When companies stop making computers

Research

Business Week

January 19, 1976

In August, 1971, Nathan L. Gordon, a scientist at RCA Corp.'s sprawling corporate research center in Princeton, N.J., was heading his company's biggest R&D project. His job was the massive one of creating the basic "architecture" -- the over-all system design and programming -- for RCA's next generation of computers. But within a month RCA had decided to drop out of the computer industry, and not only Gordon but nearly 40% of the corporate research staff had little idea of what their future would be.

RCA was not alone in facing a massive R&D reorganization because of a decision to drop out of computers. When General Electric Co. quit the computer business in 1970, not only the divisions involved but about 20% of its corporate R&D programs felt the impact. And Xerox Corp., which folded its Xerox Data Systems this year, is now facing fundamental changes in priorities in its R&D efforts.

Cumulatively, the switch away from R&D spending on data processing systems by these three industrial giants is redirecting a significant portion of industrial R&D toward other fields. Together, RCA, GE, and Xerox are spending about $700 million annually on R&D related to their commercial businesses. Had they stayed in the mainframe computer race, one-third to one-half of that annual expenditure would almost certainly have gone to information system projects. But now:

Xerox is refocusing on office systems, copiers, on-demand printing techniques, and word processing.
General Electric has recast its massive corporate R&D organization to concentrate on electronics, energy, and materials.
RCA is devoting half or more of its efforts to consumer products and has strengthened its work in communications and digital electronics.

RCA: Back to the consumer

To catch up and keep up with the frantic technological pace of the computer industry during the 1960s, RCA devoted a steadily increasing proportion of its research dollars and the biggest share of its top talent to bolster its Information Systems Div. By 1971 about 35% of the company's research center budget was pouring into computer projects; in contrast, Information Systems produced only 10% of corporate revenues. So when RCA dropped computers, the impact on its R&D effort was wildly disproportionate.

At the David Sarnoff Research Center, few think of the computer era as the good old days. Says William M. Webster, RCA's vice-president of research: "Closing down the computer operation was analogous to cutting off a gangrenous leg -- it was terribly traumatic, but eventually the body got healthier."

The biggest immediate problem that RCA faced was what to do with its research talent. "There was tremendous pressure to cut back corporate research expenditures," Webster recalls. "After all, when you have just recorded a $500 million write-off, it does tend to dull the spirit of adventure."

How RCA did it

How to make the cut was a critical decision. On the basis of manpower devoted to computers, the R&D effort could have been cut by 40%. On the other hand, the computer division, assessed for corporate research on the basis of its proportion of corporate sales revenues, had contributed 10% to the research center's support. "Management chose to cut the smaller amount, I think wisely," says Webster.

Convinced that the technical staff was flexible and broadly talented, Webster managed to keep most of his people. Most of the economies came out of cutbacks in nonresearch support and administration services.

Gordon and his staff of computer designers more than survived the cuts. Recalls Gordon, "Almost instantly, Conrad [Anthony L. Conrad, the new chief executive of RCA] invited us to help other divisions with their computer applications." For some computer professionals, such an invitation might be like asking an automobile designer to help out for a while in the repair shop. But Gordon, after tidying up his group's work on fourth-generation computer architecture in preparation for turning it over to the Univac Div. of Sperry Rand Corp., marched off to apply his computer knowhow to problems at Hertz Corp., an RCA subsidiary; RCA Global Communications; and, within a year, the Consumer Products Div. and the world of television sets. This final move bloomed into a lasting affair.

A career in TV

Gordon, who now carries the more general title of director of systems research, found a whole new career in the TV side of his company's operations. "I went to the Bloomington [RCA TV assembly] plant," he says, "and did some analyses using mathematical models to find out where we could improve things like the cost effectiveness of product testing and quality control. We really began to understand the business and its problems. It's a mature field, and at that time we found a great proliferation of products -- we were trying to make everything for everyone."

The Consumer Products Div. tended to believe that RCA's Research Center was a place where abstracted scientists made obscure but patentable discoveries of little relevance to the nitty-gritty problems of low-cost mass production. But the division's engineers raised few objections to the influx of research people because they saw that the scientists could provide significant technical inputs for product improvement. It turned into far more than that. With the Research Center's participation, the Consumer Products Div. went through a basic strategy revision. "We decided that RCA should court technical leadership in color television," says Gordon.

"We went after everything we could do with color TV as a system. In the course of it, we at the Research Center became part of the color television business, a fundamental change," he says. "I'm basically a shirt-sleeves guy, and we spend a lot of time at the factory now. It's not trouble-shooting. It is the only way I know to cross-correlate manufacturing problems and engineering problems. You have to know in the lab what the basic problems and possibilities are out on the manufacturing floor."

Webster agrees. "There is practically no boundary now between Indianapolis [Consumer Products Div. headquarters] and Princeton," he says. "We may be supporting Allegheny Airlines flying people back and forth, but it looks as if we will stay tied in like this from now on." The traveling goes in both directions. Consumer Product engineers are at the labs about 30% of the time. In addition, Gordon schedules a one-hour telephone conference between his staff and the division every Friday.

The first big job

While Princeton began contributing direct technical inputs to RCA television set production several years ago, the first set designed from the ground up under the dual effort is RCA's new ColorTrak, the top-to-the-line series for 1976. The set has many computer-related features in its automatic tuning and color control circuitry, and Webster, Gordon, and others at the lab are as enthusiastic about it as new parents. Perhaps more important, they respect it as top-notch technology. If some of its features -- a remote tuning unit that resembles a hand calculator, the on-screen display of channel numbers and time, and the complete absence of controls on the set itself -- betray the computer designer's influence, so much the better. For Gordon, they hint of more computer-like innovations to come -- perhaps more logic and storage to provide such features as built-in electronic games and even message storage functions.

In addition to the changes in RCA's TV sets, the Research Center has also been the moving force behind RCA's biggest product effort since the company got out of data processing: the multimillion-dollar VideoDisc project for recorded television (BW -- Sept. 15). "If we hadn't dropped computers, we probably just couldn't have done VideoDisc in the time we've had," says Webster. "There wouldn't have been enough resources to take on such 'bet your company' projects."

General Electric: Emphasis on energy

When GE sold its computer divisions to Honeywell in 1970, the shock to those in the company's research center near Schenectady, N. Y., was not nearly so great as the one that had hit RCA's scientists after that company abandoned the computer business. The chief reason was that GE had assigned a larger share of advanced computer development work to researchers in the divisions themselves. Most computer engineers and programming staff went to Honeywell along with the manufacturing people. The corporate research center, with 1,600 on its staff, did have some computer R&D projects, and it completed the longer ones under contract to Honeywell.

"The last project, for a mass memory system, went on until last summer," says Robert H. Pry, director of GE's new Electronics Science & Engineering Group. Pry's operation, established last year, absorbed most of the researchers from the dissolved Information Sciences group at the research center.

But the Schenectady lab's priorities obviously have changed, and both the organization and he big projects of the center show the differences. Last summer, Arthur M. Buesche, vice-president of R&D, moved to concentrate all of GE's energy research programs in the new Energy Science & Engineering Group, one of the three major segments into which the lab is now divided. The energy group, says Buesche, meets the need for "an organization . . . to act as a focal point for trade-off decisions needed to deal with the proliferation of energy alternatives."

GE researchers are deep into many of those alternatives. Roland W. Schmidt, manager of the energy group, has a massive project under way to develop coal-fired, combined-cycle power plants. Such plants will burn coal for both heat and gas output. The heat drives a steam turbine-generator, and the low-Btu gas, burned in gas turbines, produces direct power plus additional heat. The aim of the GE project is to find ways to make combined-cycle plants far more efficient than conventional ones in turning coal into power -- and to ensure that the new plants will not pollute.

In a cavernous section of the lab where coal gasification segments of the combined-cycle units are under test, the coal piles and slightly sulphurous smell of coal gas contrast with a roomful of computers and the electronic monitoring and control equipment wired into the hot reactors and piping.

Both the design and testing are heavily dependent on computers in other energy projects, ranging from ultra-high-temperature gas turbine wheels that can work at 3,000F to developing the fuel cycles for experimental breeder reactors for nuclear power plants. Says Schmidt, "I wouldn't be surprised if we've invested as much in computers as in pressure vessels."
Change of focus. Like RCA, GE found no problem in redirecting its computer specialists, whether they were programmers, circuit specialists, or mathematicians. "Many projects under way could be used," says Pry, "but there was a change of focus from being a vendor of computer systems to being a user."

Thus GE has continued many projects that look a lot like the computer business. It continues to develop better communication links between minicomputers and large mainframe computers, for example, and the lab's own design for a minicomputer has been licensed to another company for manufacture. In the software field, the center still contributes some work for GE Information Systems, the time-sharing subsidiary. And it has many projects involving applications programs, since the company's work in design automation, aircraft engine design, and nuclear power involves extremely complex computer codes, including special programming languages and cross-compilers. To control costs, the center is trying to develop more formal disciplines for programming. "Software has been mostly an art form," says Pry, "and that has made it hard to manage."

The fundamental technology of computers, based on integrated electronic circuits, now permeates the products of most GE divisions, especially those in aerospace and control systems, so work in the field has continued, too. Pry's Electronics Science & Engineering Group now blends into one organization both digital and traditional analog technologies, as well as software development.

"The boundaries between the disciplines got very fuzzy," says Pry. "There are a lot of trade-offs between hardware and software in electronics now, and it makes sense to have them under one organizational roof."

Xerox: Automating the office

Unlike RCA and GE, which have been through the post-computer rough spots, Xerox's R&D activities, involving 5,500 to 6,000 employees in basic and applied product development work, still face some changes. "My concern is a people concern," says Jacob E. Goldman, group vice-president and chief scientist. "It is a little early to talk about it."

About 450 Xerox employees were directly involved with design and development of mainframe computers in Xerox Data Systems. For them, the uncertainty has been a wracking experience.

Xerox's move to drop computers came as a shocking surprise to its employees, despite rumors that XDS was in trouble. At the time of the announcement, no buyer for the operation was in sight, and XDS people tagged themselves with the macabre nickname of "the undertakers." On Dec. 30, Xerox and Honeywell signed the formal agreement for transferring marketing and service operations, and final arrangements for transfer of personnel and projects should be completed by February, according to Xerox spokesmen.

Market goals

Goldman expects to be able to keep most of the technical talent in R&D because most of Xerox's present and planned products incorporate computer-type logic and memory circuits. He does not expect any change in the company's major market goals. "The secretary-executive combination is still one of the most undercapitalized areas in business," he points out.

Xerox's major strategy remains focused on word processing, document reproduction, and communication markets. "We are in a growing segment of business -- office information technology," says Goldman. "In general, it utilizes a technical effort very similar to that of computers."

The company's main development thrust will not change, Goldman believes. The company plans to use more digital techniques in office equipment to take advantage of the falling cost of digital logic and memory, and the development of better "marking engines," as Goldman phrases it, is still a major R&D goal. The effect of discontinuing computers, Goldman concludes, "will put only a small notch in [the growth curve for] R&D expenditures," which now run in the range of $200 million per year.

So far, there have been very few changes at Xerox's two main research centers in Palo Alto, Calif., and Rochester, N.Y. At these centers, which are staffed by about 400 researchers, Xerox conducts a wide range of projects not tied in to product development schedules, ranging from studies of photo-electric materials to human factors studies on office systems.

Good health

The three giant companies that have dropped computers have, in effect, changed the marching orders of a sizable army of more than 4,000 people employed in applied and basic research. The major results of the change are only beginning to show in the products of GE and RCA, and it may be several years before other research projects now under way show a significant impact on the sales and profits of their sponsors.

RCA's Webster believes that the post-computer changes in research priorities bode well for the whole economy. "In research, I see our responsibility as keeping both the company and the industry we are in healthy," he says. "That means at times you have to look a little beyond this year's profit-and-loss statement because it is certain that you can't work profitably in a sick industry." If R&D is as good medicine as Webster believes, then the three giants that have dropped mainframe computers will be contributing big doses for the health of U.S. business in typewriters, television, and turbines.

GRAPHIC: Picture 1, RCA's Nathan Gordon and William Webster, vice-president for research, put color TV high on the R&D priority list, and the ColorTrak was the result. Grant Compton -- BW; Pictures 2 and 3, At GE, Buesche recast the labs into energy, materials, and electronics sections, with Pry in charge of electronics.; Picture 4, Xerox's Goldman: Office automation is still the basic thrust of company R&D.