11. A related ingredient, good software—the topic of earlier chapters.

The Canary in the Mine “How,” I asked a consultant, “do you know if someone is trying to offer a constructive suggestion about computerization or to get in the way?” “Consider the source,” he said. “Is it coming from a researcher or a union?” To some extent the man was right. Some people, unionized or not, may be no more than twentieth-century version of the Luddites, the English workers who, during the 1800s, smashed the mechanical looms that were taking away their livelihoods. Indeed, Dr. Michael Colligan, a NIOSH psychologist, says, “Opposition to computers is usually kind of veiled or disguised under more general complaints about eyestrain, headaches, and physical discomforts often associated with video display terminals (VDTs).[34] “But,” Colligan warns, “if you talk with workers awhile, major concerns begin to surface.” Legitimate concerns. Whether it’s individual workers or unions, don’t tune out gripers. Listen to people like Joan. Footnote 34: The Colligan quotes are from Joel Makower’s useful book _Office Hazards: How Your Job Can Make You Sick_, published 1981 by Tilden Press, Washington, D.C. Joan started out around 1980 as a model employee in a northeastern office of a major insurance company, whose officials declined to be interviewed. “I averaged ten hours a week overtime,” she said, “and I came in when I had pneumonia and a fever of 102.” In a modest way, helping to unravel the mysteries of the machines, she aided computerization. Later, however, in a not-so-modest way, she aided something else: unionization. By the time I reached her, she had quit her $10,000-a-year job to go into business with her husband. For more than an hour she poured out her anger against the terminals and the company that now used them. “They affected my stomach quite a bit,” Joan said. “Most days I worked, I would throw up in the ladies room. I didn’t throw up before I started working on the computer, and I haven’t thrown up since I quit. “The screens weren’t at all soft on the eyes, even with the glare screens that snapped over the top. And they were nontiltable.” She also complained that the screens were too low even for a five footer like her. “Could I have some books and put my system on top of them so the screen will be in the middle of my eyes?” she recalls asking a supervisor. “You get this constant tension in your back because you’re always hunched over.” “No,” she says the supervisor told her in effect, “that’s unacceptable, totally.” Why? “Because,” came the reply, allegedly, “the office won’t look the same. It won’t be level.” Joan says management wouldn’t even let her work with the terminal in the middle of her desk. “They wanted them all on the same side,” she recalls. “I said, ‘Hey, I don’t work on that side of the desk, sport.’ It was ludicrous, absolutely ludicrous. Even a simple request like that was too much.” At least the chair was adjustable “I lowered it so I could get my eyes as close to the middle of the screen as possible,” she says. “I was so low that if I opened the middle drawer on my desk I’d bang myself.” But it was the lights that most irked Joan. They were bright fluorescents, fine for traditional paperwork but not for the computer era. Management, pressed by employees, took out many of the tubes; and yet Joan says some glare-ridden workers still wore sunglasses. She also complains that the company wouldn’t even offer the operators eye examinations, an irony, considering that she was processing health claims. Again, Joan was a union activist—as a result of her experiences—and some managers may recoil from the very notion of clerks organizing, and yet in her litany of complaints about working conditions, there was not one with which most ergonomics professionals would quarrel. Had management listened to her on those issues, in fact, it would have been doing itself a favor. Forcing the workers to strain their necks, backs, and eyes is hardly the road to higher productivity and profit. It would have cost several thousand dollars at the very most to change the lighting. Corporate bureaucracy may or may not have allowed the workers to be consulted in selection of the terminals, but there is no reason—other than a misdirected passion for discipline and order—why the company couldn’t have let Joan change the level of her screen. Joan’s experiences are a good argument for the Canary-in-the-Mine Theory of Labor Relations. In the days before electronic gas detectors, coal miners carried canaries into the pits. The birds’ lungs were more sensitive than theirs to poisonous fumes. And watching the canaries, many a miner saved himself and others. Novelist Kurt Vonnegut, Jr., used the canary-in-the-mine parallel to describe artists and writers serving as an early-warning system to call attention to social ills. And workers like Joan, “troublemakers,” are your canaries. You may not agree with them, and you may aggressively fight them, but before making up your mind, at least listen to them. They can anticipate the concerns of less vocal workers scared of losing their jobs. “Well,” you ask,“what about my authority as a manager? You knuckle under to the troublemakers once, they’ll come at you again and again.” Confident executives, however, will avoid this misplaced machismo. They will assert their authority more productively, letting workers have a voice in working conditions but, in return, demanding good, steady production, which, after all, is the bottom line. When your stockholders receive dividends, how many will think, What slobs—their VDTs don’t even line up at the same level! Ideally, of course, your people won’t have to prop up their machines at all, because they’ll have helped you choose easily adjustable screens and other equipment. “When I help design new computer systems,” says Robert Waters, an ergonomics specialist with a background in psychology, “I try to find out who’ll be using them. Then, if possible, I’ll seek out their opinions. “If you do that, people won’t feel so threatened. It will be their equipment, too, not just the company’s. They won’t think, I don’t like it, it’s bad for me.” Richard Koffler, editor-publisher of the _Ergonomics Newsletter_, also recommends giving the worker a voice in equipment selection: “If you can do it, try it. They’ll tell themselves, ‘I wouldn’t want to look like a fool because I chose something that wasn’t good.’” In Washington, D.C., Richard G. Barry, now senior advisor for management systems at the World Bank, has successfully let employees help pick some of the computer gear the bank was using. Barry didn’t set out to experiment. He just decided that giving the workers a strong voice would be the most logical way. It happened, among other places, in the offices of the Eastern African region at the bank’s D.C. headquarters. “We had a few word-processing machines,” Barry said, “and there were a few word-processing centers around the bank where we could go if we wanted something done, but essentially we were starting out from zero.” The office included two hundred professionals and one hundred secretaries and other clerical workers, and Barry’s task force represented them in several ways. One was by job function; lending officers were there as well as those overseeing the existing projects. People also came from different parts of the office’s bureaucracy, from different countries, and, obviously, from different job levels—everyone from secretaries to division chiefs. Barry didn’t go out of his way to pamper clerical workers in particular. He simply felt that like any other bank employees, they should decide what tools they needed to get the job done. Secretaries, far from being anticomputer, asked for equipment to reduce repetitive typing. Managers sought fast word processing, in different languages, which complicated keyboard requirements. The secretaries themselves wanted terminals with whose keyboards and screens they’d be happy. In a microcomputer installation without separate terminals, of course, there may be less choice. It took a year for the task force to line up the equipment, and wisely the actual equipment shopping came last. Instead, the task force focused on what characteristics it wanted the gear to have. Then it learned what was available—how products compared to the ideal system within budget. As a result, the Eastern African region ended up with good, sharp computer screens, ergonomic furniture from the very start, and a lighting system fit for computer work. And Barry says the teamwork helped promote ... ... Good Job Design A friend of mine, a mid-level manager, was almost salivating over the economies of computerized word processing. “You know what I’d do?” he said of an office where he’d worked. “I’d take out just about all the secretaries and dump them into a central office. I’d reduce their salaries. They’d be doing less demanding work and should be paid less.” He wasn’t an ogre. He was reflecting the thinking of a good many modern managers. It makes sense, doesn’t it, turning offices into factories, which, indeed, might run twenty-four hours a day, to make the best use of equipment? Well, not quite. In fact, not at all. Nor should you try obnoxious gimmicks like =computerized pacing=. They’re all threats to good =job design=, which is nothing more than sensibly deciding what each job includes and how the employee does it. Consider typing pools—er, centralized word-processing operations. If not passé, they at least deserve to be; most, anyway. Employees’ salaries keep rising, even with slowed-down inflation, while the price of computers is falling. By the end of this decade an average worker’s salary may exceed $20,000 annually. And yet micros and dedicated word processors will easily cost less than several thousand dollars. Even now, for a typical business, Barry says, centralized word processing simply doesn’t pay; it tears the employee away from his or her boss and harms the managers’ own work. Executives must vie with other offices of a company for the word processors’ time. They can’t effortlessly set their own priorities. And quality is more of a problem, too, because, as Barry notes, “someone in centralized word processing won’t normally get hell for not catching an error. I expect my secretary to challenge anything she’s not certain about. For instance, my secretary will come in and say, ‘You left such-and-such person off the list of people receiving the memo.’ Which matters. Often who a memo goes to matters as much as what’s in the memo.” And what about the question of knowing which calls to put through to the boss? Around the time Barry and I talked, he was obsessed with a paper projecting the World Bank’s office needs for the next twenty-five years. He didn’t have to tell his secretary of the importance of any call about the paper. She knew. For she’d been helping him put it together. “It isn’t a question anymore of the efficient use of the machines,” Barry says in recommending the traditional secretary over the pool arrangement. “It’s effective use of the people. It’s like the phone. Even if people use the phone just twice a day for business, we still give them their own. We don’t put all the phones in one room, do we?” You may disagree with Barry and still insist that central word processing is the most efficient use of both equipment and people; and maybe your company’s size, economics, or policy will allow nothing else. But if so, be prepared for more job complaints. Shut off from the outside world, perhaps in a windowless, poorly ventilated room, the pool members may feel like Dickensian factory hands. And the claustrophobia may not just be physical. With no one around but others in drudge jobs and their immediate supervisors, the workers feel cut off from chances for advancement. They become obsessed with their immediate surroundings and may be more apt to notice shortcomings of their equipment. “You look at studies of typing pools,” says Waters, “and you’ll find they may actually get in the way of productivity. More workers hate their jobs, and they’re more likely to dislike their equipment, feel backaches, other pains.” He says that sick leave, deservedly or not, may increase. A gray-haired word processor in Manhattan said she disliked even “clusters,” which were smaller and presumably more pleasant than a giant typing pool in a back room. She missed the “pretty cozy relationship” she enjoyed as secretary to one executive.[35] Footnote 35: The material on the Manhattan word processor comes from Barbara Garson’s article in _Mother Jones magazine_, July 1981, p. 32. “He might send me uptown to return a blouse that didn’t fit his wife. So he couldn’t very well tell me not to make personal calls. Not that I stayed on the phone all day. My job was to get his work out, which included staying late if I had to. He appreciated me. And I expected to stay with him till he retired.” Responding to consultants’ recommendations, however, her company one day offered her three choices. She could work as a mere administrative assistant, join a word-processing operation, or quit. So she forsook the familiar typewriter for a Wang. “Ah,” she reflected, “when I think of those years of correction tape, white-out, retyping!” Then, continuing her remarks to writer Barbara Garson, she added: “I want you to take this down: I love my Wang. It’s not the machine that gave me the shaft.” What, then, did? “Straight typing, seven hours a day,” said the word processor. “That’s the shaft.” Like Barry, she criticized the quality of work from word-processing pools—or the lack of it. “The worst thing,” she said, “is when things come by me with wrong spelling, wrong information. Sometimes I think I should correct it. I did at first. But that’s not how they judge me.” It’s by the number of keystrokes. “So why should I take time to correct their work? And why should I stay two minutes past lunch if they’re timing me that way?” Michael Smith, however, makes a good case for careful, sensitive monitoring. “Management would be foolish,” he said, “if they didn’t monitor performance. It’s what you do with the monitoring that counts. If you use it in a negative way, that hurts performance. It might improve it for a minute or a half hour while you were watching them, but it might hurt in the long run.” He recommends, instead, systems that give the workers themselves “direct, immediate feedback. The faster the better.” Joan’s old insurance company, shortly after she left in 1982, started a monitoring system in offices across the country. Some of her ex-colleagues found it more palatable than a simple keystroke count. Every six months it adjusted their job levels and salaries. And production—the number of health claims processed—was just 30 percent in the formula. Accuracy was another 30. The remainder of the formula plugged in attendance and supervisors’ observations on job knowledge and the quality of correspondence. The employees weren’t just at the mercy of a machine. “It’s a more equitable system,” said a claims processor whose salary leaped from $220 to more than $280 a week. Still, the union claimed that most members weren’t so well disposed to the new system. That may or may not have been true. Just the same, the claims processor qualified her own enthusiasm. She wanted faster feedback from the monitoring. “It would be constructive,” she said, “if we got it at the end of the day or even weekly.” The woman, too, hoped to know the exact math behind the salary system: “I’ve asked my manager a number of times. She said, ‘You wouldn’t understand it.’ I said, ‘Tell it to me, and I will understand it.‘” Big questions persisted. Was the insurance company, for instance, really crediting her extra for writing letters to doctors who haven’t submitted diagnoses with claims? The company told her that it was—that it still paid to handle such cases. But she wasn’t so sure. Management, she claimed, had lied on other matters. Another worker said the company did not audit enough forms to discourage slipshod work from high producers. Even in its first few weeks, with the normal start-up wrinkles, the system did show some promise. Production was dramatically up. “If I was a manager and my job was to get the maximum amount of work out of an employee,” said a claims processor, a union member, “yeah—in that position I’d use it.” But she worried, justifiably, about her own future. “Maybe a person has been producing ninety claims a day. How much more can they push them? They can’t push them past a hundred, so what else is there for them to do? “There aren’t many [promotion] opportunities here because supervisors have been here for years and years,” she said. And she wondered: What about turnover? It was winter 1982. A depression or near depression was in full swing. Would the system have worked well in normal times? Moreover, how much of a trade-off existed between the claims processors’ productivity and health? NIOSH studied 130 occupations in the 1970s and found that clerical workers were second highest in illnesses like coronary heart disease. Also, many of the women in Joan’s former insurance office had children and blue-collar husbands, and a Massachusetts study showed that women in such families suffered nearly double the heart-disease rate of their husbands. And in Joan’s office, did health insurance claims—for her coworkers’ own ills—cancel out some of the productivity increases? Moreover, how about the moral issue? Suppose the monitoring system taxes the employees’ health for short-term gains; isn’t it almost the same as if the company were belching carcinogens into the air? If you’re working in a low-level clerical job without control over your fate, you’re a good heart-attack candidate, according to R. A. Karasek, a professor at the Department of Industrial Engineering and Operations Research at Columbia University.[36] He cites studies showing that people in “strain occupations” may show twice as much “definite and suspected myocardial infarction and angina pectoris.” Footnote 36: See “Job Decision Latitude, Job Design, and Coronary Heart Disease,” by R. A. Karasek, who participated in a 1981 Purdue University conference on job stress. The Karasek findings appeared on pages 48-55 of _Machine Pacing and Occupation Stress_, a book published in 1981 by Taylor & Francis Limited, London. In Karasek’s words, “Job-design strategies advocating limited skill usage and decision authority for the majority of the workforce appear to be associated with a host of undesirable, unintended consequences ranging from skill under-utilization (and consequent productivity loss) to increased risk of coronary heart disease.” Computerized pacing can increase the stress and health risks. Smith says it’s rarely if ever the answer in white-collar work, not even mail sorting—nothing “involving letters, letters, and so on. It doesn’t improve efficiency because there are a lot of errors.” Pacing isn’t just monitoring, after all. The work is electronically moving past the worker as if it’s on a factory conveyer belt. Think of all the lemons Detroit put on the road in its eagerness for new production records. “With older workers the error rate isn’t as high,” Smith says of white-collar jobs paced by machines. “Maybe it has to do with the nervousness of young people. Older workers always report more job satisfaction. They become used to the drudgery. It’s a hell of a way to put it, but that’s what happens.” Not that many older workers love machine pacing, either. So, in a large, factorylike office, what’s a happy compromise between lax discipline and mechanized martinets? You might try =participatory monitoring=. Hear employees’ suggestions for a fair monitoring system. You needn’t agree. Just listen. Your people may know of complications you wouldn’t consider in arriving at your production goals; and you may also get a better inkling of how long it will take for people and machines to adjust to each other. Also, setting goals, keep remembering that tasks vary. Don’t let your monitoring system penalize people, for instance, who fill out forms more complicated than other workers‘. Keep your commitment to quality. Be willing to give employees some time away from arduous work at the tube. It’s a good way to reduce expensive errors. And tube breaks might not cost you that much in the end. Can you, for instance, design your people’s jobs to use them fully while limiting their time in front of the screens? Maybe you can’t. Perhaps, with thorough computerization, there aren’t many off-line jobs left. But try. Maybe, for instance, some data-entry clerks, showing high motivation, can work part time in low-level telephone sales. The clerks won’t feel so trapped; and you may discover some top-flight talent. Ask your employees for their ideas. Who says every tube break absolutely has to be a _work_ break? Plan, however, for breaks of one kind or another. NIOSH favors breaks “of at least fifteen minutes every two hours” for moderately heavy terminal work and the same breaks every hour for workers in the most demanding, repetitive tasks. A British labor group even suggests structuring the workday so that people spend no more than half of it at the terminal. You might chafe under those restrictions—many American companies would—but don’t scoff at your own people’s ideas on breaks as long as they do the work. Here again, think about the Canary-in-the-Mine Theory of Labor Relations; do not tune out the complainers: do not misplace machismo. Terminal Happiness In the Dark Ages, the pre-VDT days of newspapering, I worked on a rickety manual typewriter. How I envied Darlene at the desk behind me! I was a reporter-feature writer, while she was stuck with grinding out TV listings and obits; but Irving Leibowitz, the editor, had favored her with a Selectric, and I demanded to know why. “Well,” Leibo said, “she’s a neater typist.” It mattered. Darlene would feed her Selectric copy to an =optical character reader=, which helped convert the typing to newspaper print. “There’s another reason, too,” he said. “What?” “Darlene has a lousier job than you do,” said Leibo, himself a manual typewriter user, “and just as much typing to do. In fact, more.” Budgeting for VDTs, you might keep Leibo’s wisdom in mind. He gave out the best equipment not to his higher-ranking people but to those who needed it the most. I thought of Leibo and Darlene when I read of the Grid Compass executive computer and its original $10,000 price. It might be a splendid machine, but what a waste of money in many cases. The money instead might go to buy the right screens and keyboards for subordinates. So often the difference is just a few hundred dollars, if it exists at all. Don’t give your employees a say in the selection of equipment and then restrict their choices through unreasonable penny-pinching. If you can’t do it the right way immediately, maybe you should wait before you computerize. Mind you, no terminal or micro is going to be ergonomically perfect. “I’d flunk them all,” said Waters, and he started with the DEC VT100 terminal he was using on his job at the time. The display was plain, old white on black, not the best ergonomically, and it was too dim for many offices. Fortunately, the lighting in his room was subdued. The terminal, however, had other shortcomings—for instance, the lack of simple knobs to adjust the screen’s brightness and contrast. Waters instead had to control them with a series of keystrokes that he was always forgetting. His loudest groans, though, were over the numbers pad to the right of the keyboard. “I count 1-2-3-4-5-6-7-8-9,” Waters said. “I don’t want anything where I count 7-8-9-4-5-6-1-2-3.” That’s how his numbers pad was; that’s how my machine’s is; that’s in fact how most computer pads are. “Well,” I asked, “isn’t that just like a calculator’s?” “But,” said Waters, “isn’t it a little confusing to switch back and forth between that arrangement and the numbers of a telephone? Which are in the normal numerical order.”[37] Footnote 37: Before Bell adopted the standard touch-phone numbers pad, it did a study showing the superiority of the 1-2-3-4-5-6-7-8-9 arrangement. The study had strong financial motivations. After all, the more wrong numbers people dialed, the greater would be the cost to the phone company, since Bell at the time probably wasn’t charging anyone for wrong numbers. You may disagree with Waters, but his message comes through. Even ergonomics experts can’t always end up in front of the terminal of their dreams. What you and your people want, however, is a sensible compromise. Ideally, rather than simply seeking the biggest discount through the most massive purchase, you’ll remember that different people need different terminals. You may in fact save money that way. An executive who hates to type, for instance, and who doesn’t have to, won’t need the same keyboard as a data-entry clerk. It still should be a good board in case he changes his mind, but it needn’t be the most expensive. With a World Bank-style selection committee, you’ll be more sensitive to the needs of different offices, different levels of employee. Encourage some committee members to spend at least two or three hours with the equipment you’re planning to buy; they may change their minds later—but that’s better than no tryout at all. Testing my Kaypro at dealers, I wrote test letters, composed sample articles, and tried as closely as I could to duplicate my routine. Impressed by the low price of the Osborne 1, I gave the little keyboard every benefit of the doubt. And yet, even after several hours, I simply could not adjust. “Just a matter of operator retraining” was how one saleswoman put it after I complained that I kept hitting the return when I wanted a quote mark. And yet I wasn’t about to gamble $1,800 on a machine with a keyboard I might forever hate. Ergonomically, I had an advantage: I was buying just for myself. But a selection committee, made up of different kinds of prospective users—clerks and executives alike—is the next best thing to a lone customer shopping with only himself in mind. Avoiding the HAL syndrome, your committee should consider, among other things, the following: THE SCREEN For heavy-duty viewing, a terminal or computer should have at least a nine-inch screen and ideally a twelve incher. I emphasize the words “heavy-duty.” An executive using a terminal half an hour a week obviously has needs different from a clerk doing tedious data-entry work most of the day. Even the executive, however, should have a screen big enough for unexpectedly long sessions in front of the tube. On the other hand, suppose a bank teller must view just a few columns of numbers every now and then; he may never need a screen bigger than a few inches. Too large a screen, in fact, even in heavy-duty work, may expose one to glary reflections. Size preferences can be quirky. I’ll make do for the moment with the Kaypro’s nine inches, but I wouldn’t mind a bigger screen; another writer once said he’d like to hook up the same model to a twenty-one-inch monitor. The next basic is the color. The two sexes have somewhat different tastes. Men, says John van Raalte, an RCA scientist, are less sensitive to colors in the reddish-orange range than are women. That doesn’t mean, however, that, as a rule, women should automatically have reddish-orange monitors. For men and women the optimum range of light sensitivity is usually amber or green. Comfort is high in those ranges, too, at least for most people.[38] Footnote 38: For observations on the merits of various colors, see _Ergonomic Aspects of Visual Display Terminals_, edited by Etienne Grandjean and E. Vigliani and published by Taylor & Francis Limited, London, 1982. Amber screens are popular in Europe, where one study showed a lower error rate compared to green and other statistics said more users liked amber. American computer magazines have breathlessly praised amber. It’s as if they were fashion publications thrilled by the latest from abroad. But _some_ U.S. experts question the controls in at least one proamber study—for example, the number of volunteers, fewer than two dozen. And Bruce Rupp of IBM points out that it’s harder to produce a steady, flickerless image with amber than with green. What’s more, an executive with a company planning to sell amber terminals said screens of that color burned out faster than did green ones. Meanwhile, NIOSH says dark characters against a stable white background are especially promising; and Etienne Grandjean, a leading ergonomist with many admirers in the labor movement here and in Europe, agrees. Perhaps there’s a less jarring transition when your eyes move between the screen and the printed material you may be working with. It sounds logical enough, and in fact that’s what the Macintosh and many other computers use. That’s also the display style I saw in the Newspaper Guild offices in Washington. “Do you see any flicker?” asked David Eisen, the guild’s research director, who, in the labor movement, is one of the best-informed people on VDTs. I studied the white background. “The little lines,” I said, “seem to be rolling into each other.”[39] Footnote 39: If you don’t want lines on a CRT to seem to be rolling into each other when you‘re using a bright background, you should worry about something called a _refresh rate_. That’s the number of times the picture “repaints” itself on the screen. Etienne Grandjean, the prominent Swiss expert on VDTs, recommends a rate of at least 80 cycles a second; others say it needn’t be so high. The Xerox 860’s rate is 70 cycles, according to David Eisen, and that’s better than average. Still, I noticed the roll, anyway. In a letter to me in January 1983 Grandjean also recommended a slow phosphor for use with the white background. That means the images would take longer to vanish from the screen than they would otherwise—reducing the perceived flicker. The machine’s regular operator also saw the roll but said the screen was comfortable, anyway. The Guild machine, moreover, a Xerox 860, had the ability to revert to white letters against a black background, which would have eliminated the rolling. Would that all screens be as versatile, especially those used in many tasks or by people with different tastes. No matter what the task, however, don’t get caught up in terminal fashion. Stress the basics. Are the characters, for instance, shaped well? How big? At least .12 inch and preferably .16? Is the =dot matrix= at least seven by nine? Harry Snyder at Virginia Polytechnic Institute found that seven-by-nine matrixes produced “significantly fewer errors” than when subjects worked with those that were five-by-seven. If you can’t find the information in a machine’s literature, you might use a magnifying glass to count the number of dots making up the maximum widths and heights of the characters. Trying the large “N” and several more letters on my Kay pro, I learned that dot matrix was a mere five by seven—the bare minimum for light use; for a writer it could be better. You might also back off twenty inches to see if the dots merge almost completely into the letters and numbers. Eisen suggests looking for flaws like “flicker, character blurring at the edges of the screen, and adequate space between the characters and lines.”[40] Eisen obviously is looking at VDTs from the viewpoint of the union members. But unionists’ desire for comfort and ease of use often will overlap with employers’ desire for maximum productivity. Woe unto the employer who doesn’t take advantage of the overlap. In fact, Eisen’s observations here on keyboards and screens are within the ergonomics mainstream and shouldn’t be dismissed because of his affiliations. Footnote 40: Eisen’s advice on VDTs can be found in the booklet _Humanizing the VDT Workplace: A Health Manual for Local Officers and Stewards_, published jointly by the Newspaper Guild and the International Typographical Union (the price is $1.50 from the Guild, at 1125 15th St., N.W., Washington, D.C. 20005). “The two things you must remember,” sums up Waters, “are legibility and comfort. You may think at first that a five-by-seven matrix is fine. But after six months your eyes may say, ‘No, No!’—and you may get headaches and feel you need new glasses. Actually, you simply may need a better screen. You’ve just got to consider how often, long, and intensely you’ll be looking at the tube.” When shopping around, you also should ask, “How many letters and numbers—how many columns—can fill the screen?” For word processing you’ll normally want at least 80 columns across, the standard. Some screens, like that of the Victor 9000 microcomputer, can display 132, which can be just the ticket for spreadsheet aficionados with the right software. You’ll also want to know if you can highlight selected parts of the screen, for instance, categories of information stashed away in the computer. You might use =reverse video=—a light background and dark characters if your normal background is dark—to keep track of columns showing names and addresses. Not interested in phone numbers? Then those columns for the moment would be the normal light-dark combination. In word processing, reverse video could show you the blocks of texts that you planned to move or erase or underline. Two other amenities—more than amenities, in fact—would be easily reachable and adjustable controls for changing brightness and contrast. The Kaypro lacks a contrast control, much to my chagrin. At least I can adjust brightness—a “must” normally and one still “mustier” for people using antiglare filters, which reduce light from the screen. Keep thinking what you or others will most use the computer for. Color graphics? Fine. But your multicolor monitor won’t display characters as sharply as a black-and-white or =monochrome= screen would, and the screens might lose their crispness faster and add to the eyestrain of heavy word processing. One-color screens, then, would be best for typical writers and secretaries. On the other hand, if you’re an executive or clerk working with many figures alongside each other, the color monitor might help you and your eyes keep the numbers apart. Might. Some people may object to color except for graphics. Two kinds of color monitors are common. The =composite= monitor is technically more like a home TV than the =RGB= or “red/green/blue” kind. It’s cheaper. But colors blur into each other more than with the latter type. Whatever monitor you buy, keep the screen clean. And don’t be surprised if, in a year or so, the characters start looking a little fuzzy. Harry Snyder, a leading expert on the ergonomics of monitors, says the half-life of a tube in heavy use is normally about a year. Don’t stint on replacements and jack up your people’s error rates. THE KEYBOARD Remember Raquel toiling at a keyboard attached to her computer? Maybe she’s happy. She may have a well-built back, a graceful, flexible neck, wrists that don’t quit, or simply a body compatible with her machine and furniture. But I still question the attached board for most people. Unless Raquel spent many hours on the machine before buying it, she might not have known if she’d be comfortable. The more time you spend at a keyboard, the more likely you‘ll end up with the right one. Test keyboards thoroughly. If you‘re buying several dozen, you or a staffer might spend a good three or four hours with the machine you’re about to purchase. A difference of just 10 percent or so in the number of keystrokes per hour per operator could mean thousands of dollars annually to your business. Even if you’re buying a machine just for your own use, you should still put in a good half hour checking out the keyboard. That’s true with any computer. But the odds are stacked against you if you buy one with an attached board. In Norway, in effect, computer keyboards by law must be detachable, and in Germany official standards require them in most cases (an exception exists for, among other things, limited-use portables with built-in screens and keyboards).[41] Increasingly, union people here are raising the issue—and quite justifiably, considering the feeling of most ergonomics experts. Detachable keyboards, in fact, were one of the issues over which a California union struck the local Blue Cross-Blue Shield organization. Footnote 41: The German requirement for detachable keyboards appears in paragraph 4.3.1 of Standard ZH1/618 of “Safety Regulations for Display Workplaces in the Office Sector,” as released by the Trade Cooperative Association, Central Office for Accident Prevention and Industrial Medicine. “Well,” you ask, “what about keyboards on infrequently used machines?” You might, however, end up using the keyboard more often than you expected. So play it safe. Make all boards detachable. It won’t cost that much more. Also, keep the cords at least four feet long. And unstrap the keyboards from the VDTs in the first place—which sounds insultingly elementary, except that I’ve heard of an absentminded company that, until a labor dispute, had forgotten to do so. Also, don’t let the best keyboards become status symbols in reverse. A well-meaning but wrong executive on the West Coast brags he does _not_ have a terminal with even a good board. He’s a fast typist but thinks he could spend his time better giving dictation. “An executive typing,” he says, “is like hiring a brain surgeon to give out pills for a sore throat. If you’re paying him $50,000 a year, why should he do a $20,000 job two hours a day?” Most executives would agree. But this is rapidly changing; more and more executives are tapping out spreadsheets and eventually will do their own word processing. They may bat out rough drafts on computers or word processors, then have secretaries whip the documents into shape electronically without having to retype all of them. Besides, you can learn typing at any age. And you needn’t be an excellent typist for computers to help you; indeed, with mistakes easier to zap, the worst typists will profit the most. Take a Maryland architect-consultant named Jess McIlvain. Before he got an IBM PC, he just couldn’t write or type at length—worrying that the result “looked like the cat had walked over it.” Two years later he was cranking out reports several hundred pages long.[42] Footnote 42: The McIlvain example comes from _PC Magazine_, May 29, 1984. The new breed of managers, of course, may feel at home on the keyboard from day one. Some colleges even now require students to buy computers, many of which they’ll almost certainly use for word processing. What’s more, as you’ll read in Chapter 11, executive keyboards can ease the transition in the future to telecommuting. So don’t impose your own work habits on a junior executive with fast fingers. Why shouldn’t he or she have a deluxe keyboard if it helps him serve your company better? And you yourself shouldn’t feel awkward with at least a good board for quick memos. Mysteriously, IBM forsook the Selectric tradition in designing the keyboard of its original Personal Computer, where the left shift key, oddly, isn’t immediately next to the “Z.” The deviation left some computer magazine editors flabbergasted. Companies even started making keyboards for the PC—at several hundred dollars apiece—that retained the traditional layout. “They’re not like normal keys,” my consultant friend Michael Canyes complained of the ones on the PC itself. “They click on and off like switches. “A good key,” he said, “gets stiffer as you press down. But the IBM’s are either up or down.” Some cynics have a theory. Maybe IBM didn’t want the PC to steal too many customers away from the Displaywriter costing several thousand dollars more. Yes, keyboards are a matter of taste, and unlike Michael, I can stomach the switchlike =tactile feedback=. But the shift location is inexcusable. IBM in effect admitted its goof; the IBM AT, a more powerful cousin of the IBM PC and XT, appeared with a much-improved keyboard. At least the original IBM PC’s keyboard surpassed that of the tiny, chiclet keys on the first PC_Jrs._ Reluctantly—after sales were flagging—IBM replaced the chiclet-style boards with more conventional ones. Pity us “users.” Even terminals intended for mainframes and minis can be turkeys. I actually favor my Kaypro board over a terminal—from Digital Equipment Corporation—which Waters once was using with a big mini at work. Blame the old-style programmers who didn’t grow up with word processing. “Most can’t type like a secretary,” says Canyes, “so the manufacturers often skimp on the touch.” And isn’t it strange? You can buy an electric typewriter with adjustable touch for $200 but not get it in a computer costing fifty times as much. I asked David Eisen why he hadn’t raised the touch issue. Wouldn’t bad keyboard matches hurt productivity? “That’s more the companies’ problem,” he said. “We’re more worried about health.” He can’t recall a flood of complaints from Guild members about keyboards with the wrong touch. Union people, however, have talked about another change, flat keyboards, now popular in Europe. The tops of U.S. boards commonly slant at fifteen degrees. And Waters tells of an experiment where, given the ability to change the angle, most participants settled for the fifteen-degree one. A caveat, however, is in order. Maybe the people ultimately would have done better with a flat board and plenty of practice on it. Perhaps you want an adjustable board. But for me, anyway, it won’t matter as much as touch. Another issue is the QWERTY keyboard versus the more modern layouts. “QWERTY” means the first six letters found on the boards of nearly all American computers and typewriters. It’s a nineteenth-century legacy—a bad one. The early typewriters couldn’t keep up with the nimbler typists, so the machines’ designers thwarted the humans. QWERTY isn’t alphabetical, it doesn’t bring together commonly used letters like “t,” “h,” and “e,” nor does it use finger muscles properly. The Dvorak board and other latecomers should work better. In practice, though, they might confuse typists, so here’s a solution. Let us old-fashioned people QWERTY away. But think about buying machines you can switch over for people trained on more efficient layouts. A program like Smartkey—which electronically changes the keyboard—might be the answer if you also relabel keys on machines used by Dvorak typists. The Apple II(c) even has a switch on the top of the machine to go from QWERTY to Dvorak. Some say the Dvorak improvement is dramatic. Waters isn’t so sure. He says some studies have indicated that the Dvorak board offers as little as a 5 percent increase in speed and little reduction of errors. “So in the end,” says Waters, “most people in the industry have decided not to tinker with the key arrangement.” No matter how they’re arranged, your keys shouldn’t be shiny; a matte surface is good. So are neutral colors rather than black and white, except, says Eisen, for function keys. My Kaypro sins. Its keys, like those on Digital Equipment’s VT100, are a gleaming black, which, however, doesn’t matter that much to me, since I can control my lighting and normally don’t look down at the board when I’m typing. (Some other Kaypros have black matte keys.) Function keys—the ones that let the operator delete a word or add a paragraph with a single touch—should ideally be a different color from the main keyboard’s. The same for number pads. The function keys if possible should have labels indicating their purpose. If not, a chart near the function keys might show, for instance, that “P1” “Deletes Word.” Some people, however, say that function keys really slow you down, that they make you take your fingers off the main keyboard. I basically agree. The same would hold true for the use of Macintosh-style mice. SPECIAL ERGONOMIC FEATURES A dream VDT tilts. It swivels. It can move up and down on a pedestal. It helps you avoid back strain, stiff necks, glare, and headaches. It adjusts, so you don’t have to. A good VDT can help compensate somewhat if you lack ... Good Furniture “People will adapt very nicely to automation,” an IBM official once said, “if their arms are broken, and we’re in the twisting stage now.” That was in 1975. Since then, however, in a literal way, some computer makers have been gentler with people’s arms—and wrists and backs. A variety of good computer furniture exists now, some in walnut, some in formica. With all the trimmings, you can vary: