--------------------------------------------------------------------------------

Authors: Paul De Palma

Volume: 68

Issue: 1

Start Page: 61

ISSN: 00030937

Copyright American Scholar Winter 1999

Full Text:

In the misty past, before Bill Gates joined the company of the world's richest men, before the mass-marketed personal computer, before the metaphor of an information superhighway had been worn down to a cliche, I heard Roger Schank interviewed on National Public Radio. Then a computer science professor at Yale, Schank was already well known in artificial intelligence circles. Because those circles did not include me, a new programmer at Sperry Univac, I hadn't heard of him. Though I've forgotten the details of the conversation, I have never forgotten Schank's insistence that most people do not need to own computers.

That view, of course, has not prevailed. Either we own a personal computer and fret about upgrades, or we are scheming to own one and fret about the technical marvel yet to come that will render our purchase obsolete. Well, there are worse ways to spend money, I suppose. For all I know, even Schank owns a personal computer. They're fiendishly clever machines, after all, and they've helped keep the wolf from my door for a long time.

It is not the personal computer itself that I object to. What reasonable person would voluntarily go back to a typewriter? The mischief is not in the computer itself, but in the ideology that surrounds it. If we hope to employ computers for tasks more interesting than word processing, we must devote some attention to how they are actually being used, and beyond that to the remarkable grip that the idol of computing continues to exert.

A distressing aspect of the media attention paid to the glories of technology is the persistent misidentification of the computing sciences with microcomputer gadgetry. This manifests itself in many ways. Once my seatmate on a plane learns that I am a computer science professor, I'm expected to chat about the glories of the new DVD-ROM as opposed to the older CD-ROM drives, or about that home-shopping channel to the computer-literate, the World Wide Web, or about one of the thousand other dreary topics that fill PC Magazine and your daily paper, and that by and large represent computing to most Americans. On a somewhat more pernicious level, we in computer science must contend with the phenomenon of prospective employers who ask for expertise in this or that proprietary product. This has had the effect of skewing our mission in the eyes of students majoring in our field. I recently saw a student resume that listed skill with Harvard Graphics but neglected to mention course work in data communications. Another recent graduate in computer science insisted that the ability to write WordPerfect macros belonged on her resume.

This is a sorry state. How we got there deserves some consideration. few words of self-disclosure may be in order. What I have to say may strike some as churlish ingratitude to an industry that has provided me with a life of comparative ease for nearly two decades. The fact is that my career as a computer scientist was foisted upon me. When I discovered computers I was working on a doctorate in English at Berkeley and contemplating a life not of ease but of almost certain underemployment. The computer industry found me one morning on its doorstep, wrapped me in its generous embrace, and has cared for me ever since. I am paid well to puzzle out the charming intricacies of computer programs with bright, attentive students, all happy in the knowledge that their skills will be avidly sought out the day after graduation. I can go to sleep confident that were tenure to be abolished tomorrow, the industry would welcome me back like a prodigal son.

Yet for all its largesse, I fear the computer industry has never had my full loyalty. Neither did English studies, for that matter, but this probably says more about those drawn to the study of texts than about me. My memories of the time I spent in the company of the "best which has been thought and said" are hazy, perhaps because the study of literature is not so much a discipline as an attitude. The attitude that dominated all others when I was a student, that sustained my forays into the Western Americana of the Bancroft Collection, is that there is no text so dreary, so impoverished, so bereft of ideas that it does not cry out to be examined-deconstructed, as a graduate student a few years my junior might have said. But the text I now propose to examine, impelled, as it were, by early imprinting in the English department, goes beyond words on a page.

From an article here and a TV program there, from a thousand conversations on commuter trains and over lunch and dinner, from the desperate scrambling of local politicians after software companies, the notion that prosperity follows computing, like the rain that was once thought to follow the settler's plow, has become a fully formed mythology.

In his perceptive little book Technopoly, Neil Postman argues that all disciplines ought to be taught as if they were history. That way, students "can begin to understand, as they now do not, that knowledge is not a fixed thing but a stage in human development, with a past and a future." I wish I'd said that first. If all knowledge has a past-and computer technology is surely a special kind of knowledge-then all knowledge is contingent. The technical landscape is not an engineering necessity. It might be other than it is. Our prospective majors might come to us, as new mathematics or physics majors come to their professors, because of an especially inspiring high school teacher, because of a flair for symbol manipulation, or even because of a (dare I use the word?) curiosity about what constitutes the discipline and its objects of study-not simply because they like gadgets and there's a ton of money to be made in computing.

The misidentification of computer science with microcomputer gadgetry is a symptom of a problem that goes far beyond academe. Extraordinary assertions are being made about computers in general and microcomputers in particular. These assertions translate into claims on the American purse-either directly, or indirectly through the tax system. Every dollar our school districts spend on microcomputers is a dollar not spent reducing class size, buying books for the library, reinstating art programs, hiring school counselors, and so on. In fact, every dollar that each of us spends outfitting ourselves with the year's biggest, fastest microcomputer is a dollar we might have put away for retirement, saved for our children's education, spent touring the splendors of the American West, or even chosen not to earn. In the spirit of Neil Postman, then, I'd like to speculate about how the mythology of prosperity through computing has come to be and, in the process, suggest that like the Wizard of Oz, it may be less miraculous than it looks.

The place to begin is the spectacular spread of microcomputers themselves. By 1993 nearly a quarter of American households owned at least one. Four years later, the Wall Street Journal put this figure at over 40 percent. For a home appliance that costs at least a thousand dollars, probably closer to two, this represents a substantial outlay. The home market, as it turns out, is the smaller part of the story by far. The Census Bureau tells us that in 1995, the last year for which data are available, Americans spent almost $48 billion on small computers for their homes and businesses. This figure excludes software, peripherals, and services purchased after the new machines were installed.

The title of an article in the Economist-"Personal Computers: The End of Good Times?"-hints at the extraordinary world we are trying to understand. In it we learn that annual growth in the home computer market slowed from 40 percent in 1994 to between 15 percent and 20 percent in 1995. By the fall of 1998, market analysts were predicting 16 percent growth in the industry as a whole for the current year. Those of us involved in other sectors of the economy can only look on in astonishment. When a 20 percent, or even 16 percent, growth rate-well over five times that of the economy as a whole-is "the end of good times," we know we're in the presence of an industry whose expectations and promises have left the earth's gravitational pull.

To put some flesh on these numbers, let's try a thought experiment. The computer on my desk is about sixteen inches by seventeen inches. The Census Bureau tells us that the microcomputer industry delivered over eighteen million machines in 1994, the year when, according to the Economist, good times ended. Of these, perhaps a third went to the home market, the balance to business. At the 40 percent growth rate in the home market cited for that year and the more modest 16 percent growth rate for the business market, the boys in Redmond and Silicon Valley will have covered the United States's 3,679,192 square miles with discarded microcomputers well before my daughter, who is now thirteen, begins to collect Social Security.

Fabulous as they seem, these figures come from only part of the industry. Microcomputers do not define computing, despite their spectacular entry on the scene. The standard story goes like this. There was once a lumbering blue dinosaur called IBM that dominated the computer industry. In due course, smaller, more agile, and immensely more clever mammals appeared on the scene. The most agile and clever of these was Microsoft, which proceeded to expand its ecological niche and, in so doing, drove the feeble-minded IBM to the brink of extinction.

The business history in this story is as faulty as its paleontology. IBM may be lumbering and blue, but in 1997 its sales were nearly $78 billion. Compare that with Microsoft's $9 billion. The real story is not in the sales volumes of the two companies but in their profit margins. In 1997 IBM's was 7.7 percent, while Microsoft's was a spectacular 28.7 percent. This almost mythical earning capability is expressed best in Forbes' annual list of very rich Americans. We don't hear much about IBM billionaires these days, but Microsoft fortunes are conspicuous in the Forbes list, with Bill Gates's $51 billion, Paul Allen's $21 billion, and Steven Ballmer's $10.7 billion. These fortunes were accumulated in less than twenty years from manufacturing a product that requires no materials beyond the inexpensive medium it is stored on-not so different from a pickle producer whose only cost, after the first jar comes off the line, is the jar itself. It's a tale of alchemical transmutation if ever there was one. Is it really a surprise that most people don't know that IBM is still a very successful company or that computer science does not begin and end with Windows 98?

This joyous account of fortunes waiting to be made in the microcomputer industry has a dark side. Just as Satan is the real hero of Paradise Lost, as C. S. Lewis observed, so is popular fascination with computers due as much to the dark side as to the light. Despite generally good economic news for the past few years, Americans seem gloomy about their prospects. Our brave new world, paved over with networked computers from sea to shining sea, may well be one in which we are mostly unemployed or have experienced a serious decline in living standards. Computers, if not always at the center of the problem, are popularly thought to have been a major contributing factor.

Look at the substantial decline in manufacturing as a segment of the workforce in the United States. Between 1970 and 1996 (the last year for which data are available), the number of Americans employed increased by about fifty million. During this same period, the number of manufacturing jobs declined by about two hundred thousand. The culprit here is often thought to be computer technology, through assembly line robots or through U.S.-owned (or -contracted) manufacturing facilities in developing countries. Asia and Latin America, of course, would have less appeal to American corporations without worldwide data communications networks.

This analysis of the decline in manufacturing employment is perhaps more appealing than true. I will return to the relationship between computers and productivity. For now it's enough to observe that most people believe there is such a relationship. So if the money to be made in the computer industry is not sufficient inducement to vote for the next school bond issue that would outfit every classroom in your city with networked computers, then the poverty your children will certainly face without such a network should do the trick. With those staggering Microsoft fortunes in the background and the threat of corporate retrenchment in the foreground, I suppose I'm naive to expect the strangers I chat with on planes to know that the computing sciences are more like mathematics and the physical sciences than like desktop publishing-or, for that matter, like the rush to the Klondike goldfields.

The emergence of the microcomputer as a consumer item in the past decade and a half has prompted a flood of articles in the educational literature promoting what has come to be called "computer literacy." In its most basic sense, this term appears to refer to something like a passing familiarity with microcomputers and their commercial applications, rather like the ability to drive a car and know when to get the oil changed. Sadly, the proponents of computer literacy have won the high ground by virtue of the term itself. Who would argue with literacy? It is, after all, one of the more complex human achievements. Not only is literacy a shorthand measure of a country's economic development, but as the rhetorician Walter J. Ong has long argued, once a culture becomes generally literate, its modes of conceptualization are radically altered. Literacy-like motherhood and apple pie in the America of my youth-is unassailable.

But what about the transformative nature of literacy? I am fully aware that similar claims have been made about computers: namely, that computers, like writing, will alter our modes of conceptualization. Maybe so, but not just by running Microsoft Office. I've developed a rule of thumb about claims of this sort. If the subject matter is computers and the tense of the claim is future (and, therefore, its truth value cannot be ascertained), look at the subtext. Is the claimant a salesman in disguise? To recognize the nonsense in the claim that computers will transform the way we think, we need only indulge in some honest self-examination. I would give up my word processor with great reluctance. This doesn't mean that my neuronal structure is somehow fundamentally different from what it was when I was writing essays similar to this one on my manual Smith Corona. It does mean that the computer industry is a smidgen richer because of my contribution. It also means, as was recently pointed out to me, that it is a good bit easier to run on at great length on a computer than on a typewriter.

Not surprisingly, the number of articles addressing computer literacy in the educational literature has kept pace with microcomputer developments. ERIC is a database of titles published in education journals. When I searched ERIC using the keywords "computer literacy" and "computer-literate," I found 97 articles for the years 1966-81, or an average of about 6 per year. The decade from 1982 to 1991 produced 2,703 hits, or about 270 per year. At first glance the production of articles since 1991 shows welcome signs of dropping off. But the Internet has come to the rescue of both the microcomputer industry and its prognosticators. When I add the terms "Internet," "World Wide Web," and "information superhighway" to the mix (subtracting for duplicates), the total rises to an astonishing 4,680 articles from 1992 through the first half of 1998. This works out to about 720 articles per year. The bulk of the recent articles, of course, are full of blather about the so-called information superhighway and how all those school districts that cannot give every child access to it will be condemning the next generation to lives of poverty and ignorance.

Since computer literacy advocates are eloquent on the benefits of computers in our schools (and equally eloquent on the grim fate that awaits those students not so blessed), a brief look at how microcomputers are actually used in primary and secondary schools is in order. Microcomputers are now a solid presence in American education. The U.S. Census Bureau put the number at nearly seven million in 1997, or just over seven students per machine, compared with eleven students per machine in 1994 and sixty-three per machine a decade earlier. Picture a classroom richly endowed with computers. Several students are bent over a machine, eyes aglow with the discoveries unfolding on the screen. Perhaps there is a kindly teacher in the portrait, pointing to some complex relationship that the computer has helped the budding physicists, social scientists, or software engineers to uncover. If this is the way you imagine primary and secondary school students using computers, you are dead wrong. Several important studies have concluded that primary and secondary school students spend more time mastering the intricacies of word processing than they do using computers for the kinds of tasks that we have in mind when we vote for a bond issue.

Programming, in fact, was the one area that school computer coordinators saw decline over previous years. I would be the first to acknowledge that programming does not define computer science. This simple fact is what makes the endless discussion of programming languages in computer science circles so tedious. Nevertheless, if computer science does not begin and end with programming, neither will it give up its secrets to those who cannot program. I greet the news that high school students do not program our millions of microcomputers as an English professor might greet the news that the school library is terrific but the kids don't read. Here is a puzzle worth more than a moment's thought. There is an inverse relationship between the availability of microcomputers to primary and secondary school students and the chance that those students will do something substantial with them. I am not saying that the relationship is causal, but the association is there. Draw your own conclusions.

Though the jury is still out on the potential educational benefits of computing, we all agree that skill with computers is necessary for success in business. Even here there's a problem. Recent studies have assembled evidence that should give computer enthusiasts some sleepless nights. It appears that most businesses would be better off had they taken all that money they spent on computer technology and put it into bonds at market rates. This investment has been substantial, as anyone knows who has seen the piles of unopened software, the manuals still in their shrink-wrapped plastic, and the stacks of obsolete hardware accumulating in storerooms around the country. By 1995, it had totaled over $4 trillion. This sum, it should be noted, excludes the public money involved in training (and employing) academic computer scientists and engineers.

It also excludes another hidden expenditure. The time employees spend rearranging icons on their screens, the time they spend wondering why their spreadsheets will not recognize their printers, in fact all those minutes here and hours there spent fiddling with hardware and software, is time they do not spend on the tasks they are being paid to perform.

Let me tell a story. I have been a computer science professor for seven years. Before that I spent a decade working for some of the largest firms in the computer industry. I am, by any reasonable measure, computerliterate. One recent Sunday afternoon I thought I might pop into my office, copy this essay to a floppy disk, and work on it at home, where I was also caring for a child with chicken pox. In other words, I took the microcomputer industry up on its central promise: workers will be liberated from the tyranny of place. Able to function as parents and workers simultaneously, we will prosper along with our employers.

Here's what really happened. I promised my wife I'd be gone no more than thirty-five minutes, twenty-five for the drive to and from the university, ten to copy the file. As it happens, I am the last remaining member of the professional middle class without Microsoft Windows running at home. What I have is a 286 IBM clone running DOS and WordPerfect 5.1, equipment my last employer gave me when I left seven years ago. This setup was well on its way toward obsolescence when I acquired it. Were my students to learn that I wrote with a quill pen by the light of an oil lamp, they would think me hardly less quaint.

However, my reluctance to part with hard-earned money for a shiny new computer that I would use only as an abundantly outfitted typewriter did pose a small challenge. I would have to get the file from my office computer, a fancy Pentium workstation (courtesy of my current employer), to run on my ever-faithful home machine. Not a problem, I thought. I can easily transform the Microsoft Word file in my office to ASCII text, copy it to a 5.25-inch floppy disk, and read it into WordPerfect at home.

As we have all come to know, painfully at first and finally with resignation, when the subject is personal computers, things are not always as promised. (It has occurred to me more than once that the computer industry should have the honor of Iago and display these words boldly on every screen: "I am not what I am.") First I learned that my document was infected with the Word macro virus. No matter how I tried, Word would not let me transform it from a template (a term known to all Word users, happily ignored by most) to a text file. So I called a colleague who gave me what is known in the computer industry as a "workaround." A workaround is what you do when your machine is not running the way the manufacturer promised. By analogy, a workaround for faulty automobile brakes might be to open the door and drag your feet. In any case, my colleague is a clever fellow and the workaround did allow me to work around the handiwork of the disgruntled Microsoft employee who infected Word with the virus. Having transformed my essay into a generic text file, I was ready to copy it to a floppy and return home, safe in the knowledge that I could be both productive and parental.

Unfortunately, our former systems administrator, for reasons that must have made eminent sense to him, had disabled my A drive. But as I said earlier, I am computer-literate. Though annoying, a disabled drive is not catastrophic. I need only invoke a special setup routine to let Windows know that, in fact, there is a 5.25-inch floppy drive on my machine. But since this machine is a castoff from our department's lab, the systems administrator had, wisely, password-protected the setup routine. He had also, in the meantime, decamped for the vastly more remunerative pastures of the computer industry. In a word, he was unavailable and so was my machine. I returned home, nearly two hours after I had left, to an unhappy wife and a sicker child-and without the file.

This is not an isolated story. Anyone who has dealt with a microcomputer has a store of similar tales. There is another story here as well. Even if one is inclined to stick with the tried and true, the computer industry-and its minions across the land-will not permit it. By the time this article goes to press, the last computer in my department with a drive that accommodates 5.25-inch diskettes will have gone to wherever old computers go. My well-worn and well-loved 286 will then be an island cut off from the main, and I, its single inhabitant, will speak a language that is fast becoming extinct.

The price of computing equipment has dropped dramatically in recent years. For under $2,000 you can buy a microcomputer that processes millions of instructions per second and is equipped with a stunningly large memory and disk space. At that price we can all be equipped at the office, and most of us will choose to be equipped at home. As it happens, that $2,000 (plus a bit more for networking components) is the smallest slice of the great pie of microcomputer costs. The Economist cites a study by the Gartner Group, a respected consulting firm, that puts the annual cost of a microcomputer connected to a network at $13,200. Of this, only 21 percent goes to the purchase of hardware and software. Administrative costs absorb 36 percent. We have to pay all those people who come to our rescue, after all. This figure alone should slow down the headlong rush to outfit every desk on the planet with a microcomputer.

That administration costs more than the machine itself is not the biggest surprise. Recall my story. Just how much was my two hours worth? On the average, 43 percent of the cost of a microcomputer is consumed in what Gartner calls "end-user operations." Just what are end-user operations? They are all those things one does with a computer in order to do the things one gets paid to do. This includes rearranging icons, coaxing disk drives into action, loading and setting up software, avoiding viruses, listening to Microsoft's music as you wait helplessly on hold for advice from someone in technical support who probably knows less than you do, and so on.

Though the Gartner Group has done us the service of quantifying those long hours spent mastering yet another Microsoft user interface, the effect of that time on worker productivity has been known for some years now. Many studies, including some done by the National Research Council and by Morgan Stanley, the New York investment bank, fail to indicate any correlation between productivity growth and information technology expenditures. Distressingly, the opposite appears to be true. As Thomas Landauer has pointed out in The Trouble with Computers, those industries that invested most heavily in information technology, with the exception of communications, seem to have the most sluggish productivity growth rates. Though one still could argue that schools and colleges should continue to teach courses in microcomputer literacy because microcomputer usage has grown like a fungus after a heavy rain, our time might be more profitably spent breaking the bad news to the public that pays the bills. In the process, we might also come to understand how a machine so patently clever as the microcomputer could have done the business world (outside of the computer industry itself) so little good.

Given the several thousand articles on computer literacy and the emerging inverse relationship between productivity growth and computer expenditures, it seems reasonable to ask just who does benefit from the computer literacy movement-and who pays for it. The commonsense answer is, "Students benefit." Well, common sense is right, but, as usual, only partially so. Students, of course, are served by learning how to use microcomputers. But the main beneficiaries are the major producers of hardware and software. The situation is really quite extraordinary. Schools and colleges across the country are offering academic credit to students who master the basics of sophisticated consumer products. Granted that it is more difficult to master Microsoft Office than it is to learn to use a VCR or a toaster oven, the difference is one of degree, not of kind.

The obvious question is why the computer industry itself does not train its customers. The answer is that it doesn't have to. Schools, at great public expense, provide this service to the computer industry free of charge. Not only do the educational institutions provide the trainers and the setting for the training, they actually purchase the products on which the students are to be trained from the corporations that are the primary beneficiaries of that training. The story is an old but generally unrecognized one in the United States: the costs are socialized while the benefits are privatized.

I have described a bleak landscape in this essay. Let me summarize my observations:

Schools and universities purchase products from the computer industry to offer training that benefits the computer industry.

These purchases are both publicly subsidized through tax support and paid for by students (and their parents) themselves.

The skill imparted is at best trivial and does not require faculty with advanced degrees in computer science-degrees acquired, by and large, through public, not computer industry, support.

As the number of microcomputers in our schools has grown, the chance that something interesting might be done with them has decreased. The stunning complexity of microcomputer hardware and software has had the disastrous effect of transforming every English professor, every secretary, every engineer, every manager into a computer systems technician.

For all the public subsidies involved in the computer literacy movement, the evidence that microcomputers have made good on their central promise-increased productivity-is, at the very least, open to question. If my argument is at least partially correct, we should begin to rethink computing. The microcomputer industry has been with us for a decade and a half. We have poured staggering sums down its insatiable maw. It is time to face an unpleasant fact: the so-called microcomputer revolution has cost much more than it has returned. One problem is that microcomputers are vastly more complex than the tasks ordinarily asked of them. To write a report on a machine with a Pentium II processor, sixtyfour megabytes of memory, and an eight-gigabyte hard disk is like leasing the space shuttle to fly from New York to Boston to catch a Celtics game. Though there are those who wouldn't hesitate to do such a thing if they could afford it (or get it subsidized, which is more to the point), we follow their lead at great peril. The computer industry itself is beginning to recognize the foolishness of placing such computing power on every office worker's desk. Oracle, the world's premier manufacturer of database management systems; Sun Microsystems, a maker of powerful and highly respected engineering workstations; and IBM itself are arguing that a substantially scaled-down network computer, costing under $1,000, would serve corporate users better than the monsters necessary to run Microsoft's products.

Please don't misunderstand. This is not a neo-Luddite plea to toss computers out the window. I am, after all, a computer science professor, and I am certainly not ready (as the militias in my part of the country put it) to get off the grid. Further, the social benefits of computing-from telecommunications to business transactions to medicine to science-are well known. This essay is simply a plea to think reasonably about these machines, to recognize the hucksterism in the hysterical cries for computer literacy, to steel ourselves against the urge to keep throwing money at Redmond and Silicon Valley.

Putting microcomputers in their place will also have a salutary effect on my discipline. We in computer science could then begin to claim that our field-like mathematics, like English literature, like philosophy-is a marvelous human creation whose study is its own reward. To study computer science calls for concentration, discipline, even some amount of deferred gratification, but it requires neither Windows 98, nor a 400megahertz Pentium II processor, nor a graphical Web browser. Though I am tempted, I will not go so far as to say that the introductory study of computer science requires no computing equipment at all (though Alan Turing did do some pretty impressive work without a microcomputer budget). We do seem, however, to have confused the violin with the concerto, the pencil with the theorem, and the dancer with the dance.

I am afraid that we in computing have made a Faustian bargain. In exchange for riches, we are condemned to a lifetime of conversations about the World Wide Web. An eternity in hell with Dr. Faustus, suffering the torments of demons, would be an afternoon in the park by comparison.

Paul De Palma is Associate Professor of Mathematics and Computer Science at Gonzaga University.

Reproduced with permission of the copyright owner. Further reproduction or distribution is prohibited without permission.