From the WSJ Opinion Archives
BITS AND BYTES
How to Build a Better PC
IBM throws in the towel. That's a mistake.
IBM has just sold its PC operation to a Chinese company called Lenovo for (a relatively paltry) $1.25 billion. If I were an IBM stockholder who cared mostly about short-term profits, I'd be thrilled. If I were an IBM board member, or anyone who cared about the long-term health of IBM or the U.S. technology industry or the whole blooming U.S. economy, I'd be unspeakably depressed.
IBM invented the PC as we know it in 1981. But today PCs are a commodity, like toothpicks or telephones. It's hard to make money in the PC business, and IBM wants to concentrate on more profitable lines; wants to focus its efforts on software, servers and services. Which sounds like a good business strategy, at any rate for the short term. IBM's server business has been a consistent profit leader for the company. The PC business, on the other hand, has evidently tended, in recent years, to decrease the company's overall growth rate and pretax margin.
IBM's move speaks loudly about personal computer technology. So far as IBM can figure, PCs have reached a plateau. IBM can't think of any practical way to sell its PCs for significantly more money than other companies charge. And IBM is no random group of bums off the street. Once upon a time it was the most powerful force in the technology world; it still employs some of the smartest people in the field. If IBM has consigned the personal computer to Commodity Limbo, the prognosis is bad. Which is a shame, even a tragedy--because the modern PC is in fact a primitive, infuriating nuisance. If the U.S. technology industry actually believes that the PC has grown up and settled down, it is out of touch with reality--and the consequences could be dangerous to America's economic health.
PCs are roughly a quarter-century old. People who think of them as mature commodities might have thought the same thing about television in the 1970s--when TV was in fact on the brink of all sorts of revolutions. The airplane was 25 years old in the late 1920s; luckily, airplane companies kept inventing, developing and selling new types. The automobile turned a quarter-century old in the early '20s--and Henry Ford did consign it to Commodity Limbo. He figured that the Model T was grown-up, settled-down, fully-evolved. He almost wrecked his business, but finally got the message and produced the Model A and a long line of subsequent new designs. Obviously there are big differences between the PC and these other technologies. But there is also a big similarity: all were (or are) destined to take a lot longer than 25 years to reach maturity.
What's wrong with today's PC? Plenty. All sorts of functions that ought to be built-in are available only as add-ons or not at all.
Like many people, I have several PCs in my life--and I constantly need to ask such ridiculous questions as, "Where did I leave the latest version of that file? By what clumsy method should I move it from where it is to where it's needed?" Such questions are like asking "Where did I leave the starter crank for my Huppmobile?" If you have to ask, your (formerly) hot-shot machine is ready for the folk-art museum.
IBM might have done well selling PCs with built-in "transparent information sharing." As soon as you connected such a machine to the Internet, all your electronic documents would immediately be available--no matter where you created or last worked on them. If all your computers had transparent information-sharing, you could start composing an e-mail at work, touch it up during your drive home (using a--theoretical--in-car, audio-interface IBM PC) and finish it up on a laptop in your backyard. Lots of businesses and people would have shelled out for such PCs.
Many old and decrepit PCs would be replaced tomorrow if bringing new PCs up to speed weren't such a colossal nuisance. IBM PCs with transparent information-sharing would have made that problem disappear. Connect a new machine to the Internet and all your electronic information would have materialized automatically.
Many computer users are overwhelmed by e-mail. Whenever you start work on a computer, you ought to find a one-page e-mail summary ready and waiting. It would tell you at a glance (even if you haven't touched a computer in weeks) which new e-mails look important, which look like junk, and which have been acknowledged but not yet answered. (To acknowledge an e-mail is to send a one-line "I got it" message.) IBM might have offered a PC with "e-mail summary" built-in, along with a new key--press it and you cause the current message to be acknowledged; the computer updates its reminder list accordingly. Warning: This is not an e-mail cure. But it would help manage the pain.
There are dozens more possibilities. Why should anyone waste time throwing out e-mail (or any electronic document) when data storage is dirt cheap? Why are we wedded to a windows-menus-mouse interface that is flat, as if it were stuck to the back of the screen, when computers are easily powerful enough to turn the screen into a viewport that lets us "peer through it" into an imaginary 3-D landscape? (Information can be more clearly and effectively arranged in a 3-D space than on a restricted flat surface.) Large-screen and projection technology is cheaper all the time; why aren't large-screen computers (and living-room computers) a growing (high profit!) segment of the industry? Why doesn't every computer I use show me the exact same desktop, with the same layout of the same icons?--or (at any rate) the same picture, no matter what interface I use? I could go on.
Know this for sure: Some company will build all this and more into a radically more powerful, radically simpler PC. Will it be an American company? Don't count on it.
Mr. Gelernter is professor of computer science at Yale and senior fellow at Scientific Computing Associates in New Haven.
