IBM Research is celebrating 60 years of breakthroughs in computer science, physics and semiconductor design on Tuesday, as it steps up its efforts to scientifically study how organizations operate.
Originally housed in a renovated fraternity house at Columbia University, the then-named Watson Scientific Computing Laboratory has become one of the pre-eminent technological research centers in the world--and it has given IBM an edge over competitors in many fields.
Five IBM employees have won Nobel Prizes for, among other achievements, the discovery of electron tunneling and the invention of a microscope that captures images of individual atoms. Add to that seven National Medals of Technology, five National Medals of Science and four A.M. Turing Awards.
IBM inventions and discoveries include the programming language Fortran (1957), magnetic storage (1955), the relational database (1970), DRAM (dynamic random access memory) cells (1962), the RISC (reduced instruction set computer) chip architecture (1980), fractals (1967), superconductivity (1987) and the Data Encryption Standard (1974). In the last 12 years, IBM has received 29,021 patents--more than any other company or individual in the world.
And, unlike like the storied Bell Labs or Xerox PARC, IBM has striven to ensure that its research adds to the bottom line through enhanced products, services and intellectual property licensing.
"While we do exploratory stuff, we count on the research to help grow the parent," Paul Horn, senior vice president of research at IBM, said in an interview. "While Bell Labs spent a lot of money, they never really had a strong model on how research impacted the company."
The practical streak goes back to the beginning, Horn added. Although the group was started in 1945, the company didn't form it as a response to World War II or postwar reconstruction. "Thomas J. Watson Jr. felt there would be really hard problems that computers could solve," Horn said.
The organization, however, is becoming a far different place than it was back in the 1970s when IBM devised a way to use regular TV monitors as computer displays (1968) or unfurled the first speech-recognition application for computers (1971).
For one thing, scientific research is no longer as heavily concentrated in the U.S. as it once was. The number of peer-reviewed papers written outside the United States, as well as the number of citations to these papers, is rising rapidly.
"There are contributions of consequence that are occurring across the world," said Chris Murray, manager of nanoscale materials and devices. "I don't think we (the U.S.) are in a position at our current levels of investment in education to control or even have a strong influence on how innovation develops."
The direction of IBM's lab efforts is also changing. Years ago, the company primarily concentrated on pushing the frontiers of hardware. This resulted in such machines as the 1947 IBM 603 Electronic Multiplier, the first electronic calculator put into production, and the Sabre reservation system in 1962.
While IBM remains a major center for nanotechnology research, the company's push toward services and software has prompted it to dedicate more of its laboratories toward solving business process problems: supply chain management, application integration and transactional inefficiencies.
The ultimate question is, "How do people in an existing network operate?" Horn said. "We estimate that business process transactional services could become a half a trillion dollar market in the next couple of years, and the whole IT industry itself is only $1.2 trillion."
Right now, one of the basic challenges is coming up with a framework for studying these issues. "It touches on social sciences. It touches on business. It touches on economics," he said. Software programming and game theory are also crucial applications.
One of the first steps in developing a larger body of knowledge in this area occurred a few years ago when IBM began to set up supply chain management curriculum at various universities. Now, the company is working with North Carolina State to develop curriculum around what it calls SSME--social science and management engineering.
Although "services science" may sound squishy, Horn asserts that every new discipline does.
"A long time ago, people didn't think there was science in computer science. If you were a member of the IBM Academy, you were in hardware. There was no deep intellectual depth in software," he said. "Now people say the same thing about services."
The 60th anniversary celebration will take place at the T.J. Watson Research Center in Yorktown, N.Y. Speakers include Horn, Nick D'Onofrio, executive vice president of innovation and technology, Bob Dennard, inventor of the DRAM cell, and Fred Brooks from the University of North Carolina.