Dan Sears '74

Virtuous Reality

Nobody's got connections like Fred Brooks. Some of the computer's most basic functions
are part of his legacy — that, and UNC's prestigious
computer science department

by Kevin O'Kelley '92 (MA)


Georgia Parker/UNC Design Services
 

In Fred Brooks' office is a metal box with six switches. Above each switch is a label: "COM," "CLASS," "MY MEETING," "O MEETING," "EVAL," "SOLO TECH." On the shelf above the box sit six clocks, each one set to a different time.

It's Brooks' way of keeping track.

"Every Monday we set the clocks to 12," he explained, as he looked at each clock. "Now you can see this week that we've had 17 hours and 20 minutes of administrative work and two hours and 15 minutes of teaching and two hours and 20 minutes of research, and no general University committee work and no school work and no teaching."

He turned, sat down and smiled.

"The best way to find out where it goes is to know where it went."

The work and thought of Frederick Brooks—the founder of UNC's computer science department and one of the fathers of modern computing—seem governed by an acute sense of interconnection. A slice of a day, a single computer design project, even the discipline of computer science itself—each ultimately is judged by its place in a larger schema.

There are no throwaway moments. Everything is for keeps.

Brooks has many accomplishments to his name. His work on three-dimensional computer graphics has driven the technology's development, as well as establishing UNC as a world leader in the field. He wrote what is now considered the definitive work on managing software development projects, The Mythical Man-Month, which still is in print after 25 years.

The citations of his innumerable awards—among them the 1999 A.M. Turing Award ("The Nobel Prize of Computing"), the 1995 Bower Award and the 1985 National Medal of Technology—often describe his work in vague, albeit impressive terms.

Writers—and readers—crave tangibles: Crediting a scientist with the sole invention of a specific object has a satisfying completeness. But Brooks' accomplishments as a computer scientist come as much from his capacity to work with others as it does with his gift for individual research.

Furthermore, his accomplishments are difficult to describe—and that difficulty testifies to the magnitude of his work: To put it bluntly, contemporary personal computers are unimaginable without the achievements of Fred Brooks.

Many of the features we take for granted in personal computers—their use for word processing, the capacity to call up one of hundreds of programs in a fraction of a second, the capacity to load a software package on any number of different computers—are all traceable to the work of Brooks and to his sense of connection.

'Absolutely fascinating'
To grasp Brooks' role in the development of computers, it helps to understand how new computer technology was when his career began.

He still remembers the day—in August 1944—he learned what a computer was.

"I was 13 years old and growing up in Greenville, North Carolina. I was sitting in the Sheppard Memorial Library and reading, I believe, Time."

He was reading about the dedication of the Mark I—the first "universal calculator," a machine that could perform mathematical computation without a human operator.

"It was a whole new concept to me, and absolutely fascinating."

He had always wanted to be a scientist, and when the teenage Brooks went to Duke he studied physics, which had been his favorite high school subject.

But he never forgot his fascination with the Mark I. When it came time to pursue graduate work, he went to Harvard to study with the man who designed it, Howard Aiken.

"Aiken had in the computation laboratory at Harvard one of the handful of programs in what we would today call computer science," Brooks said. "There it was called applied mathematics, but it was indistinguishable from today's computer science.

"At the time I went to graduate school," Brooks added, "there were no computer science departments."

Brooks' work at Harvard was a fitting prelude to his later work ushering in the age of the general use computer.

Aiken wanted his students to do dissertation research on the non-scientific uses of computers. He encouraged Brooks to design a computer that could do payrolls. "So I designed a payroll computer," he says matter-of-factly, "and concluded that there was very little to gain from specializing a computer for payroll.

"There was a lot to gain from specializing a computer for file maintenance applications—so the same computer could do utility billing and insurance premium accounting and payroll and Social Security record keeping and so forth all equally well."

Some journalists who have written about Brooks assert that this project led to his job with IBM. Brooks demurs. "No. The education led to the job." He clearly considers it an important distinction.

IBM recruited Brooks in 1956 to work on the Stretch computer, one of the first supercomputers. "We built the first machine for Los Alamos laboratory. We built the second for the National Security Agency—and for that we built an attachment called Harvest, designed for language processing."

He starts to smile.

"Stretch was about five feet high, five feet deep and 15 feet long. Harvest added another 20 feet."

The S/360
In 1961, Fred Brooks was on the losing team in a long and brutal fight at IBM.

"We had a set of products coming along called the 8000 series. [It was] a higher-end family of computers."

An IBM vice president, Vincent Learson, thought the company should delay introduction of the 8000 series and instead focus on producing a more general purpose computer that could be sold to both higher- and lower-end markets as well as use the recently developed integrated circuit technology, Brooks said.

To develop this new computer, Learson brought in one Bob Evans as machine development manager to work out a plan for producing it.

Brooks led the fight to focus company resources on putting the 8000 on the market as soon as possible. The struggle lasted six months. The 8000 was put on hold.

Then Brooks got a surprise. Evans asked him to lead the design and production team for the new general-purpose computer.


IBM Corporate Archives
The IBM S/360 has been called one of the most important technological developments of the 20th century. Brooks led the design and production team for the new general-purpose computer. It hit the market in April 1965. Many IBM salesmen sold their year's quota within a few hours.

 

"You could have knocked me over with a feather," Brooks laughs now. "When Evans called me and asked me to take over his crown jewels after we had been fighting for six months, I was absolutely astounded."

The team Brooks assembled developed the computer, which became known as the IBM S/360 and has been called one of the most important technological developments of the 20th century.

It hit the market in April 1965. Many IBM salesmen sold their year's quota within a few hours.

When pressed, Brooks acknowledges what his team accomplished in producing the IBM S/360. "Today's general purpose computers are the result of putting together the evolution of the scientific computing and business computing strands. We did that in the IBM S/360."

But even that doesn't do justice to the enormity—let alone the complexity—of what he and his colleagues accomplished.

When the 360 was in development, Brooks insisted the operating system employ eight-bit bytes. He always had been interested in the computer's potential for what we now call word processing. With eight-bit bytes (as opposed to six-bit bytes, which many previous computers had used) a computer could run software that produced both upper and lower-case characters.

(A tale has made the rounds that Brooks not only worked with bytes but coined the word itself. Not true. It was Werner Bucholz, chief architect of the Stretch computer.)

Brooks' team also included disks—as opposed to just tapes—in each computer. This one step endowed the IBM S/360 with power and speed unprecedented in a general use computer.

"[Having a disk] meant you could have a big library of programs and call [one of] them up in a fraction of a second rather than having to run them in a pre-planned sequence [on a tape]."

Another of his achievements altered the nature of both computer use and the industry itself. "What we did was essentially make six different computers, all what we would call compatible. It means that two machines will run the same program and get the same answer."

What it also meant was that software and hardware became separate production processes—and ultimately, different industries.

When Brooks won the Franklin Institute's Bower Award in 1995, the citation made it clear how revolutionary the development of the IBM S/360 was. The citation credits Brooks with defining "a concept of computer architecture that separated computer software from hardware allowing these two fundamental realms of the computer age to develop dynamically and independently."

In other words, no IBM S/360, no software industry. No Adobe, no MicroSoft, no RedHat.

'I was led by very clear steps'
After shepherding the 360 into actuality, Brooks had an extraordinarily bright future at IBM. But in 1965, he was invited to come to UNC and found the University's computer science department.

As difficult as the decision was, he decided to accept.

"I'm a Christian, and I was led by very clear steps to make this shift in life," Brooks said. "I had always been teaching. I like to teach. While I was at IBM I taught one year at Vassar and one at Columbia, and I had maintained a steady publication schedule."

How did he find the time to do that while laying the foundations of an IBM career?

"Nights and weekends," he laughed.

He and his wife also had taken an extra measure to make it easier to move to academic life if the opportunity arose. They didn't want to get used to the expensive lifestyle that comes with a corporate career. His wife worked for IBM as well—so they lived on her salary and saved his.

The offer was even more attractive, coming from UNC. "My kid's grandparents were alive in Greenville. And any high schooler growing up in North Carolina made a lot of trips to Chapel Hill for scholastic debates and music festivals."

In addition to his natural inclination to teach, and the attraction of living in his native state, Brooks felt the time was right to found a department devoted purely to computer science.

"I was convinced that computer science had reached the point where making it an independent discipline made sense." He saw a clear need for the conduct of research and the design of curricula "that would be judged by the standards of computer science and not another discipline.

"This was the opportunity to do that."

'Out here in the backwoods'
Chapel Hill and its larger neighbors were distinctly different towns when Fred Brooks arrived at UNC. The very idea of a "Triangle" was only a decade old. But the area's future as a center of research was evident. In 1965, both IBM and the predecessor of the Environmental Protection Agency opened facilities in the seven-year-old Research Triangle Park. Brooks realized that the future of computer science at UNC was inseparable from that of the newly conceived "Research Triangle."

When he arrived at UNC, IBM was prepared to grant $100,000 to the new computer science department. He had a better idea: Grant only a third of the money to UNC. He asked the company to give one-third to N.C. State and a third to Duke, to help them establish curricula in computer science.

"I'm still amazed [he did that]" says Lib Moore Jones, the first secretary of the computer science department. Jones, who was departmental secretary for 14 years, believes the act ultimately led to the establishment of the Triangle Universities Computation Center in 1965. Until closed in 1990, the center, a joint venture of the three universities, provided mainframe computing services to all three as well as the Research Triangle Institute.

"That was the money that made the difference and fostered a sense of relationship between all three," Jones said.

Brooks dismisses the act as simple pragmatism: "It's very clear today—as it was then—that the only way an institution out here in the backwoods of North Carolina can achieve excellence on the scale of the Harvards and MITs was to do it jointly; so building that team together was of crucial importance.

"That's the major component of any strategy in the Triangle today."

Seeing and Feeling

In virtual reality (VR) experiments, people wearing special headsets are able to "see" an environment. This is especially helpful in designing new environments, whether the floor plan of a building or the layout of a warship. A UNC research project not only allows you to "see" what an environment looks like but also what it "feels" like. Hybrid reality allows users to actually feel and bump into the objects seen in the virtual environment. Researchers hope to find out if virtual environments seem more real when users cannot walk through virtual walls or solid objects, as they have been able to do in most VR environments.

UNC researchers set up a model kitchen made of styrofoam, simulating countertops, work islands, a sink and other features of a kitchen. Wearing the VR headset, a user sees a synthetic visual kitchen, which replicates, with half-inch accuracy, an actual kitchen. However, the visual model and the styrofoam model have been carefully aligned so that when the user reaches out to touch something such as a countertop corner, the corner in the styrofoam model is where the visual information indicates it will be.

A reasercher moves around the styrofoam model kitchen (above) while seeing the synthetic visual model (above right) through the headset. Both models are based on an actual kitchen (below right). Photos by Larry Ketchum

 

 

Alien images
The popular image of computer scientists depicts them as delighting in technology for its own sake—like the MIT graduate students who play the hand-to-hand combat game DOOM on the Media Lab's floor-to-ceiling computer screen.

Our cultural imagination also toys with the notion of a self-sufficient artificial intelligence—a computer with all the cognitive capacities of a human mind.

Nothing could be more alien to the man who has spent 35 years developing virtual reality.

"Computers are tools," Brooks said. "And the question you ask when you have a tool is, 'What can you do with it?'"

The notion of computers as tools is critical to understanding Brooks' belief in the limitations of artificial intelligence.

He prefers to work toward what he calls the goal of intelligence amplification.

In 1969, after Brooks decided to make computer graphics a focal area for the UNC computer science department (along with computer architecture, software engineering and natural language processing) he arranged a meeting with the provost, Charles J. Morrow. He told Morrow he had a computer with substantial graphics capabilities and a very capable graduate assistant. He added that he was interested in determining the capacity of computer graphics to solve research problems.

He then asked Morrow a question: "Who on the faculty most deserves to have his intelligence amplified?"

Morrow came up with a list of faculty and projects. Brooks chose biochemistry professor Jan Hermanns, who was pursuing the problem of protein-folding—the still unsolved mystery of how ribosomes bind together into amino acids to form proteins, the basic building material of organic cells.

Working with Hermanns bore an unexpected dividend. The biochemist introduced Brooks to Duke University scientists Sung-Hou Kim and David and Jane Richardson, who were working on another, more soluble project: the development of a system that could derive the three-dimensional structure of a molecule, given crystallographic data from X-rays.

That work, Brooks said, resulted in findings used both by "academic labs studying protein structure and pharmaceutical laboratories designing drugs."

Brooks' pragmatic approach to computer graphics and virtual reality has set the tone for the department's work in those areas for the past 30 years.

"I had a choice of some places much better known than here [when I was looking for a job]," said Professor Henry Fuchs, whose research interests include using ultrasound imaging to enhance the effectiveness—and reduce the invasiveness—of surgery.

"But I immediately felt in Fred a kinship in our approach to the field, and I suspect that most other people who came here did so for similar reasons."

Indeed, in an attempt to understand how UNC's computer science department has achieved its current eminence, it's difficult to underestimate the influence of Fred Brooks.

"Like most of the people here [in computer science], I get offers from other places, places with more prestige, with more support, in more exotic locations," Fuchs said. "And I'm still here after 22 years. And a large reason for that is Fred, who established and continues to influence this department so it has a certain magic to it."

"It's a place where computer scientists build tools for people, without any apology, in which people are free to work on the same problems for years even though they may not be fashionable."

The value of that philosophy is evident in the strides in computer graphics (of which virtual reality is a sub-field) made in the past decade.

When Brooks set out to make computer graphics a departmental focus, few of his colleagues thought it was important, Fuchs said. But he pursued it, in part, because of a speech he heard at the 1965 meeting of the International Federation of Information Processing Societies. Ivan Sutherland, the inventor of Sketchpad (the first graphics drawing program), exhorted his colleagues to pursue computer graphics research.

Sutherland urged his colleagues to think of the computer screen as a window and to work to make what they saw through that window "look real, sound real ... feel real."

It was a speech, Brooks recalls, "that wonderfully fired the imagination."

And it led to what we now know as virtual reality.

The much-hyped virtual reality essentially is a process by which a computer-generated environment is substituted for an actual one "either by a head-mounted display, which occludes the real world and substitutes stereo images, or by means of a cubicle of screens on which you project a virtual world that surrounds you."

"That's the key," Brooks said, "this immersion.

"And you should be able to interact with objects in the virtual world.

"So if you 'grab'—dangerous word—a virtual object, the object should 'move' as if you were holding it."

Brooks, however, dislikes the term virtual reality. He prefers "synthetic environments." But he smiles a little sadly as he says it.

"I'm afraid it [virtual reality] has stuck."

Thirty years after Sutherland's speech, Brooks' faith in computer graphics and virtual reality as worthy pursuits clearly has been vindicated.

Virtual reality has gone from being a distant goal of long-term research to a real technology with proliferating applications.

And in the past decade, The New York Times, New Scientist, Business Week, The Economist and Scientific American all have covered computer graphics projects at UNC.

Today these projects often are collaborations between UNC computer scientists and researchers at Harvard, Johns Hopkins and other institutions—as well as faculty in other UNC divisions such as the medical school.

Many industries use virtual reality for vehicle simulations, Brooks notes—simulating airplane flight, truck driving, ship piloting.

Among their other projects, Brooks and Fuchs currently research ways to use virtual environments to train people for dangerous occupations, such as fire fighting or military operations.

"We are also working on ways of measuring precisely when the VR illusion is working, and how well, by measuring bodily functions such as heart rate and increase and palm sweat in a stressful virtual situation," Brooks added.

Brooks' work contrasts starkly with the popular image of virtual reality as an entertainment medium. When asked about the increasing use of virtual reality for games, he said, "I think it's inevitable."

Then he paused. "It's not something I feel particularly called upon to spend my time on."

'I love what I do'
In addition to his research work, Brooks usually teaches two classes a year. In his department, he notes, two or three classes is a standard load for professors with active research programs.

Unlike some faculty who are accomplished researchers, Brooks doesn't begrudge any time he spends teaching.

And it's fair to add that his teaching, as well as his research, has had far-reaching effects on modern computer design. One of his former students, John H. Crawford '77 (MS), was the chief architect of the Intel 386, 486 and Pentium processors.

He does, he admits, "feel torn" about all the committee work that "takes more than half my time"—for both the University and national scientific organizations. Committee work, however, comes with seniority, he says.

Committee work or no, he has no intention of retiring.

"I love what I do. I can't think of anything I'd rather do than what I do."


Dan Sears '74
In Brooks' office is a metal box with six switches. On the shelf above the box sit six clocks, each one set to a different time. He resets the clocks every Monday. It's Brooks' way of tracking how he spends his time.



Kevin O'Kelly '92 (MA, '00 MSLS) is a writer based in Carrboro and a columnist for The Chapel Hill News.

Article first published in the Carolina Alumni Review, 90(1), Jan/Feb 2001, 22-30. For more information about the Review, visit the UNC General Alumni Association's home page. Presented here by kind permission of the Carolina Alumni Review.