Printer Friendly

Industry report - information technology.

Industry report-- Information Technology

America's standard-bearer, the hightech computer/software/semiconductor and communications industry, is big and outwardly healthy. Inside, though, it almost qualifies for the critical list.

As recently as last year, the Information Technology (IT) industry's rosy cheeks were deceiving people. In their report called Information Technology R&D--Critical Trends and Issues, for instance, the US Congress's Office of Technology Assessment (OTA) reported that IT "is generally robust as measured in a variety of ways, and in comparison with US industry as a whole.

"For example, during the 1972-82 period, sales revenue grew by 66 percent, compared with 40 percent for the composite US industry, and profits grew by 36 percent, compared with 6 percent for the composite.'

(The OTA defines Information Technology as electronic hardware that is used to create, communicate, store, modify, or display information, and as programming or software that is developed to control the operation of that hardware.)

Figures compiled by the US Dept of Commerce (see graphs) indicate that this large aggregation is now a major component of US manufacturing, and that its growth rate has been phenomenal, especially when compared to other, older industries. But recent reports--including DoC's own 1985 US Industrial Outlook --reveal that our IT manufacturing base is eroding at an alarming rate.

Computer hardware--slipping into trade deficit

During 1984, shipments by US manufacturers of electronic computing equipment (SIC 3573) reached about $53 billion in dollar value, up 20 percent over 1983 (after adjusting for inflation). This year, the industry's shipments are expected to grow by nearly 17 percent over 1984, to about $62 billion.

Despite this vigorous growth, US makers of computers and peripherals are deeply concerned. For one thing, since 1981 the balance of trade has been shifting in favor of Japan and other foreign producers (see graph). Exports of computing equipment should rise at nearly 30 percent this year, but imports will grow faster than exports. Early this year, the DoC forecasts that the US trade surplus in computing equipment could drop from 1984's $5.4 billion to about $2 billion for 1985.

A large segment of the computer industry's market--office automation, or OA--has escaped Japan's onslaughts. This segment is coming to be dominated by the large, full-line systems suppliers such as IBM and AT&T.

"A third possible contender in this segment is what I would call Japan Inc,' says David Power, partner and IT industry specialist at Temple, Barker & Sloane Inc, a management consulting company based in Lexington, MA. "Japan Inc includes the large suppliers, notably Hitachi, Fujitsu, and NEC. These companies have the commitment, financial strength, components, and technical skills to put together large, integrated office systems.

"Thus far, however, they have not demonstrated success in system integration,' Power continues. "People here worry about Japan Inc, particularly NEC, but at this point it is unclear whether or not they can succeed.

"Language and alphabet differences may present major problems for them. In addition, the Japanese companies may suffer from a lack of adequate representation in the US. Certainly, at this point, they do not have a strong position in the US marketplace.'

Despite their current relative security in the OA market, US producers have problems with foreign competition. The problems lie primarily in those portions of the market involving price-sensitive, standardized products. Losses to foreign producers have been occurring in lowcost disk drives, printers, and display terminals. US producers continue to dominate the high end of the peripherals market, and the computer market as a whole.

In an effort to meet the low-end price competition from foreign companies, many of our computer manufacturers have not only been buying parts and subassemblies offshore, but have also been moving production en masse to countries offering lower labor costs and taxes. A notable example that made the news this year is Motorola Inc, based in Schaumburg, IL.

According to a story in Business Week (April 15, 1985), more and more of the company's new design, production, and assembly facilities are going to countries such as Japan, South Korea, Singapore, and Taiwan.

"Motorola's work force, currently 30-to 35-percent offshore, could shift to 50 percent over the next five to ten years,' says the report. "Galvin (Chief Executive Robert Galvin) insists Motorola and other large multinationals have no choice.'

The story quotes Galvin as saying, "We will do what we have to do in order to survive, but that survival includes a process of defection. We are defecting from this country.'

The manufacture of computers and peripherals involves hundreds of types of metal parts, including stampings (as for housings and backplanes), die castings (as for disk drives), and screw machine products (for instance, connectors). Over the past decade, many US producers, looking for lower parts prices, switched from domestic to foreign suppliers. Lately, though, a partial reversal has begun to occur, at least for some manufacturers.

"We've heard reports that some buyers of castings have discovered that the grass on the other side of the pond is not quite as green as they thought,' says Peter Findlay, president of the American Die Casting Institute (ADCI), Des Plaines, IL.

"When a buyer procures castings offshore, he may find, for one thing, that just-in-time deliveries are not practical,' Findlay continues. "Then too, he may have to send purchasing agents overseas often, and communications costs go up. It's also more costly to replace rejects, and to make design modifications.

"In other words, these buyers are finding that there is more to the total cost of castings than the quoted price. As a result, some members of our organization --particularly West Coast die casters--report that they are regaining business from customers who went offshore a year or two ago.

"It's a mixed picture, though,' he adds. "We can't say that this return to domestic sourcing applies to all suppliers across the country.'

Partially stifling overseas sales by US computer manufacturers is the desire by many foreign governments to develop and protect their own fledgling computer industries. Sales of US-made computer products have been brisk in the Far East and Latin America, for instance, but in some cases the trade doors are swinging shut.

The Third World nations have seen the success of US, Japanese, and European marketers, and naturally are attracted to computer production. Startup costs are relatively low, so the potentials for indigenous production as well as sales look inviting.

To encourage their own computer industries, certain governments have been imposing policies that limit local selling to local producers. This practice not only shrinks the potential exports of US computer manufacturers, but also makes it safe for the locals to turn out inferior, outdated, and overpriced products. Over the long run, that hurts the worldwide computer industry as a whole.

In competing among themselves as well as with foreign producers, US computer makers have been automating many phases of production. The goals: to raise quality levels, reduce rework and rejects, and cut costs.

Westinghouse, for example, is well along in a $50-million program to automate their printed circuit board (PCB) production facilities in Texas and Maryland. Thus far, the company has boosted first-pass yields of PCBs from the previous miserable 12 percent to over 80 percent. The long-term goal is to hit 95 percent.

On the West Coast, Apple Computer's two-year-old Macintosh plant in Fremont, CA, has become known as a showcase of advanced computer production. Robots, automated storage/retrieval systems, and bar coding are but three of the technologies employed here. According to Apple, the key ingredient is total, real-time integration of shop-floor processes with a plant-wide control and information network.

Continuous tracking of bar-coded components and assemblies throughout the plant gives the company full, current data on production status and inventory levels. These data also provide continuous updating of the plant's manufacturing resources planning (MRP-II) software.

Examples such as these abound throughout the industry. IBM Corp, Hewlett-Packard, Digital Equipment Corp--these and dozens of other manufacturers, small as well as large, can boast of achievements in automation. But the main reason the US has been able to achieve leadership in computers is that we innovate faster and better than do our competitors.

Today US producers find themselves in a race with Japan, France, the UK, and West Germany, to develop bigger, faster, more versatile computers--the so-called supercomputers. Much of the R&D in this country is aimed at developing hardware (and software--the two work together) for military applications such as battle management and missile guidance.

As the OTA report indicates, since 1950 the vast majority of computers have been designed around one architectural model. This is the von Neumann machine, invented by mathematician John von Neumann. In this architecture, data and program instructions are stored, retrieved one by one sequentially, and acted upon individually by the processor.

Lately, though, computer designers have been studying ways to make it possible for the computer to break large chunks of data and calculations into smaller chunks, and then process them simultaneously in parallel. One approach, called the vector design, has been adopted by Cray Research Inc, a leading supercomputer maker based in Minneapolis, MN. This year, Cray will introduce the Cray 2, a four-processor machine that can work on four chunks or vectors of data at one time. Coming next year will be the Cray 3, an 8- to 16-processor machine using gallium arsenide (GaAs) rather than silicon in the chips.

Over a dozen US companies are now actively pursuing new techniques, components, and products in supercomputing. Also being pursued are new, economical designs in custom computers for specialized applications such as signal, image, and graphics processing.

One of the major government activities in supercomputer R&D is being conducted by an agency called DARPA-- for Defense Advanced Research Projects Agency. This agency has begun a Strategic Computing Program to develop artificial intelligence hardware and software for the Dept of Defense. The National Science Foundation, NASA, and the Dept of Energy have also launched programs to increase computing power for their researchers.

But the Japanese and others are hard on our heels. Led by Hitachi and Fujitsu, the Japanese have been pursuing what they call their Fifth Generation Computer Project. In the other direction, Britain's government supports the Alvey Program, engaged in R&D for new semiconductors and advanced software. The French are well into their La Filiere Electronique Program, a five-year national R&D effort, and several European nations have banded into Esprit, a program intended to draw together the continent's technical resources for information R&D.

Thus far, US funding for computer R&D has been rising to help meet the growing competition. According to the OTA, however, efforts in this country to outpace the competitors are being hampered by a worsening scarcity of qualified researchers and technicians, and by an unduly large outflow of technical information.

Other challenges facing our computer industry's R&D efforts include accelerating obsolescence of research equipment and software, and too much emphasis by corporate management on short-term developments and dollar returns.

Because of its expertise in R&D, the US computer industry will likely retain a lead in the big-ticket end of the industry. At the smaller-ticket end, however, imports will probably continue to rise, and more production by US-owned companies will move overseas. As the DoC points out in their report, "The US trade surplus in computer equipment could fall into deficit, perhaps as early as 1986.'

Computer software--challenged, but relatively secure

Now large enough and sufficiently discrete to carry its own Standard Industry Classification (SIC 7372), the US software industry grew by 33 percent in 1984 over 1983, posting sales of nearly $18 billion by year's end. Employment has also been rocketing; growth in the number of software jobs rose during 1983 by 27 percent, to 247,000. Estimates of the total number of programmers in the US today range upward of half a million.

One of the factors behind this strong growth is the rapid increase in the size of the installed computer base. Another is the high level of software investment by US companies attempting to increase productivity. In large computer systems, the cost of the software may exceed that of the hardware by four times.

The US software industry receives over 20 percent of its revenues from overseas customers. This is an average figure for all forms of software; overseas sales of packaged software account for about 30 percent of total revenues, with overseas custom and integrated-package sales bringing in much smaller percentages.

Research and development is considered a crucial activity by software producers. As a percentage of total revenues in 1983, R&D spending was 10.4 percent --considerably higher than the industry composite of 2.6 percent for that year.

Both the federal government and private industry have initiated major efforts to speed development and modification of software. Advanced program editors and debuggers, new programming languages, new design methods, and improved programming environments: all promise to help US industry develop a more efficient, systematized software engineering discipline.

Spending by industry for software development alone topped $10 billion in 1984. Some companies, such as Hewlett-Packard, spend nearly two-thirds of their R&D budgets on software development. Over 40 percent of the technical people at AT&T Bell Laboratories, including about 300 PhDs, are engaged in software development. IBM reportedly has about 150 researchers working on new software engineering techniques.

One of the world's largest users of software is the US government. According to the OTA, a recent study found that 95 to 98 percent of the government's application software is custom-developed. The Dept of Defense alone spends $4 to $8 billion a year on software.

To deal with problems of software engineering and maintenance, the DoD has proposed a new program called Advanced Software Technology. DoD sees this as a 10-year effort costing $250 million. Congress approved funding of $10.5 million for 1984.

In addition, the National Science Foundation funds research in software engineering by universities and corporations. Total NSF funding in this field ranges between $5 and $10 million a year.

As in the computer hardware industry, software developers are not without competition from abroad. The Japanese, for example, have had a software engineering project underway since 1970. Funded by the government and credit banks, a consortium called Information Technology Promotion Association spent an estimated $25 million in 1983 for software R&D.

An interesting facet of Japanese software R&D is the setting up of "software factories.' Staffed by people from member companies in consortia, these factories provide integrated environments for software design, development, testing, and maintenance. According to one report, the factories have achieved notable productivity gains, partly through reuse of an average 30 percent of existing code in new applications.

As part of the Alvey Program mentioned previously, Britain plans to spend about $100 million for software R&D over the 1984-89 period. The British intend to establish an Information Systems Factory by 1989; here researchers will focus on integration of the tools and methods for improving development of both hardware and software.

In the past, the French have been noted for the quality of their software. Two French government agencies run major software R&D centers employing some 100 researchers. About 11 other government and private labs are engaged in software engineering R&D.

Aside from engineering, one of the other major fields of software R&D is artificial intelligence (AI). This field includes expert systems, natural-language processing programs, and image- or vision-processing programs. The last-named includes developments for robot guidance as well as for inspection, metrology, and other vision applications.

The OTA reports that in 1983, US industry spent as much as $75 million on AI products. An estimated 15 major US corporations are engaged in product R&D in this field.

Among government agencies, the DoD is the leading funder for AI research. The DARPA Agency mentioned previously, along with the Office of Naval Research and US Air Force, have been deeply involved, and the National Science Foundation has been awarding grants for AI research since 1950.

Overseas, the Japanese, British, and French are vigorously pursuing R&D in AI. Probably the most ambitious effort is part of Japan's Fifth Generation Computer Program, which aims to ultimately produce an integrated prototype involving supercomputers and AI software.

"It is still too early to measure the potential market for commercial AI products,' points out David Power. "A few companies have brought out pioneering products in natural-language and vision processing, and in elementary expertsystem software. Much research and development work remains to be done, though, in supercomputers as well as in software.

"Further, large masses of data need to be gathered, sorted, and loaded to provide databases for AI systems. The big push in AI probably won't come until the 1990s.'

Our domestic software producers must struggle with formidable problems. These include:

Piracy. US software vendors claim that illegal copying has reduced their sales volume by several billion dollars a year. "In one Far Eastern country,' says the DoC report, "a copy of a popular $500 US software package currently sells for less than $10.'

A lack of applications software for supercomputers. This has been a significant barrier to their widespread adoption.

A shortage of personnel skilled in developing software for advanced computer architectures. The shortage is especially acute in the areas of scientific and mathematical applications.

Difficulty in attracting university faculty members who can train the next generation of software researchers. The comparatively low salaries and restricted opportunities offered by universities are blamed for this problem.

Risk and difficulty in introducing new techniques into software production. Conversion often calls for work on large existing inventories. Further, most of today's software development relies on outdated methods and attitudes, and is oriented toward single, customized projects.

Mounting demand for software from a production base that cannot keep pace. The lag between definition of a large application and delivery of software for it now ranges up to two and one-half years, and the length of the lag keeps growing. In addition, much programming time is being wasted on debugging old programs because of design errors, coding errors, or poor initial specifications.

Inability to systematically reuse parts of old programs in the generation of new ones. Much time is being squandered on re-invention.

A high rate of turnover among computer system professionals, particularly those who write code. The rate has been estimated at 15 percent a year.

Despite all these problems, however, and despite the slowed growth in the US economy this year, our software industry expects a 1984-85 revenue growth topping 30 percent. Sales this year could exceed $24 billion.

Long-term, the industry anticipates continued sales growth of over 30 percent annually through the balance of the decade. According to the DoC report, foreign competition in software will mount throughout the 1980s, "but it should have a minimal effect on the US industry.'

Foreign software industries ought not be ignored or taken lightly, though, cautions the DoC. The governments of South Korea, Taiwan, Singapore, and India--as well as Japan, France, and the UK--have all made commitments to establishing strong, local software industries.

Semiconductors--brilliant, but in deep trouble

If there is one industry that typically illustrates the American dilemma--inventive genius coupled with vulnerability in production and finance--it is the semiconductor industry (SIC 3674).

On the one hand, US companies and universities continue to lead the world in design and prototype production of new, more powerful, more compact microchips and other, related devices. The rate at which these products are being invented and improved is nothing short of dazzling.

As pointed out in the OTA report, the number of transistors that could be packed on a single chip doubled each year--from 11,000 in 1972 to 600,000 in 1981. The technology has now achieved the capability to pack a million components on one silicon chip.

What's more, the minimum widths of circuit connecting lines have been reduced by a factor of two every six or seven years. "Using photolithography with visible light,' says the OTA report, "line widths have been reduced from 25 microns in 1972 to about 1.5 microns currently.' Other methods of lithography --x-ray, electron beam, and ion implantation--may soon produce line widths in the 0.1 to 0.01 micron range. These widths are equivalent to rows of 20 to 200 atoms.

One result of increasing the packing density has been an increase in memory capacity of random-access memory (RAM) chips. For example, just this past February, IBM's General Technology Div, Essex Junction, VT, announced that it had fabricated a new, one-million bit memory chip. It measures only 7/32 3/8, and is said to be one of the fastest megabit chips produced anywhere. Normal operating speed is 80 nanoseconds; initial samples have operated as fast as 60 nanoseconds. (A nanosecond is one billionth of a second.)

Dataquest Inc, semiconductor market research company based in San Jose, CA, predicts that at some time during the next decade, the world market for RAM chips will hit $15 billion, and that about 66 percent of that will be for DRAM (dynamic RAM) chips. By 1995, the DRAM market alone may top $30 billion, about half of that going for 4-megabit chips. Some observers think that the Japanese are already ahead of us in this field, particularly in the financing and development of facilities for low-cost mass production.

US researchers, as well as the Japanese, are pursuing ways to achieve even greater component density by stacking chips in 3-D arrays. Spearheading the drive in this country is the DoD, with its Very High Speed Integrated Circuit (VHSIC) program. The DoD has already committed about $500 million to this R&D program, with funding spread among nine contractors.

The microprocessor chips currently in use contain several hundred thousand switching transistors. VHSIC chips, on the other hand, will contain millions of transistors, each with submicron dimensions. Designed to provide higher switching speeds and greater circuit density than previous processors, VHSIC chips will be used initially for military applications such as battle communications, radar signal processing, and "intelligent' avionics.

One of the latest developments in private R&D programs is the "butterfly chip,' from Bolt Baranek & Newman, Cambridge, MA. This special type of microprocessor will link 128 processor chips in a supercomputer, enabling it to run 60 million instructions a second. At latest estimate, a 128-chip system will cost about $1.2 million--one fifth to one seventh the cost of a 1985 supercomputer with equivalent speed.

Another promising development is the so-called "ballistic transistor.' In this device, the signal-bearing electrons theoretically don't collide with any atoms in the semiconductor material. Collision-free travel could enable the signals to move hundreds or even thousands of times faster than they can today.

AT&T Bell Laboratories, one of the major researchers in this field, is working toward a ballistic transistor that can be turned on and off trillions of times a second. Last fall, Bell Labs built a device that switches at 9.4 picoseconds. (A picosecond is one millionth of one millionth of a second.) The Japanese are hot on the trail, too, however; researchers at Tohoku University reportedly will soon unveil a 1.0 picosecond transistor.

Gallium arsenide (GaAs) is finding an increasing number of uses in microelectronics, but another use--being explored here by Bell Labs, Honeywell, and Rockwell International--is that of photoelectric transmitting, receiving, and switching devices for fiber optic systems. These devices will permit more efficient integration of laser and other light-transmitted signals with microelectronic circuits.

Partly to counteract the collective power of Japan's Ministry of International Trade and Industry (MITI), a dozen US companies have set up a cooperative R&D venture called Semiconductor Research Corp (SRC). Located in North Carolina's Research Triangle Park, SRC will channel about $22.5 million this year into semiconductor research. The money goes to some 450 researchers in universities. Funneling of R&D funds and assignments through this coordinating venture is said to help prevent duplication of research efforts.

Despite all its brilliant inventive work, though, the US semiconductor industry finds itself in serious financial difficulty. Early predictions of growth rates and shipment volumes have been pruned drastically.

Lately, weakening computer sales, along with the general softening of the US and world economies, have caused microchip inventories to pile up at alarming rates. As a result, prices of chips, particularly the ubiquitous 64K RAM chip, have plummeted. This past May, chips previously selling for $3.50 apiece on the spot market were going for 75 cents or less.

Then too, competition from foreign plants, particularly Japanese, has been forcing this country into an ever-deepening semiconductor trade deficit. Figures compiled by Business Week indicate that the balance crossed the zero line in May, 1981, and is plunging steeply. As of the end of last year, we had nearly a $3 billion trade deficit in semiconductors and related devices.

Many US chipmakers have been fighting back, countering the offshore labor and tax advantages with new, highly automated plants. One problem, though, is that the cost of erecting and equipping a chipmaking plant has increased roughly tenfold during the past 10 years.

The larger producers can afford to build, but the smaller companies cannot. Some of them have resorted to forming joint ventures with foreign chipmakers. This may solve the problem for the US companies, but it also creates large holes through which our technology can leak. (Of course, leakage can flow in either direction.)

A recent example of this sort of arrangement involves RCA Corp and Japan's Sharp Corp. Each company felt it was too small to raise the capital for individual chipmaking plants, so the two pooled their resources in a $200-million joint venture. The new plant will be built somewhere in this country.

According to a story in Business Week (May 6, 1985), RCA hopes the new venture will enable it to sell more of its chips in Japan, and Sharp yearns to capture a larger share of the US market. Too, RCA wants to learn more about production of CMOS chips from Sharp's manufacturing people.

Despite all the problems endured by the industry today, the DoC forecasts that shipments of semiconductor products by US manufacturers will grow at a 15-percent compound annual rate through 1989. "Continuing price declines and technological progress will increase the constant dollar figure by 25 percent each year,' says the DoC report.

Slowdowns in economic growth, economic downturns, and Japanese competition will continue to exert pressures on the industry. Because of the likelihood of sustained growth in demand, even further inroads by the Japanese will still allow for sizeable growth rates in US semiconductor production, the DoD predicts.

Radio and TV communications-- healthy, but also dipping into deficit

Due in large part to the currently high levels of military spending, the US radio and TV communications systems and equipment industry (SIC 3662) is now in fine fettle. The dollar value of the industry's shipments this year should top last year's by about 8 percent, according to DoC forecasts.

Included among products of this industry are radio and TV communications gear of all types; intercom systems, alarms, and signals; search and detection, navigation, and guidance systems and equipment; space satellite communications systems; fiber optic cable, devices, and systems; and miscellaneous such as ultrasonic equipment, simulators, teaching machines, and laser communications systems.

The largest category--search and detection, navigation, and guidance systems and equipment--accounted for roughly 57 percent of the total dollar value of shipments in 1984. This category grew more than 12 percent over 1983. As long as demand from the military continues, the industry will thrive.

As for foreign sales and the trade balance: in the DoC's estimate, exports were up about 9 percent in 1984, but imports increased by 15 percent. Some 47 percent of the imports came from Japan, but Canada, Mexico, and Taiwan also shipped in significant amounts.

Researchers at McGraw-Hill say that the US industry hit the zero line in early 1983, and that we ran a $700-million trade deficit last year.

One of the industry's categories that shows particular promise for growth, independent of defense needs, is fiber-optics communications. Approaching $1 billion in 1984, the world market for fiber optics is projected to reach $3 billion by 1989. Chief competitors in this field are manufacturers in the US, Canada, West Germany, and Japan.

Already being employed extensively in telephone communications, fiber optics will also find widespread application in local-area networks for towns, building complexes, and large factories and offices. In the OTA's opinion, heavy investments in existing copper wire and coaxial cable will retard adoption of fiber optics in existing large telephone and cable-TV systems.

Another category with great potential for growth is that of space-satellite communications systems and equipment. In 1984, industry shipments approached $2 billion, says the DoC report. The total could reach $5 billion by 1989.

According to Dr Jerry Waylan, president of GTE Spacenet Corp, McLean, VA, these dollar figures can be misleading because the satellite industry is in a state of rapid change. "The cost for earth stations is going down fast,' he reports, "and we don't know when it will bottom out.'

Today, about 60 percent of the traffic carried by satellite communications systems is video. This includes network and cable TV broadcasts, and newshaul to media. Voice communications account for about 30 percent of the traffic, and data communications the remaining 10 percent.

"Over the next eight to ten years, we see data transmission taking the lead,' says Waylan. "By 1994 or so, data will account for about 60 percent of the traffic. Video and voice will continue to grow, but at slower rates.'

He points out that many large companies are becoming interested in satellite data transmission for computer timesharing and inventory control applications. Xerox Corp, for example, is using a number of 4-ft-dia antennas and a GTE satellite to link timesharing terminals at many scattered locations with a company mainframe in Hawthorne, CA.

Recently there has been speculation that the number of suitable orbit positions over North America may be running out. Waylan points out, though, that the Federal Communications Commission (FCC) has implemented a plan to address the potential problem, mainly by reducing the spacing of satellites from 4 degrees to 2 degrees apart.

In addition, the FCC and NASA have undertaken programs that will increase the capacity of each satellite. Refinements will include the multiplying of beams, to permit using the same frequency for simultaneous retransmissions in two directions--for instance, East Coast and West Coast.

Another refinement would add a major frequency band to satellite communications. If proven feasible by NASA, a third band--the Ka, or 20-30 gigahertz band--will be added to the current C and Ku bands. (A gigahertz is one billion hertz or 1000 megahertz.)

Approaching crisis

On balance, it appears that our Information Technology industry--not long ago regarded as the paragon of US inventiveness and know-how--is rapidly sliding into trouble. Much of the industry's manufacturing base has been unable to match Japan and other foreign nations in cost-effective production. Even the crafty computer industry is headed for a trade deficit.

In the aggregate, US IT posted a $6.8-billion deficit last year. Forecasts from the American Electronics Association indicate that by the first of next year, the industry could be over $12 billion in the red.

What can the industry do about it?

One insider who offers a uniquely qualified opinion is William Taylor, leader of the Office of Advanced Systems and Software Technologies, Gould Inc, Troy, MI. An acknowledged expert in artificial intelligence, Taylor grew up in Japan as the son of American missionaries. He holds an MSEE degree from MIT, and has been involved in military and industrial electronics since 1968.

Speaking at the Assembly Engineering Conference in June last year, Taylor said the problem has nothing to do with technical capabilities, or a lack of qualified people. "The real problem is a management failure,' he stated. "Management in the electronics industries is permitting a cancer to grow in their markets which will destroy them if they don't do something about it.'

He identified the cancer as Japanese determination to "eat our lunch.' They are zeroing in on narrow markets, he said, and taking away market share less rapidly than the total market is growing. As a result, many of our companies still see sales increases, but are unaware that they are losing percent of market.

"After a while, the Japanese will have more sales power than we do,' said Taylor. "They will be larger and gradually overpower us.'

And the solution? Taylor urges that we consider the advice of a real expert on the Japanese, General Douglas MacArthur. "Never wait for the Japanese to attack you,' the general said, meaning that if you sit in a fixed position and wait for them to come, they will blow you away.

"If we move around, jump around, attack when they don't expect it,' said Taylor, "they can't do anything because they don't improvise well.

"Why haven't they taken over the personal computer market? They can't get set for it. They are about to attack Apple and IBM comes along. Bang!--20 percent of the market overnight. Then, they are about to build an IBM-compatible, and another new one comes along.

"Float like a butterfly, sting like a bee,' Taylor added. "We have to be the Harlem Globetrotters of manufacturing.'

Taylor also advised that we should be nice to our manufacturing people. "The Japanese are,' he said, "but in the US, product design engineers lord it over production engineers.'

He added that we'd also do well to tap the brainpower of factory workers. "They rent their bodies at so much an hour; they'll throw in their brains for free, if we let them. Shutting people's minds out of their work costs us their allegiance.'

Taylor ended his talk with an upbeat note: "We can win,' he said. "My Japanese friends tell me that they are worried that some day America will get its act together.'

The impending crisis in IT has been brought to the attention of the White House. In late January this year, the President's Commission on Industrial Competitiveness--after a year of study--submitted a comprehensive list of recommendations to Mr Reagan. Heading the commission was John Young, president of Hewlett-Packard Co, Palo Alto, CA.

The commission said in its printed report that we as a nation, and the IT industry in particular, need to:

Apply a higher percentage of profits to R&D.

Direct more R&D toward commercial, civilian applications.

Demand that the government create special incentives for R&D.

Put more emphasis on factory automation. "It does us little good to design state-of-the-art products, if within a short time our foreign competitors can manufacture them more cheaply,' stated the report.

Create safeguards against the theft of intellectual property for commercial purposes, "especially by the newly industrializing countries.'

Simplify regulations and procedures that slow the adoption of new products.

Create a cabinet-level Dept of Science and Technology.

Provide federally funded support for America's universities for basic research and the training of future scientists and engineers.

Invest more in development of manufacturing technology, and put a higher value on excellence in production. In past years, manufacturing engineering has suffered from low status and compensation levels, the commission noted.

Do all that we can to lower the cost of capital for manufacturing industries. "A study by the Semiconductor Industry Association shows that the Japanese success in semiconductors during the 1970s was due, in large part, to the advantage of low capital costs, not superior technology,' said the report.

Forge new understandings between labor and management, improve our ability to redeploy labor, and invest more in worker training.

Focus more on excellence in education at all levels.

Make world trade a national priority, in part through creation of a new US Dept of Trade.

Conduct trade negotiations to improve the free flow and fairness of world trade.

All in all, the report placed a great deal of the burden on corporate management and the federal government. Something the report did not dwell upon, however, may be our Number One need: a sense of urgency, a national unity, a determination to do the job. Perhaps we don't hurt enough yet; perhaps we must sink still deeper.

What will it take to get us moving?

Appliances next

Next month Survive-85 will report on the appliance-manufacturing industry.

Photo: Technicians test a new VHSIC matrixswitch chip. Transistors being developed in Phase 2 of the Dept of Defense's VHSIC program will measure less than one micron in the largest dimension. Photo courtesy Military Electronics Div, TRW Inc.

Photo: Estimated percent growth of selected US manufacturing industries-dollar value of US shipments

Forecast figures published early this year in the Dept of Commerce report, 1985 US Industrial Outlook, show the relative growth rates of four major Information Technology industries, left, and eight older industries. Since they were issued, the figures for growth in semiconductors and computers have been revised downward.

Photo: Estimated dollar values of shipments of selected US manufacturing industries

As indicated in forecasts published early this year by the Dept of Commerce, the aggregate of Information Technology industries accounts for a significant portion of US manufacturing. Also considered part of the IT grouping, but not shown here, are connectors, vacuum tubes, capacitors, resistors, coils, and transformers--all members of the Electronic Components complex (SIC 367). Growth in these five industries is predicted to be sluggish this year. Since they were published, the figures for semiconductors and computers have been revised downward. The "nec' after Electronic Components in the graph stands for "not elsewhere classified.' This category includes diverse components such as printed circuit boards, magnetic tapes, liquid crystal displays, and bubble memories.

Photo: A Cray 1-S supercomputer, installed at the NASA Lewis Research Center, Cleveland, OH. Researchers use the computer in fluid mechanics, to study the flow of gases around turbine engine components. Another use is in structural dynamics, for prediction of vibrations in structural members of turbine compressors. Photo courtesy NASA Lewis Research Center.

Photo: Electronic computing equipment --trade balances

From a strong trade surplus in 1981, the US computer-manufacturing industry has been diving rapidly towards a trade deficit. The Dept of Commerce predicts that the industry will show a negative trade balance sometime in 1986.

Photo: Automation of computer manufacturing extends to the inspection and test stations. Here a technician connects a printed circuit board to a system that can automatically test thousands of electronic circuits in seconds. Photo courtesy Hewlett-Packard Co.

Photo: Technicians check the wiring in a supercomputer module. Supercomputers break data into chunks, or "vectors,' and process a number of vectors simultaneously. Japanese and European companies, as well as over a dozen US companies, are developing supercomputers. Photo courtesy Cray Research Inc.

Photo: To boost quality and reduce costs, computer manufacturers have been automating many aspects of production. Here a robot in a Series 7565 manufacturing system automatically inserts microchips into a printed circuit board. Photo courtesy Advanced Manufacturing Systems, IBM Corp.

Photo: This matrix-switch chip is one of several types developed recently by TRW Inc, and its partner Motorola Inc, under Phase 1 of the Dept of Defense VHSIC program. Though being developed for military applications, the new chips will make it possible to design much more powerful desktop computers by the end of the decade. Photo courtesy Military Electronics Div, TRW Inc.

Photo: Preparing silicon wafers for deposition of metal. After loading wafers on trays, the operator swings them into the evaporation chamber, right. Here conductive metal is evaporated from a crucible and deposited in uniform layers on the wafers. After being cooled and removed, the wafers are masked and photo etched, leaving only the tiny conducting lines. The wafers are then cut into chips. Photo courtesy General Technology Div, IBM Corp.

Photo: Launched from French Guiana this past May by an Ariane 3 rocket, the GSTAR I communications satellite carries 16 Ku-band (12 to 14 gigahertz) transponders. The satellite will provide voice, video, and data communications service to users in all 50 states. The Dept of Commerce predicts that the satellite communications industry will grow from $2-billion sales in 1984 to about $5 billion by 1989. Drawing courtesy GTE Spacenet Corp.
COPYRIGHT 1985 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1985 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Quinlan, Joseph C.
Publication:Tooling & Production
Date:Jul 1, 1985
Previous Article:Help us help you communicate.
Next Article:LANs: manufacturing's prodigal child.

Related Articles
Tricam Wins Major Industry Report at NepCon Show.
Sage Research Introduces CPE Customer Satisfaction Benchmark Program; Study Shows Gigabit LANs Get Top Satisfaction Scores, PBXs Score Lowest.
Research and Markets: Global Technology Gaps, White Spaces and Trends in Patent Filings Uncovered.
Research and Markets: View this Analysis of Key Trends in Biopharmaceutical Manufacturing Technology.
Technology could improve health system.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters