Click for next page ( 19


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 18
Hardware The U.S. computer hardware industry, a diverse collection of small and large firms, is composed of three major and integrally related industries: semiconduc- tor manufacturing equipment producers, semiconductor manufacturers, and manufacturers of computers ranging from personal and portable systems to supercomputers and associated peripheral equipment. The $4.6 billion U.S. semiconductor-manufacturing-equipment industry makes the tools- wafer-pro- cessing, testing, and assembly equipment that semiconductor manufacturers use to fabricate integrated circuits and related electronic devices, the source of revenues totaling $24 billion in 1988.1 In turn, 40 percent of all semiconductors sold in the United States are bought by domestic computer manufacturers, who, according to the Department of Commerce, were estimated to have sold $70 bil- lion worth of computing equipment in 1989.2 Perhaps double that amount may have been earned by domestic manufacturers in 1989 if systems software, ser- vice, maintenance, and leasing revenues are included, according to an industry trade group.3 Each of these industries is, in tum, divided into segments defined by the products they make and sell. Colloquium participants did not systemati- cally survey the hardware industry, but rather drew on their experiences to com- ment on what they saw as critical issues and trends. Serious cracks have developed in the chainlike relationship between these important domestic industries. Weakest in the chain is the semiconductor indus- try, which at the start of this decade held commanding leads in world markets for virtually all Does of integrated circuits. The Japanese pulled even with, and then surpassed, the approximately 250 U.S. merchant semiconductor manu 18

OCR for page 18
HARDWARE 19 facturers in percentage of world sales beginning in 1985. By 1988 Japanese manufacturers held a commanding 48 percent of the world market, whereas the world share for U.S. firms had fallen to 37 percent.4 Observers quoted in the media and at congressional hearings have described the decline as ruinous, and evidence from the domestic industries that are sup- pliers and clients of the U.S. semiconductor producers suggests that the descrip- tion is apt. The position of the U.S. semiconductor-manufacturing-equipment industry has fallen precipitously relative to that of Japanese equipment makers. U.S. firms still claim more than half of the $7 billion world market, but their Japanese counterparts now lead in sales of lithography equipment (for imparting circuit designs on silicon wafers) and other critically important tools.5 Some steps have been taken to arrest this decline, primarily through the formation of the SEMATECH consortium. More recently, several of the computer industry's leading hardware manufacturers have joined together in an ambitious effort to establish another major domestic supply of state-of-the-art dynamic random access memory (DRAM) chips. But this effort has only just begun, and its suc- cess is by no means certain. As it stands now, however, the nearly complete exit of U.S. merchant semiconductor suppliers from the market for DRAM chips by far the world's largest market for semiconductors weakens prospects for the equipment industry. Computer manufacturers have felt the demise of U.S. DRAM production from the client end. Forced to buy the memory chips from Japanese producers, which account for 90 percent of the world DRAM market, most U.S. computer makers faced serious shortages in 1988. Some were forced to delay introduc- tions of new products and to cut production.6 Consumers also felt the pinch as rising DRAM prices translated into a 15 percent increase in the cost of some computer models. (IBM, which produces DRAM chips for use in its products, was less affected by the shortage.) In contrast, Japanese computer manufactur- ers, large integrated fops that produce semiconductors for their own diversified line of products and sell the rest on the world markets, avoided the bottleneck and continued to increase their relative share of international sales of comput ers. The ripplelike effects and the contributing causes of what is now called the "DRAM fiasco" are instructive, and they will be discussed in greater detail below. Colloquium participants suggested that other segments of the hardware industry may experience comparable crises during the 1990s. A comment by Gordon Moore of Intel, the U.S. firm that invented the DRAM but no longer makes the chip, underscores the gravity that many ascribe to the situation. "If you take the whole sequence of things," Moore said, "in the end we are all going to be dead, and the end may be nearer than we think because the whole structure necessary to remain competitive is, frankly, getting out of our control increasingly." 1

OCR for page 18
20 KEEPING THE U.S. COMPUTER INDUSTRY COMPETITIVE SEMICONDUCTORS Invented in the United States in 1959, the integrated circuit launched what many call the Computer Revolution, and a seemingly endless series of advances in device technology have sustained the revolution. Of the hundred-thousand- fold improvement in computer performance since the 1950s (measured as a function of the cost of information-processing capacity), the greatest share stems from innovations in component devices (e.g., central processing units and internal memory) built from integrated circuits. These devices are often referred to as "semiconductors," a reference to the base materials from which they are made. For 20 years after the invention of the integrated circuit, U.S. fops' mastery of semiconductor technology translated into preeminence in world markets, and the rapidity of innovation was viewed as perpetuating the industry's dominance. In recent years, however, the relative importance of innovation as a competitive advantage has waned, and the value of manufacturing efficiency has risen, set- ting the stage for the U.S. industry's decline. "ET]o a very great degree, the United States arena functions as a kind of pub- lic service organization for worldwide industry," said Charles Ferguson, post- doctoral associate at the Massachusetts Institute of Technology's Center for Technology Policy and Industrial Development. "We innovate and people thank us for the innovations that we develop and then incorporate them into their com- mercial products, which they then sell, often to us." The "most spectacular and visible symptom" of this situation, according to Ferguson, is the nearly complete exodus of U.S. merchant semiconductor manu- facturers (firms that produce electronic devices for sale to other companies) from DRAM markets. Indeed, Ferguson and many others at the colloquium suggested that this episode embodies the nature of the challenge confronting all of the interdependent segments of the hardware industry. The "DRAM Fiasco" Intel's Moore believes that Japan's rise in DRAM manufacturing began when the evolution of the technology shifted from a series of rapid, seemingly random advances to a predictable course. In the early 1970s, theory and practice demonstrated that memory capacity would quadruple about every three years, from 1 kilobit, to 4 kilobits, to 16 kilobits, and so on. Japan established a series of programs aimed at developing world-competitive capacity for manufacturing 64-kilobit DRAMs, but, aided by a booming market for memory chips during the late 1970s and early 1980s, that nation's firms beat the timetable. Japanese semiconductor manufacturers capitalized on the shortage to achieve a signifi- cant share of the market for l~kilobit DRAMS. "Beyond that," Moore said, "they went to their usual strategy of overinvest

OCR for page 18
HARDWARE 21 meet," adding considerably to Japan's DRAM-manufacturing capacity. The firms then "engaged in dumping, selling significantly below their total cost," to sell the excess product that resulted from the overinvestment in capacity. In the mid-1980s, demand for semiconductors dropped, triggering an indus- ~ywide recession and exacerbating the price-reducing effects of excess DRAM capacity. During that period, according to Moore, the Japanese firms demon- strated another attribute to go along with their manufacturing prowess a "tremendous tolerance for pain." During the DRAM glut of the mid-1980s, the Japanese industry lost an esti- mated $4 billion, about double the loss sustained by U.S. manufacturers. "Clearly, they won," Moore said. "They came out with a 90 percent market share in DRAMs, and we abandoned the business." According to MIT's Ferguson, "In 1986, the world DRAM market was $1.5 billion. This year [1989], it is probably going to be $10 billion" approximately a sixfold increase. Today the United States is home to three merchant manufacturers of DRAMs Motorola, Micron Technologies, and Texas Instruments (which man- ufactures most of its components abroad) and most observers believe the con- tinued existence of the firms in this market is tenuous. With the exception of IBM, which has its own DRAM-manufactunng facility in Burlington, Vermont, U.S. computer makers are forced to shop overseas for much of their needed sup- ply of memory chips. Why did U.S. merchant semiconductor manufacturers choose to exit from the DRAM market? Moore's recounting of the factors underlying Intel's deci- sion not to compete is illustrative. By mid-1985, Intel had completed the design of a "good 1-megabit DRAM" and had worked out the manufacturing process, Moore said. Instead of pro- ceeding to the mass-manufacturing stage, the company chose to forego making the $400 million investment in equipment required to compete for about a 10 percent share of the world market. According to Moore, the company could not afford the investment and decided to leave the market, opting to concentrate on logic circuits and erasable programmable read-only memories (EPROMs). "I think had we made the DRAM investment at that time," Moore said, "we would be a much weaker company today." Today it is highly unlikely that an individual U.S. semiconductor firm would opt to enter the market and help bolster domestic production, according to Moore and others at the colloquium. In addition to Japan, South Korea and other Asian nations, as well as Western European countries, are building plants that will add greatly to world DRAM-manufacturing capacity. "For a company to go into a business like that with the idea of making money," Moore said, "frankly seems like a real folly. It is going to require something more than the usual market motivations to get a significant reentry in the United States."

OCR for page 18
22 KEEPING THE U.S. COMPUTER INDUSTRY COMPETITIVE Free market economics would appear to suggest that the United States does not need to manufacture DRAMs, given that other countries appear to do so more cheaply and that the world semiconductor industry seems destined for another glut of memory chips. Computer makers and other customers can pur- chase their chips from overseas suppliers, and U.S. semiconductor producers can devote their resources to products with higher profit margins. While acknowledging that this perspective is embodied in the tactical think- ing of individual firms, many colloquium participants pointed out that the col- lective result of these tactical decisions can be sectorwide harm. The nearly complete loss of the DRAM market, they maintained, inflicted strategic damage on the entire hardware industry, and perhaps the entire U.S. computer sector. Why? An answer is that DRAMs are technology drivers. Although their designs are far less sophisticated than those for logic devices, DRAMs have the densest circuitry of any integrated circuit. The same standard-cell unit may be repeated millions of times on a single chip. Thus DRAMs are the commercial test bed for tools and methods that achieve ever-finer line widths. Once perfec- ted and mastered, this manufacturing technology allows chip makers to squeeze more and smaller transistors and other subcomponents onto their microproces- sor chips, gate arrays, and other types of integrated circuits-the new devices that are the basis for the next round of improvements in computer equipment. Consequently, in all but abandoning DRAM production, U.S. merchant semi- conductor manufacturers lost the primary means for "addressing the kind of manufacturing issues that we have not done as well as the Japanese," Moore said. Moreover, the developers of the tools needed to correct these manufactur- ing weaknesses are suffering from the tandem effects of the loss of a major domestic market and the growing prowess of their competitors. Hewlett Packard's Doyle reported, for example, that a Japanese semiconductor company executive told him that 70 percent of the equipment in the firm's plant for mak- ing 1-megabit DRAMs had come from U.S. suppliers. In the company's facility for making the next generation of memory chips~-megabit DRAMs-only 30 percent of the equipment had come from the United States. "It is an extremely disturbing situation," said MIT's Ferguson. "If you are an American semiconductor producer, you have to ask yourself how deeply in your heart of hearts do you really trust Canon, Nikon, and Advantest. The answer cannot be terribly comforting." Nearly a quarter of Advantest, the world's largest maker of digital semiconductor test equipment, is owned by Fujitsu, a Japanese conglomerate that makes semiconductors, computers, and consumer electronics goods. Even a company as large as IBM, which has a tradition of meeting many of its needs internally, is feeling the effects, finding that it must become more self- dependent than it might otherwise prefer. "We use the best of the U.S. infra- stmcture, which, unfortunately, is not enough," said Toole. "We have to supple- ment what is available in tools and materials by our own developments within the company."

OCR for page 18
HARDWARE 23 The DRAM situation is emblematic of what many have characterized as the "hollowing out" of U.S. industry, the loss of the manufacturing infrastructure that supports many sectors of the economy.7 The computer sector and the fed- eral government have recognized this problem, and they have combined their resources to form the SEMATECH consortium, aimed at developing the advanced manufacturing technology needed for fabricating 4-megabit DRAMs and the next generation of logic circuits. Even if SEMATECH meets all expectations, its success may not reinvigorate the semiconductor and semiconductor-manufacturing-equipment industries. In the area of DRAMs, for example, it is unlikely that an individual firm will take the risk and make the capital investment required to achieve a major presence in DRAM markets. As a result, Moore and others said, the strategic commitment that building such a facility would require will not be forthcoming without external incen- tives or a means of distributing the inherent risk. Interestingly, shortly after the colloquium seven hardware manufacturers Intel, IBM, Hewlett Packard, Digital Equipment Corporation, National Semiconductor, Advanced Micro Devices, and LSI Logic announced their intention to build a joint manufactur- ing facility for producing memory chips. Called U.S. Memories, Inc., the facili- ty would require about $1 billion in capital, half of it to be provided by the seven founding firms and by additional partner firms that are being sought. The rest would be raised through debt financing and other means. Although busi- ness details have not been made public, initial plans called for production of 4- megabit DRAMs based on a design licensed from IBM to begin in 1991. However, less than five months later, observers questioned the future of U.S. Memories after IBM announced it would license some 4-megabit DRAM tech- nology to Micron and because of the absence of commitments by other major U.S. computer manufacturers.8 While any one venture must be evaluated on its merits, these circumstances and the shadow they cast on U.S. Memories once again called into question U.S. industry's capacity for sustaining a strategic, cooperative effort. Another means to help secure the still fragile position of U.S. semiconductor manufacturers, Moore and others noted, would be to subsidize memory-chip production to assure adequate supplies. Like federal subsidies for agricultural products, a DRAM subsidy would set a guaranteed minimum price for U.S.- made chips. Beyond DRAMs The U.S. semiconductor industry continues to maintain significant leads in microprocessors and other advanced areas of integrated circuit technology. In addition, colloquium participants predicted that the industry would continue to be the source of many of the innovations that drive the evolution of semicon- ductors and the products that use them, including computers. Participants were

OCR for page 18
24 KEEPING TlIE U.S. COMPUTER INDUSTRY COMPETITIVE far less certain, however, about the marketing advantages that would convey with innovation, given the rapid diffusion of technology and the demonstrated manufacturing strengths of Japanese firms and other Asian nations. "We had as big an advantage in DRAMs [as] we now have in microproces- sors," said Moore. "I think it is a matter of time until the competition is just as strong there. We cannot keep moving to the edge and living there very effec- tively." Signs of this progression are already evident, particularly in markets for application-specific integrated circuits. DRAM-manufacturing capabilities are directly transferable to these fully or partially customized electronic devices, and Japanese firms now control about 40 percent of this high-growth market.9 Speeding the competition's catch-up efforts, according to several colloquium participants, are technology-licensing agreements that, while advantageous to individual firms, are strategically damaging to the hardware industry. Among the technologies transferred to Japanese semiconductor manufacturers in recent licensing agreements are two advanced microprocessor designs based on reduced instruction set computing (RISC) architecture. Licensing payments can alleviate individual firms' cash-flow problems, of particular concern to small start-up firms. Also, by accelerating the diffusion of a new technology, licens- ing can expand the relevant market. However, licensing can also work like a Trojan horse: Often, foreign licensees emerge as the top competitors in licen- sors' markets. A particularly troubling aspect of the increase in licensing to foreign firms is that it reflects a relative dearth of licensing and other strategic alliances within the United States. Colloquium participants noted that large U.S. firms in the industry are often approached for financing from innovative start-up firms. But nearly as often, those large firms are not responsive, forcing the start-up compa- ny to look to the queue of foreign investors who come to the United States to shop for promising new technology developed by entrepreneurial companies. (See Chapter 5 for a discussion of this phenomenon.) COMPUTER EQUIPMENT Global competitive factors shaping the semiconductor industry are also influ- encing the world computer industry. Before the late 1970s, U.S. computer imports were virtually nonexistent; in 1988, sales of foreign-made computers and related equipment captured 39 percent of the $51 billion U.S. market.~ As is the case for semiconductors, U.S. firms hold the dominant position in markets for technologically advanced products, such as supercomputers, superminicom- puters, workstations, and high-performance machines with new processing architectures. But in the industry's largest unit-volume market, personal com- puters, foreign competitors account for more than 40 percent of worldwide sales, which are expected to total $25 billion in 1989.1 ~ Whether personal com- puters become the computer manufacturing industry's equivalent of the DRAM

OCR for page 18
HARDWARE 25 may well be determined by the manufacturing performance and strategic actions taken by U.S. firms during the l990s. The similarities between the competitive factors shaping the semiconductor and computer industries do not end there, however. The foremost foreign com- petitors in computer markets include many of the same Japanese firms that now account for more than about half of the worldwide sales of semiconductors. Not coincidentally, these firms have not experienced the same DRAM shortages that have beset many U.S. computer manufacturers. "There is a great deal of evidence," said MIT's Ferguson, "that the memory market is subject to strategic control whether coordinated or not is unclear.... It is subject to strategic control by the collection of vertically integrated Japanese electronics firms that dominate not only Japanese semiconductor pro- duction, but also Japanese computer, telecommunications equipment, and, to some extent, consumer electronics production. What is interesting about that, I think, is the striking degree to which this situation demonstrates the continuing ability of Japanese firms to both cooperate and compete." Similar trends are emerging among the growing computer sectors of other Asian countries, includ- ing Taiwan and South Korea, which are gaining a growing share of the personal computer market. Like the pattern of competition facing U.S. semiconductor makers, the tech- nological trends steering the evolution of the computer industry also are coming into view. These trends will promote significant changes in industry structure. Most significant, perhaps, is the commoditization of computers, particularly at the low end of the product spectrum. Commoditization Whereas individual companies may now manufacture thousands or several hundred thousands of machines in a year, soon the annual production levels of "each competitive, successful firm" will exceed millions of units, Ferguson pre- dicted. "As a consequence, the ability to manufacture in very large volumes, the ability to design products rapidly, to insert them into production rapidly, to have a worldwide sourcing, manufacturing, and distribution network twill become] ever more important. That is an area in which . . . the United States is not doing very well." C. Gordon Bell of Stardent Computer outlined the technological factors underlying this trend. According to Bell, all of the components that will be used in the commercially available machines in the year 2000 are now under development. His reading of the field suggests that personal computers at the start of the next century will be as powerful as today's supercomputers (while supercomputers of the day will themselves be far more powerful). Speeding this transition will be the already-visible movement to distributed computing environments, in which users can simultaneously exploit the resources and

OCR for page 18
26 KEEPING TlIE U.S. COMPUTER INDUSTRY COMPETITIVE applications not only of their local machines but also of any number of comput- ers that are accessible through a network. Both trends will increase demand for smaller computer systems, driving down their price and stimulating supply. According to Bell, while the performance of all levels of computers (mea- sured by the standard metric of millions of instructions per second, or MIPS) has been increasing at double-digit rates since the 1960s, the processing capa- bilities of microcomputers have been increasing at the astounding rate of about 40 percent annually, as compared with approximately 14 percent for minicom- puters and mainframes. With the anticipated widespread use of RISC technolo- gy in microcomputers, performance of what traditionally has been the industry's low-end offerings will increase at an annual rate of 70 percent. Bell did not foresee the extinction of high-end markets, such as for scientific uses and other computation-intensive tasks. But he raised questions about how demand for computers with different levels of performance will change. Will many people who require supercomputers today find that the powerful personal computers of tomorrow-operating alone or in combination with others- meet their needs? Will new applications and new approaches to problems emerge that cannot even be addressed with today's most powerful machines? Answers to these and other questions will determine the size, structure, and profitability of the computer industry in the future. Bell suggested that change will be rela- tively radical and that the vast majority of computers for sale at the beginning of the next century will resemble consumer electronics goods and will be produced in similar volumes. Technological Convergence of Industries The comparison of computers to consumer commodities was also made by Ferguson, who foresaw similarities that went beyond volumes of production. According to the MIT researcher, a common technology base of "digital optics, digital microelectronics, and digital magnetics" the technology areas that have been the wellspring of advances in computer and telecommunications equip- ment are "going to be used in a wide spectrum of products." Included in this category of products, by Ferguson's accounting, are digital facsimile machines, printers, high-definition televisions, and home communications and information systems. Early commercial versions of these items are already available; other products are under development. An example of this new trend toward horizontal integration in design and production is Canon, a Japanese firm. Known to most American consumers as a manufacturer of cameras, Canon also makes laser printers, photocopiers, and photolithography equipment. With this diversification into new product lines, which exploit the same optical and electronic technologies, the company's rev- enues have increased tenfold to $10 billion-during the last 10 years, Ferguson said. While IBM has a broad product range, the U.S. computer-hardware industry

OCR for page 18
HARDWARE 27 overall is not characterized by extensive horizontal integration. Moreover, the virtually complete loss of the nation's consumer electronics industry makes it difficult for U.S. companies to exploit the converging technology base and the production economies of scope it implies and to diversify into many product lines. Manufacturing Both commoditization and convergence to a common technology base make manufacturing efficiency and flexibility essential for competing in domestic and global markets. "This scenario is manufacturing- and semiconductor-intense and, hence, subject to intense competition in areas in which we perform poorly," said Bell, adding later that the nation's general "lack of manufacturing ability" is the primary source of the hardware industry's declining competitiveness. According to Ferguson, "Industries already are making a transition to a mass- produced, low-unit-cost structure. Those industries are increasingly dominated by Asian forms." Ferguson predicted that U.S. manufacturing weaknesses will become an increasingly costly liability as manufacturing capabilities become the primary "determinant of long-run competitive success" in computer manufacturing and the growing cluster of related industries. Moreover, the continuing-and, some would say, accelerating-erosion in the upstream industries that supply materi- als, manufacturing equipment, displays, and electronic components will handi- cap U.S. computer manufacturers' abilities to develop the necessary level of manufacturing proficiency. "We have to find ways to improve and to protect our fragile manufacturing base," IBM's Toole said. "We are at a huge disadvantage because of the virtual absence of consumer electronics." As pointed out in several studies, Japanese firms have achieved higher-quali- ty products at lower cost through continuous improvement in processing. Incremental improvements, accomplished by fostering the translation of innova- iions into new products, can help a firm maintain a competitive advantage. A telling indicator of the different manufacturing perspectives of Japanese and American firms is their spending patterns. Citing a study by Edwin Mansfield, Doyle of Hewlett Packard noted that the average U.S. company spends 70 per- cent of its R&D budget on product development and only 30 percent on the manufacturing process. In the average Japanese firm, the percentages and prior . . . sties are reversed. STANDARDIZATION Hardly separate from the industry-shaping trends described above, the inter- national push for standardization (affecting computer hardware, software, and data communications) is a catalyst in each hardware industry. Although carried

OCR for page 18
28 KEEPING THE U.S. COMPUTER INDUSTRY COMPETl71VE out by domestic and international standards-setting bodies, the push is prompted by users seeking compatibility and interconnection among the machines of dif- ferent manufacturers. Companies may also advocate standardization if a partic- ular set of standards improves their competitive position or at least diminishes the advantages of the market leaders. Two examples, using IBM products, illus- trate how standardization can be a boon or a bane to an individual company. Based on internally developed standards, the 360 System family of compatible computers cemented IBM's leadership position in the large international market for mainframes. The company's proprietary control of what became a de facto industry standard gave it a clear advantage, and other computer manufacturers directed their attention to market niches not filled by IBM mainframes. IBM chose a different tack when it made its relatively late but hurried entry into the personal computer market, revealing its microcomputer architecture to the rest of the industry and purchasing the machine's operating system from an outside vendor. The openness of the company's standard was effective in pro- moting the development of software and peripherals by other firms, assuring, as IBM intended, that customers had ample support The strategy made the IBM PC a marketing success. But within about five years after the computer was introduced, cheaper PC "clones," most assembled in the Far East, began to claim significant shares of the market, and they continue to do so. U.S. sales of imported clones in 1988 were up 50 percent over the previous year.12 To con- found clone manufacturers, IBM has introduced a new family of personal com- puters (the Personal System/2 series) with proprietary features. While IBM was reasserting more proprietary control over its line of personal microcomputers, Sun Microsystems was exhibiting yet another approach to standardization, actively encouraging foreign and domestic firms to adopt its Sparc microprocessor. Some observers predict that Sun is destined to repeat IBM's experience with its personal computer, but Sun Chief Executive Scott McNealy espouses what others view as a pragmatic view of the drive for stan- dardization. "I don't believe the world can go back to proprietary systems," McNealy said after his company licensed two Taiwanese feds to manufacture personal computers based on Sun's microprocessor design.13 Sun is a staunch advocate of standardization, and through its licensing agreements with foreign and domestic firms, it is maneuvering to have its system adopted as the industry standard, thereby creating a relatively high-volume market. Critics of the open-architecture stance taken by Sun and other U.S.-based companies argue that these firms have made it easier for foreign competitors to perfect the technology, speeding their entry into high-end computer markets now dominated by U.S. firms. Although colloquium participants did not agree on the point at which standardization handicaps innovation, most agreed that user demand for increased compatibility among the machines of different ven- dors and for a greater choice of applications will result in more hardware stan- dards. Most also agreed that as the number of software and hardware features

OCR for page 18
HARDWARE 29 shared by different vendors increases, the importance of manufacturing efficien- cy increases, and more firms may enter computer markets, competing on the basis of cost and product quality rather than innovation. Given the naiion's manufacturing weaknesses, IBM's Toole said, "Standardization will probably hurt both the technology and systems competitiveness of the United States, as it moves the contest to the area of Japan's greatest strengths." NOTES 1. U.S. Department of Commerce. "Electronic Components, Equipment, and Semiconductors," 1989 U.S. Industrial Outlook (Washington, D.C., 1989), p. 30-9; 1990 U5. Industrial Outlook (Washington, D.C., 1990~. U.S. Department of Commerce, 1990 U5. Industrial Outlook, 1990; figure pro- vided via personal communication with a Department of Commerce analyst. CBEMA. Information Technology Industry Global Market Analysis, Industry Marketing Statistics Committee, Washington, D.C., 1989. 4. Dataquest as cited in "Japanese Solidify Dominance of Semiconductor Market," Washington Post, January, 4, 1989, p. F1. 5. U.S. Department of Commerce. 1989 U3. Industrial Outlook, 1989, p. 30-9. 6. U.S. Department of Commerce. 1989 US. Industrial Outlook, 1989, p. 26-2. 7. See, for example, Jodi T. Allen. "Regaining America's Dulled Industrial Edge: the Answer is Right Here at Home," Washington Post, May 10, 1989, p. F3. 8. Richards, Evelyn. "Future of Joint Chip Venture Now in Doubt," Washington Post, November 10, 1989, p. F1. 9. Dertouzos, Michael L., Richard K. Lester, and Robert M. Solow, MIT Commission on Industrial Productivity. Made in America: Regaining the Productive Edge (Cambridge, Mass.: MIT Press, 1989), p. 249. 10. U.S. Department of Commerce, 1989 U5. Industrial Outlook, 1989, p. 26-1. 11. U.S. Department of Commerce, 1989 U.S. Industrial Outlook, 1989, pp. 26-9 to 26- 11. 12. U.S. Department of Commerce, 1989 US. Industrial Outlook, 1989, p. 26-7. 13. Fisher, Lawrence M. "2 Taiwan Licenses Set by Sun Microsystems," New York Tunes, June 22, 1989, p. D4.