by Richard Cole
"Working in so many areas--that's been the beauty of my experience in the industry."
Business Advantage: "You might say that being a father has helped me in business. I can sometimes see a great deal of similarity between negotiating with my five and seven year-olds and negotiating with business divisions."
In many productive lives, nothing is wasted, and this fact is especially true for Dennis Roberson's career in technology. In a recent interview, he provided an example by recalling his days as a young electrical engineering student at Washington State University. At the time, he was working extensively with lasers (dubbing himself the department's "Head Laserologist"). Although his interests soon moved from physics to other areas, his laser background now informs his understanding of NCR's bar-code scanners, a critical product area which he helps to develop as the company's CTO.
With a background that includes both computers and telecommunications, hardware and software, design, manufacturing, research and productization, Roberson finds that he can leverage experience in one area to better understand and manage projects in another. This ability is critical in his current role. He and his office serve as a review body for all research and development at NCR. They also act as spokespersons on matters of technology directions and industry trends. Roberson works comfortably with many different divisions that might not speak the same dialect or even the same language, making sure that their activities are aligned with NCR's corporate directions and strategies.
Roberson planned to stay at AT&T after graduation, but cutbacks in the NASA space programs in the early 1970s meant that Bell Labs stopped hiring. Instead, he joined IBM in 1971 as a logic design engineer for their Advanced Systems Development Division. The division was involved in systems concepts and applications, exactly what Roberson wanted to study. The division was also located in the San Francisco Bay Area, which allowed him to pursue postgraduate studies at Stanford University.
Roberson remembers his first years at Big Blue as a "really wild, free-flowing era." Bright engineers were given the run of the labs to create first and worry about production or market practicalities later. On one project, he and his colleagues decided to design one of the first multimedia systems from scratch, building or adapting all the components themselves. The display technology was built out of fixed-head disks refreshing the displays. Scanners were built out of modified office copiers. "It was a wonderful way to work," Roberson says. "Nothing was dismissed as impossible, so people tried an amazing variety of things. They sometimes succeeded, sometimes failed, but it was always very stimulating." Roberson adds that, given a willingness to work hard and to physically move around the country, a young engineer like him could rise rapidly through the organization.
Eventually the temperature rose above zero, and Roberson continued work which he still describes as one of the highlights of his career. His group developed a number of desktop computers: the 5110, the 5120, the System 23 (which was the first Intel processor-driven machine) and finally the 5150, which became known as the IBM Personal Computer or PC.
Interestingly, Roberson says that his first vision of a PC was as a laptop. Even the first 5100 desktop unit was meant to be portable, although the machine weighed in at 55 pounds and required a sturdy case with a steel handle to transport. Roberson claims that he carried prototypes of the 5100 home on a regular basis, although he concedes that it was a feat that only the "physically strong" should attempt.
Roberson became a successful evangelist for the IBM PC, promoting it in speeches around the country as the "portable personal computer for the professional problem solver." He says with a smile, "If I could get past that first tongue-twister, I could get through the rest of the presentation with any audience." His stay in Rochester also incurred another major benefit: he met his wife, Debra Lin at a local church. Their marriage has resulted in four sons and a daughter, who play "every sport you can imagine."
In 1985, he became director of logic programs at IBM's microprocessor facility at Burlington, VT. He was put in charge of chip design, semiconductor process development, test engineering, packaging and the memory subsystem for IBM's 3090 mainframe. Once again, his varied background came in handy, especially his academic work in solid-state physics. "Chip manufacturing is essentially a heavy physics area," he explains. "It's a well-understood and well-controlled environment, and if you can handle the physical aspects in a very disciplined way, you can make steady progress."
Progress was exactly what the group made, introducing in early 1986 a 1Mbit memory chip years before any such development was expected. With a chuckle, Roberson says, "The Japanese thought they owned the DRAM business. It was great fun proving them wrong." When the chip came out, his group followed the reaction in Japan through the newspapers. "We read about leaders of the Japanese semiconductor industry being called home from vacation and everything. They were taken completely by surprise." His group also produced the first CMOS chip by IBM as well as IBM's first RISC processors.
Roberson joined Digital Equipment Corp. in 1988 as a vice president working with software development groups. In 1994, he joined NCR, then a part of AT&T's Global Information Solutions company. The move to a telecommunications environment was another big shift.
Roberson points out major cultural differences between the computer and telecommunications industries. In telecom, he says, the technology is based on a high degree of connectedness. Standards are very important and consist of more than "just checking off boxes." Because standards are stressed in a rigorous and formal manner with required certification by bodies like the International Standards Organization, technical progress can be slower than in other industries such as computing.
As an example of the deliberate pace--at least in the past--of the telecom industry, Roberson points out that his digital display telephone took seven years to appear on the market. "On the other hand," he says, "you get a high degree of reliability. In fact, 99.999 percent reliability is common for telecom." In contrast, the computer industry has often had to trade off degrees of reliability with new technical development, although, as he remarks, the differences between the two industries are growing smaller.
Roberson is excited about the recent merger of OSF and X/Open into The Open Group. He says that the merger will help accelerate the process of delivering open system standards to customers and better focus the efforts of the industry to bring new products to market.
For the future, Roberson looks forward to working with open systems and NCR in the company's retail, financial and telecommunications markets--that, and getting in a bit more waterskiing at Lake Murray.