Chuck Thacker Attains Computing’s Peak
By Rob Knies, Managing Editor, Microsoft Research
When Chuck Thacker graduated from the University of California, Berkeley, with a bachelor’s degree in physics in 1967, he envisioned a career as an engineering physicist, designing particle accelerators.
Things didn’t progress according to plan.
Thacker entered the workaday world to make money to pay for graduate school. Before long, he found himself hired as a staff engineer for a computer-research project based at his alma mater, and he soon found himself mesmerized by computing. As he jokes, “I fell in with bad companions.”
Forty-two years later, Thacker, a technical fellow with Microsoft Research Silicon Valley, has ascended to the peak of his accidental profession, being honored March 9 as the 2009 winner of the Association for Computing Machinery’s highest accolade, the A.M. Turing Award, given for contributions of lasting and major technical importance to the computer field.
The award, which is accompanied by a prize of $250,000, generally goes to computer scientists noted for conceptual or theoretical work. Thacker, 67, becomes just the second person to receive a Turing Award, widely regarded as the Nobel Prize of computing, for contributions in designing and building computer machinery, following Britain’s Maurice Wilkes, the 1967 recipient.
“I was extremely surprised,” Thacker said. “I never expected to win this one. There are several other nice awards that I’ve won that I thought were within the realm of possibility, but this one I never even thought was possible.”
Others were not quite as surprised, such as Butler Lampson, a Microsoft Research New England technical fellow and a Turing Award winner himself in 1992. Lampson, a longtime collaborator, nominated Thacker for the award.
‘An Engineer’s Engineer’
“Chuck is surely one of the most distinguished computer-systems engineers in the history of the field,” Lampson said in his nomination letter. “Chuck is an engineer’s engineer. His skills span the full range, from analog-circuit and power-supply design through logic design, processor and network architecture, system software, languages, and applications as varied as CAD and electronic books, all the way to user-interface design.”
Thacker becomes the fourth computer scientist from Microsoft Research to have won the award, joining Lampson; Tony Hoare, who won in 1980; and the late Jim Gray, who took the 1998 honor.
Rick Rashid, senior vice president of Microsoft Research, acknowledged the scope of Thacker’s influence.
“Many people in the field of computing today owe the path of their careers to Chuck—myself included,” Rashid said. “As a graduate student at the University of Rochester in New York, I began using a Xerox Alto and the Ethernet in 1975, and that led directly to the research in operating systems and distributed computing that has defined my life.”
Thacker—son of an engineer, father of daughters Christine and Katherine, and for 46 years married to Karen, with whom he lives in Palo Alto, Calif.—is best known for his pioneering work at Xerox PARC as the chief designer for the groundbreaking Alto, the predecessor of modern-day personal computers.
But, as Lampson noted, Thacker’s contributions range far and wide. He is co-inventor—with Robert Metcalfe, David Boggs, and Lampson—of the Ethernet and has achieved many other novel, successful networking explorations. His Firefly project for the Digital Equipment Corp. (DEC) in the mid-1980s produced the first multiprocessor workstation and remains relevant in today’s multicore programming environment. He has had a long, productive interest in computer architecture. And his work in tablet computing led to the prototype of Microsoft’s first Tablet PC.
A Life Spent in Labs
Along the way, he became a founding member of three major research labs: Xerox PARC in 1971, DEC’s Systems Research Center in 1984, and Microsoft Research Cambridge in 1997.
He never did, however, make it to graduate school.
“I had a very odd initial start of my career,” Thacker said a few days before the Turing Award was announced, “because I do not have a Ph.D.”
Having graduated from Berkeley and needing cash to continue his studies, he took a job with Jac Hawley, a Berkeley inventor. Thacker had worked with Hawley earlier at Berkeley Instruments, a company that made digital automated weather stations.
“I made him a deal,” recalled Thacker, a former ham-radio operator. “I said: ‘Jack, I will take a year or so and work for you. You don’t have to pay me too much, but you have to teach me how to use your machine shop, and in return for that, I will design your electronics for you.”
Hawley accepted that offer, and Thacker found himself in a builder’s paradise. As he told Al Kossow for a Computer History Museum oral history in 2007, “He had a good lathe and a good milling machine and a good drill and all the surrounding tools that you need to build mechanical stuff, so I learned how to use all that and became a fairly good machinist.”
A few months later, a friend of Hawley’s dropped by the shop and told Thacker about an opening at Berkeley for the Project Genie computer-research effort. He got in touch and hired on, but not because he had any particular affinity for computing.
“I had used computers—I had programmed them—in the course of studying physics,” Thacker said. “I’d always thought that was not very interesting, but when I actually got closely and deeply exposed to what was going on, I was seduced.”
During Project Genie, he also was introduced to seminal figures in computing history, people such as Lampson and Peter Deutsch, Mel Pirtle and Wayne Lichtenberger. They were interested in building a successor to the SDS 940, one of the first successful commercial time-sharing machines. But the effort was too big to gain funding as a university project, so the Berkeley contingent started shopping around.
“We had to go to New York for venture capital,” Thacker remembered. “You didn’t go to Palo Alto then.”
Funding secured, they founded the Berkeley Computer Corp. (BCC) and built a large, working time-sharing system, the BCC-500, in 1969. But the project lost momentum when investment-banking capital evaporated during the 1970 recession. The machine was sold to the U.S. Department of Defense’s Advanced Research Projects Agency (ARPA), which installed it at the University of Hawaii, where it was used as the school’s primary computer for several years.
“At that time,” Thacker said, “it was possible to know almost everybody in computing, and the guy who knew everybody was Bob Taylor, head of the ARPA Information Processing Techniques Office. Bob had just been hired by George Pake, who was setting up the Xerox Palo Alto Research Center (PARC).”
Taylor and Pake knew of BCC and came to Berkeley to interview the staff.
“They hired some of us,” Thacker said, “and that’s the way the PARC Computer Science Laboratory got started.”
The facility, which opened in 1970, became legendary for its computing innovations, and in early 1971, Thacker became one of its first hires. The research pursued by Xerox PARC bore fruit that shaped the future of computing: the Ethernet, laser printing, and, of course, the Alto.
“We were lucky,” Thacker told Kossow, “because we could work on new things … because we needed them. The motivation of being able to build things that you actually want is extremely high.”
The Alto pioneered the use of a bitmapped display that, in conjunction with a mouse, offered a graphical user interface that remains in use today on more than a billion computers around the world. The UI also enabled WYSIWYG—“what you see is what you get”—word processing, and the Alto was designed to be self-sufficient, with storage and computing achieved locally. In addition, Gary Starkweather, with help from Thacker and others, developed the laser printer, which was too expensive to give to everyone at the lab, necessitating a need for networking that resulted in the development of the Ethernet.
“It was very intense,” Thacker recalled. “We previously had built a time-sharing system we wanted to have for the computer-science lab at PARC, a machine more like the machine that the rest of the ARPANET community used, a PDP-10. Unfortunately, Xerox had just bought a computer company [Scientific Data Systems], and they wanted us to use an SDS Sigma V. We looked at the machine, and it was just not good for time sharing. We wanted to buy a PDP-10, but it would have been very unseemly had we placed an order with the company’s primary computer competitor, so we decided to build one.”
That they did, over the next 18 months, and the result was MAXC, the Multiple Access Xerox Computer.
Building the Alto
“When we got done with the MAXC system,” Thacker said, “we were looking around for things to do, and Bob Taylor had been trying to convince us to build a personal machine, but he couldn’t quite explain it to the point where we understood what he was talking about. But then a couple of things happened. No. 1, I figured out a way to make a very inexpensive machine, by the standards of even minicomputers of the day. The other thing was that we figured out the bitmap display.
“That came about because of the introduction of semiconductor memory. MAXC had used the very first available semiconductor memory. We had these nice little boards and this very inexpensive memory—it cost less than a tenth of a cent a bit. That meant you could afford to use the memory as backing for the display.”
The cost engineering was, indeed, one of the design choices Thacker made that helped to make the PC a practical reality.
“The normal way a computer works,” he said, “is there are a central processor and the memory, and then there’s input/output, and the input/output controllers always run the costs way up. It’s your graphics card and your Ethernet controller and all that. So I said, ‘Let’s time-share the processor at the lowest level.’ This is what people now would call multithreading. That made the cost much lower, so we could afford to build one of these things for every lab member. That worked out extremely well.”
And networking was soon to follow.
“The Ethernet didn’t come in immediately,” Thacker said. “It lagged by about six months. The signaling part of the Ethernet I figured out; that was the realm of electrical engineering. But all the other stuff, the packet format and the protocols and so on, that was Metcalfe and Boggs and, to some extent, Butler.”
The Alto was never released commercially, but it was deployed extensively within Xerox and found its way into a number of universities—and into Jimmy Carter’s White House. Its influence has been pervasive; concepts first revealed in the Alto are used daily on hundreds of millions of computers worldwide.
“It’s very gratifying,” Thacker smiled. “When people say, ‘What have you done for Microsoft lately?’ I say: ‘You don’t understand. The most impact I’ve had on Microsoft was work that was done before Microsoft even existed, when Bill [Gates] was in short pants.’
“PCs didn’t get to be as good as the Alto for about 12 years. It took a lot of work on software and a lot of hardware progress before they got on a par with scientific workstations. But they had one major advantage, and that was that Moore’s Law was on their side.”
From PARC to DEC to Microsoft
In 1983, after 13 years at Xerox PARC, Taylor left after a disagreement with management, taking several of his key personnel, including Thacker, with him. They found a new home at DEC, founding the Systems Research Center.
“Curiously enough,” Thacker said, “I stayed there for 13 years, too.”
While at DEC, he helped construct the first computer to use the DEC Alpha system, a machine credited with advancing Alpha’s emergence onto the market by at least a year.
“That was worth a lot of money,” Thacker said. “Nobody ever made money on the Alto, but that made a lot of money.”
He also led hardware development on Firefly, which enabled years of system and applications research both within DEC and in academia. And he pursued networking development of the AN1 and AN2 systems and a high-performance, high-bandwidth networking project that became a product called GIGAswitch/ATM.
But in the mid-‘80s and early ‘90s, Thacker also found himself in discussions with Microsoft. He had preliminary talks with company personnel—including, on various occasions, Gates, Lampson, Gray, and Rashid—in 1983 and again in 1991. But Thacker thought of himself as a hardware guy and didn’t see an appropriate fit at Microsoft.
But in 1997, he noted, “They finally found the right button.
“I was preparing to leave DEC,” he explained. “Our kids were grown and gone by that time, and my wife and I figured that we would take a year and do a sabbatical, either at the Computer Laboratory at Cambridge or at ETH in Zurich. I knew a lot of people at both places.”
Thacker was pitching the idea of a Microsoft Research lab in Silicon Valley at the same time that his colleague, the late Roger Needham, was lobbying for a Microsoft Research facility in Cambridge, U.K. The latter won out.
“I sent Roger an e-mail and said, ‘Congratulations, I’m glad to hear your lab was a success,’ ” Thacker remembered. “Roger was an extremely cryptic guy, and he sent me back mail in a few minutes that said, ‘Well, yes, it does seem to have happened, and you’ll hear more about this shortly.’
“Thirty minutes later, the phone rang, and it was Nathan Myhrvold [founder of Microsoft Research and then Microsoft’s chief technology officer]. He said, ‘Chuck, how would you like to come to Microsoft and take a two-year appointment to help Roger set up the lab in Cambridge?’
“I said, ‘Well, let me talk to my wife,’ so I called her up and said, ‘How would you like to go off for two years and help to set up a Microsoft lab in Cambridge?’ She said: ‘No! I hate Boston. It’s too cold!’ I said, ‘No, love, the other one,’ and she said, ‘Oh! OK!’
“That’s how I ended up coming to Microsoft.”
Thacker spent his two-year England assignment recruiting, defining the lab’s research agenda, handling publicity, and establishing operational principles. Once Microsoft Research Cambridge was on a sound footing, the Thackers began planning a return to the States.
Karen Thacker was wary of Redmond, concerned with the region’s potential to afflict seasonal affective disorder, so her husband called Rashid and said: “Well, Rick, I’m not moving to Redmond. Am I still working for you?” Assured that remained the case, Thacker returned to the Bay Area and got involved in a nascent electronic-book effort. He’d worked on such projects in the past and convinced the Microsoft team he could build a reader.
Thacker took charge of a group of Peninsula-based contractors, provided the initial design for the Tablet PC, and worked with Flextronics to get it manufactured. While the Tablet PC has never achieved wild popularity, it has gained steadily in market share ever since its debut in 2001.
“Some people at that time really loved Tablets, and some still do,” Thacker observed. “FedEx delivery people use them. We’re beginning to take over medicine, because doctors want to write; they don’t want to type. And university professors love Tablets—for one reason: When they’re giving a class, they can write on their slides as they’re teaching.
“I’m proud of the Tablet PC,” he told Kossow in 2007, “because I think, eventually, it will take over, and all PCs will be Tablet-like.”
Thacker is in his 13th year at Microsoft, but don’t expect his 13-year itch to strike again. He remains engaged and active in his work, leading a computer-architecture group in Silicon Valley and working with academia to use field-programmable gate arrays to enable multicore-computing experimentation. He’s also quite intrigued by a project called Barrelfish, a partnership between Microsoft Research Cambridge and ETH Zurich that will enable research on operating-system principles specifically for multicore systems.
Such efforts have Thacker reconsidering the way computing is done.
“We have made an awful lot of progress over the last 50 years,” he said, “but it’s the 21st century, and the technologies we have at our disposal are very different from the kinds of technologies we had when those original decisions were made. It seems prudent for us to go back and look at some of those things in light of what we have in the 21st century. You might wind up doing computing in a very different way.”
For Thacker, it’s been quite an adventure, from dreams of particle accelerators to becoming a recipient of computing’s greatest honor. Along the way, he’s enjoyed the camaraderie and counsel of friends such as Lampson, Taylor, Alan Kay, Roy Levin, and Mike Schroeder. But Thacker’s focus remains steady: making computing even better:
“My wife says: ‘The computers that I like are the invisible computers. I know my car has 200 computers in it. I don’t see them. I don’t have to interact with them. The way I see you interacting with computers is dumb.’ And she’s right! We have a tremendous number of things to do.”
What about the things Thacker has accomplished thus far? How does he measure his impact?
“Personal computing,” he concluded, “basically grew out of what happened at Xerox PARC—and I had a big part in that. That’s probably enough.”