs -- which in turn bring the computer to full possession of its senses.
The whole free-software movement, I realized, is itself a tremendously successful bootstrapping project. Starting with the simplest of objects -- the ones and zeroes that are the basic building blocks of code -- programmers have hacked together increasingly elaborate structures: programming languages, operating systems, the Internet itself.
I put aside the question of how best to introduce a hurriedly written account of a press conference. I decided that if I could just find the initial boot-up moment for the whole story of free software -- a story to which I'd been devoting my reporting career -- the rest of the narrative would no doubt unfold in logically pleasing order, like a row of falling dominoes or a sequence of coded subroutines* snapping efficiently into action.
But that raised the obvious question: Just where should I look for free software's original boot-up moment?