The Barnes & Noble Review
It's hard to imagine anyone better qualified to chronicle the computing and Internet revolution than James Gleick, who first came to wide public attention as the author of Chaos.
That's not just a cheap pun: Chaos theory sought to understand the "jagged edges and sudden leaps" that appear throughout nature, transforming systems from orderly linear growth to apparent madness. Kind of like what happened with PCs and the Web. Gleick actually founded and ran one of the earliest Internet service providers, before returning to journalism and writing the bulk of the essays collected here in What Just Happened. What better qualifications could you ask for?
Gleick ranged far and wide during the '90s and '00s: AT&T Bell Labs in 1993, where he thought he'd find the coming network revolution (that's where it had always happened before); Microsoft's Redmond campus in 1992 for a preview of Word 2.0 (which proved just as buggy as the 1.0 version he'd become obsessed with); Silicon Valley; and all over the Internet (chronicling the arrival of such phenomena as forwarded email jokes and eBay auctions). Some of it almost seems like fantasy now, if not for the fact that we've just lived through it. (Remember when a reporter scarfed up the mcdonalds.com domain, and what McDonald's had to do to gain custody?)
As with any revolution, even the smartest "micro-level" expectations often prove wrong. (Apple's Newton is best remembered as the butt of "Doonesbury" jokes; cable companies never became telephone providers en masse, and interactive CD-ROM developers never repaid their venture capitalists.) But Gleick has shown an uncanny ability to put his finger on the megastuff that really matters: from the impact of cellphones on human relationships to the collisions between fair-use rights and copyright owners, privacy advocates and the government, Microsoft and the world. On the whole, these essays hold up remarkably well, and even where they don't, they're a remarkable snapshot of times that won't soon return.
As founder of the Pipeline, one of the first portals to give PC users access to the Internet, Gleick (Faster: The Acceleration of Just About Everything) was in the forefront of providing Internet access to home users. As a New York Times editor and reporter for ten years, he also led journalism's focus on technological change. In this collection of essays, written between 1992 an 2001 and originally published in the New York Times Magazine and The New Yorker, we are led back into territory that once seemed so ominous: the buggy nature of Microsoft software, Internet homesteading, the Y2K scare, e-cash, pornography on the web, spam, and more. What is perhaps most notable is how contemporary many of these essays still seem. His 1995 article, "Making Microsoft Safe for Capitalism," for example, clarifies many of the issues at the heart of the antitrust case against the software giant, while other essays address the still unresolved privacy issues inherent in the new technology. In analyzing many of these trends with trenchant humor, Gleick entertainingly reveals their import for the lay reader. Recommended for public libraries and undergraduate academic libraries. Christopher Brennan, SUNY at Brockport Copyright 2002 Cahners Business Information.
From the Publisher
“A marvellous journey around our technology-drenched world ... The work of a master.” –The Independent
“Gleick’s a crack investigator who digs for the exceptional facts…. A worthy overview…on the brave new problems we’ve faced—and will face into the future.” –Detroit Free Press
“Invokes nostalgia for a simpler, more innocent time, before we took all this technology for granted.” –The Rocky Mountain News
“What Just Happened is a lively time capsule that examines the recent past—one that, not long ago, seemed fairly far-fetched.” –Columbus Dispatch
“Gleick is a writer blessed with a techie’s mind and insight. . . . As we further immerse ourselves into a plugged-in world, it would be wise to listen to what Glieck had to say back when.” —Book Street USA
"Gleick is the king of popular science writing." —Irish Times
"Gleick is one of America's leading exegete of the technological revolution that, like it or not, is taking over all our lives. He spends his at the cutting edge of computer and allied sciences, returning from the front with visions of the future." —The Observer
“Gleick’s essays remain pertinent.” —The New York Times Book Review
“James Gleick . . . is on the outer reaches of the electronic frontier . . . and [he] has mastered it.” —The Roanoke Times
“What Just Happened is a lively time capsule that examines the recent past— one that, not long ago, seemed mostly far-fetched.” —The Columbus Dispatch
Read an Excerpt
CHASING BUGS IN THE ELECTRONIC VILLAGE
I couldn't wait to buy Microsoft Word for Windows--rumored to be the new Cuisinart, Mack truck, and Swiss Army knife of word processing software, full-featured, powerful and, for a writer, the ultimate time-saving device. I was writing a long book, and I wanted the best. One day in January 1990, I finally got to tear open a software box bigger than some computers, and out it came. The world's preeminent software manufacturer had spent roughly as long developing this word processor as the Manhattan Project had spent cooking up the atomic bomb, but secrecy had not been quite as airtight. For more than a year, Microsoft had been leaking juicy tidbits to its waiting army of trade journalists, computer consultants, and corporate purchasers. Word for Windows (aka Winword or WfW) would be Wysiwyg (the standard acronym for What You See Is What You Get)--that is, it would display page layouts and typefaces with high fidelity to the final printed product. It would let users work with nine documents on the screen at once. It would have a macro language--a way to spend hours writing mini-programs to streamline all those little chores that can suck up milliseconds of a writer's time.
And it would also have--in some corner of my mind, I must already have known this--bugs.
Computer software is the brightest of bright spots on the American economic landscape, a consumer product evolving in a floodtide of innovation and ingenuity, an industry that has barely noticed the recession or seen any challenge from overseas. Bugs are its special curse. They are an ancient devil--the product defect--in a peculiarly exasperating moderndress. As software grows more complex and we come to rely on it more, the industry is discovering that bugs are more pervasive and more expensive than ever before. Word for Windows had big bugs and little bugs. A little bug might mean that a user would sometimes find the em dashes (—) and en dashes (–) switched. Even a little bug could send users running for their blood-pressure medication. "This bug is severe," one user railed at Microsoft when he discovered what was happening to his ems and ens. "It renders the whole setup useless for any serious work."
A bigger bug would cause unwanted typefaces to appear willy-nilly in one's documents. An even bigger bug would cause a poignant message to appear on screen
Unrecoverable Application Error
Terminating Current Application
after which the computer would crash, die, freeze, lock up or hang (the slang was evolving too fast even for Word's electronic thesaurus to keep up). The U.A.E., as it became known, would send one's current work into the digital oblivion so familiar to computer users.
Searching for help, I stumbled into an odd corner of the electronic village, a "forum" on the CompuServe Information Service devoted entirely to a permanent floating conversation about the ins and outs of Word. CompuServe is a vast electronic network that subscribers can dial into via modem to use a wide array of services from game-playing to stock quotations. Twenty-four hours a day, users from a dozen time zones dial in, read their mail and everyone else's, and post replies. Microsoft's Word for Windows forum provided instant access not only to the experiences of other users but also to Microsoft's development team--because an assortment of programmers also joined in, including the program manager responsible for Word's development.
CompuServe fosters a strange form of communication, more casual than letters, more formal than telephone conversation, and extremely public. An electronic culture has developed with its own evolving rules of decorum, its own politesse and its own methods of conveying rage, not to mention its indispensable acronyms: IMHO (in my humble opinion), FWIW (for what it's worth), PMJI (pardon my jumping in), ROFL (rolling on the floor laughing), IANAL (I am not a lawyer). The characters ;) are supposed to mean "just kidding"--on the theory that they suggest, sideways, a face that is smiling and winking.
The medium creates a modern kind of town meeting--perhaps the only real town meeting left--where opinions from the humble to the slanderous fly freely. When war breaks out in the Persian Gulf, say, CompuServe instantly opens a new forum. But never mind war--these days the waters are roiled by debate over the rivalry between Microsoft and IBM, with partisans as fierce as any religious zealot. "I have never seen such a vicious crowd," one CompuServe participant said recently, and another replied: "Hum, try the Issues forum Paranormal section where the Bay Area Skeptics and the psi-clones are mugging each other." I began spending hours each week watching Microsoft's product forum as users discussed Word's fine points, traded tips like fussy chefs exchanging recipes, and harangued the company's ever-diplomatic support crew about the essential mystery: Why couldn't the software be cleansed of all these bugs? Every so often there was progress--of sorts.
"Your vociferous complaints, while a bit overbearing, have borne fruit," the program manager, Chase Franklin, told me. "I just met with the testing engineer assigned to the problem and he has determined two things: The model 80 he was using did not hang when displaying the symbol fonts, it just took about three minutes to update the screen to a correct display, sans the symbols." Translation: My latest bug was not actually crashing the computer, as I had thought. It merely caused the computer to pause for three minutes after each character I typed.
Without quite meaning to, I had become a member of a nettlesome group of users that someone dubbed the Winword Gadfly Team. The other Gadflies were a Hollywood screenwriter teaching at Columbia University, a Toronto teacher, and a Caltech mathematical physicist. We loved our word processor--it let us fiddle and tinker until we had completely automated tasks as simple as counting the words in a chapter or as complicated as formatting a screenplay. And we hated it--because the more we used it, the more bugs we discovered. As we never tired of pointing out.
"Don't you guys ever work?" an exasperated Microsoft developer asked a few months later. "I swear you live here."
Software bugs defy the industry's best efforts at quality control. Manufacturers may spend far more time and resources on testing and repairing their software than on the original design and coding. "Debugging" is not just an integral part of the development process; it is sometimes the dominant part. Programmers are trying to combat the increasing complexity of their creations with new techniques--modular design, for example, that might contain damage as flood compartments do in a ship--but so far these have made little difference.
The problem is that software is different from other merchandise. Computer programs are the most intricate, delicately balanced and finely interwoven of all the products of human industry to date. They are machines with far more moving parts than any engine: the parts don't wear out, but they interact and rub up against one another in ways the programmers themselves cannot predict. When a program doubles in size, the potential for unexpected bugs more than doubles--far more, just as the number of potential love affairs more than doubles when the population of your office rises from ten to twenty.
So developers turn to user testing. Beta tests, as these adventures are known, have grown to enormous scales. Microsoft and IBM have each recently concluded beta tests in which tens of thousands of users participated. The Word for Windows beta test had lasted many months--but not long enough, it seemed. We customers began to feel like unwilling beta testers. And perhaps not Microsoft's favorite beta testers at that--certainly I wasn't doing my part to maintain the upbeat spirit of the company's CompuServe forum.
"When I said, 'No no no no a thousand times no' just now, you might have interpreted that to be sort of, well, negative," I conceded to one developer. A year had passed since the original January 1990 release, and strange typefaces kept showing up here and there. Some users were tempted to give up Winword altogether--but changing word processors is almost as traumatic as changing religions, and Winword's competitors had their bugs, too. Meanwhile, to temperamental writers, hoping that their word processors would become as second nature as an old typewriter, every encounter with a bug was a slap in the face. Guy Gallo, the Columbia screenwriter Gadfly, encountering the typeface problem yet again, hit his Caps Lock key: "THIS IS THE SAME BUG WE HAVE BEEN SCREAMING ABOUT SINCE WFW'S RELEASE."
That was the problem. Winword 1.0 had been updated with Winword 1.01, Winword 1.1, and then Winword 1.1a. (Microsoft was doling out version numbers sparingly.) A few bugs had been eliminated but more had been discovered, and the general impression was of sliding backward downhill. Why were we still talking about bugs that had been reported and confirmed a year before?
"You need to read your license agreement," one of the developers declared on CompuServe after I had needled him for a while. "We don't have an obligation to issue another release of the product, James, and it's warranted on an 'as is' basis. We don't even have a legal obligation to worry about your data loss." This was true. Some bug-plagued software users have tried to sue manufacturers for damages, but the courts virtually never sustain such claims.
The Gadflies discovered that Word's formatting instructions did not function as expected for footnote numbers. Worse, Winword's "templates"--documents that stored styles, customized command menus and other user information--seemed to take an increasingly long time to save. At first they had saved in seconds, like normal documents. As the templates grew larger, however, they seemed to cross an invisible threshold, and now users found themselves waiting two minutes or more--long enough to panic and reach for the on-off switch. Worst of all perhaps, typefaces--fonts--behaved in a variety of weird ways that someone gave the name "phont phunnies."
The developers said they were trying. It was a hard task: users would report problems that Microsoft's testers could not reproduce on their machines. Different computers, different amounts of memory, different documents, different combinations of software made it impossible sometimes to track down bugs, although the Winword program manager said he now had twenty test engineers tracking our reports.
But Microsoft's marketing strategists had more pressing problems. Winword was by far the dominant word processing program designed to be used with Microsoft's Windows operating environment, but a relatively small number of personal computer users use Windows. Overall, the market leader by a large margin was WordPerfect, which was known to be beta testing its long-awaited entry into the Windows market. Microsoft officials were worried. WordPerfect commands enormous loyalty, in part because--unlike Microsoft--the company makes a practice of releasing frequent free upgrades to repair even minor bugs, and in part because it maintains, at enormous expense, a toll-free telephone support line--an investment Microsoft, which says it fields 14,000 calls a day, has been unwilling to make.
In June 1991 the Gadflies arrived en masse at Microsoft headquarters, in Redmond, Washington, a suburb of Seattle. The corporate giant seemed to have disguised itself as a small college. Its bucolic 256-acre campus mingled elegant plantings with outdoor basketball courts. Its executives and developers turned out to be mostly in their twenties, wearing jeans and T-shirts. It was as if gray hair, neckties, and the words Mr. and Ms. had been banished from the earth.
Plucked from our faceless electronic existences, we had been flown in for a preview of Word 2.0--to serve as a sort of cranky focus group, we supposed. Curmudgeons, meet your new word processor. Finally, here was the major upgrade: Microsoft hoped it would keep WordPerfect from achieving a new dominance in the Windows market; we hoped it would rescue us from the bugs.
We loved it. We hated it. Winword had amazing new features. Not just the beautifully designed "mail-merge" feature--lone writers don't have much occasion to send customized mass mailings linked to thousand-entry databases. Not just the "grammar checker"--I already know my sentences are too long, thank you very much. There was an automatic envelope printer--Microsoft's research had revealed that some people still kept typewriters on hand for that purpose alone. There was a built-in drawing program, a built-in business chart program, a built-in equation editor. If I wanted to number a group of paragraphs, insert a table, zoom in on my text, or create side-by-side columns, I had merely to click my mouse on a pictorial icon at the top of my screen. No wonder my book was already a half-year late.
Yet there were signs that all was not well. A feature meant to compare two versions of a document still did not actually function; it was a shell, as one of the developers admitted privately, with little more purpose than to persuade the trade press to add one more "Yes" to the feature-comparison charts that always accompany word processor roundups. ("There were so many higher priority items that users requested, we couldn't squeeze it in," Microsoft says now, adding that the feature would "probably" work better in a later version.) The strange font problems seemed to remain--but perhaps it was too soon to tell. Another beta test was getting under way.
For the Winword team, the next months brought intense pressure. WordPerfect for Windows was rumored to be imminent. WordPerfect had run a long, large beta program. Microsoft's programmers had been able to monitor its progress and thought they could match it feature for feature. Still the original, non-Windows version had 9 million or 10 million users who would naturally be inclined to stay with a brand they knew. And meanwhile, another competitor, Lotus Development Corporation's Ami Pro, came out with a startlingly improved new version.
Winword 2.0's release slipped from September into October. The Gadflies, their ranks swelled with new volunteers, were shocked--shocked!--to discover that the developers were deliberately leaving bugs unrepaired: they were declared to be "features," or "by design." It was triage. In a candid moment on CompuServe, Microsoft programmers admitted the existence of an internal list called, evocatively, "Won't Fix."
Had twenty months of angst-ridden electronic messages been for naught? Guy Gallo begged for some assurance that "Won't Fix is more like Bug Purgatory than Bug Hell." At a Windows conference in August, one of the newer Gadflies, Ellen Nagler, a California consultant, handed out buttons: Won't Fix and It's Not a Bug--It's a Feature. The Winword developers asked for a set of the buttons. Emotion ran high.
"It IS only SOFTWARE," one of the developers typed in exasperation.
Finally, in November, the program reached stores, accompanied by an enormous wave of promotion that included, for the first time, commercials on network television. Reviewers raved. They loved the ability to drag and drop chunks of text with a mouse; they admired the new mail-merge and envelope-addressing features; and, sure enough, they fell into Microsoft's trap and gave Winword a "Yes" in the document-comparison check box. They barely mentioned the bugs. How could a reviewer on deadline be sure that any particular problem wasn't an example of another chronic and costly industry problem: the one known as User Error?