When John Savage came to Brown a half-century ago, computing was thought of as a service, not a discipline, and the idea of a Department of Computer Science at Brown was highly controversial. Few could have predicted the infrastructure necessary today, or the challenges we face in the areas of cryptography, cybereconomics, big data, and many others.
One of the founders of Brown CS, John's interests have changed as our field has changed, leading him to contribute to such areas as computational complexity, scientific computation, computational nanotechnology, and cybersecurity policy and technology. In place of our annual reunion, we hope you’re planning to join us in Room 130 of 85 Waterman Street from 2-6 PM on Friday, May 26, for a celebration of how John Savage and Brown CS evolved together, including research talks as well as personal reminiscences. A reception will follow at 115 Waterman Street on the third floor of the CIT.
In preparation for all the fun, let’s take a moment to go back in time and take a look at 50 years of co-evolution.
“I’ve had an interesting career,” says John Savage. He’s sitting in his office on the fifth floor of the CIT, looking out onto rooftops in January sunlight. “I don’t enjoy talking about myself, but I like discussing my work. I hope it provides motivation for others to reach out, to experiment.”
The impetus of that career, which has ranged as widely as any in the field, began in Lynn, Massachusetts, where John was the oldest of six children, growing up in a French-Canadian neighborhood and attending schools where some of the instruction was in French. (He still speaks the language.) “Somehow,” he says mysteriously, he acquired a BB gun in his preteen years. “But I was told it was dangerous and urged to get rid of it.”
John set aside his hopes of becoming a sharpshooter, abandoning the gun in favor of a crystal radio. “I was absolutely fascinated by it,” he says. In his teens, studying at the now-defunct Saint Jean Baptiste High School, a French-Canadian parochial school, he became a ham radio operator, upgrading to a 400 MHz receiver/transmitter and working on his Morse code skills until he could transmit 17 words per minute. Like many other enthusiasts, he enjoyed communicating with peers around the world. “But unlike them,” John says, “I was more interested in the technology.”
Inspired to pursue electrical engineering in college, John knew that he needed physics, so he took a night class at the public high school. He soon discovered that the physics teacher, an out-of-his-element basketball coach, was confused by course material about levers that he was supposed to be explaining to students. He sent out his college applications, and the acceptances started coming in: Tufts, then Northeastern. But at MIT, his first choice, John was waitlisted. This meant a visit to the Dean of Admissions, the intimidatingly-named B. Alden Thresher. “He sat at this massive desk in the middle of a big room,” John remembers. “His secretary pulled my folder, and then he looked at my letters of recommendation. ‘Nothing special!’ he said.”
Antone (“Tony”) Medeiros, Professor Emeritus of Medicine at Brown, met John in 1976, and they biked around the East Side every morning together for years. He adds an interesting detail to the story: “When the dean told John that ham radio enthusiasts didn’t often do well at MIT, John had an answer, which was that he was less interested in radio techniques and more interested in the ideas behind the technology. That says something.”
Whatever the reason, Thresher relented. It was an early success with something that John was to do again and again in his future career, making a case for something against difficult odds and delivering a convincing argument. MIT was a financial challenge as much as an intellectual one: John commuted from home, paying tuition first with a $400 scholarship from his church and earnings from a paper route, then working in the refectory, then with a $100 loan from his father, a self-made man whose parents had died young. Eventually, a Massachusetts Bay Scholarship eliminated the need for John to work, which was fortunate. “Going in,” he says, “I’d only had a half-course of trigonometry, and some of my classmates had taken calculus. At the end of the first semester, my grades were exactly the mean for my class!”
As an undergraduate John joined the Cooperative Program at MIT and spent semesters with 25,000 engineers at Bell Labs, working at locations in New Jersey and Massachusetts. Before each assignment he requested that he be allowed to work with antennas, a subject that fascinated him. He was taken to Holmdel and told that they needed to calibrate a massive antenna by pointing it at radio stars, a job known as “bore sighting”. Deciding that this was not the kind of antenna work he envisioned doing, he took an assignment in signal processing. However, the people who did work with the Holmdel Horn Antenna noticed an inexplicable hissing sound. When it was investigated it provided evidence for the Big Bang and led to the Nobel Prize in Physics for Arno Penzias and Robert Woodrow Wilson for the discovery of cosmic microwave background radiation. John had narrowly missed a big opportunity.
John earned his ScB from MIT in 1961, then stayed on to earn his ScM a year later and a PhD in 1965. A return to Bell Labs followed, where one of the projects he worked on was a digital data scrambler that altered scrambled a bitstream so that it would be unintelligible without the appropriate descrambler. He also met his wife, the descendant of multiple Brown alums, in Cambridge (John had borrowed a friend’s ID to obtain access to a Harvard event). Her father played a key role in John’s choice of both career and home.
“I thought I’d get an industry job, but my father-in-law recommended Brown. Providence was very different then: there wasn’t any Interstate 95, and we wondered where we’d live. I thought we’d be here about three years,” John laughs.
A Combined Approach
John’s industry experience proved to be extremely useful for future academic work. His PhD thesis had been on sequential decoding, and at Bell Labs, he noticed the huge disparity between the size of encoders and decoders. It was the beginning of a long-term interest in how the idea of complexity was relevant to computing in general, which John pursued across multiple research projects, including a 1972 paper (“Computational Work and Time on Finite Machines”) for the Journal of the Association for Computing Machinery.
Barrett Hazeltine, Professor of Engineering Emeritus and Adjunct Professor of Engineering, has known John from his earliest days at Brown. “It’s hard to explain this now that everyone understands the value of computer science,” he notes, “but we deal with gigabits now, and John was a real pioneer of complexity and minimization in the days when you had to build every gate by hand. There really wasn’t anyone else doing CS with his combined approach of engineering, math, and theory.”
In the years to follow, Savage’s interests grew to encompass applied theory of computation, which includes space-time tradeoffs in serial computation and area-time tradeoffs in the very large scale integrated (VLSI) model of computation, as well as silicon compilers (John notes that colleague Steve Reiss wrote code for one of the earliest of these) and the parallel algorithms used with them.
“Computational complexity is still interesting to me,” John says. “One of my early computer science results was reminiscent of the Heisenberg Uncertainty Principle. I demonstrated that the product of storage space and computation time on a random access machine could not both be made arbitrarily small simultaneously. The circuit complexity of the problem being computed puts a lower limit on the product.”
A New Framework
John’s imagined three years at Brown became six, and when his first sabbatical arrived, he brought his wife and two small children to the Netherlands. While there, an important long-distance conversation with colleagues Andy van Dam and Peter Wegner was starting. A movement had slowly begun: universities nationwide were establishing Departments of Computer Science, primarily for graduate studies.
At the time, some faculty members of Brown’s Divisions of Applied Math and Engineering considered themselves computer scientists and were teaching what were effectively CS courses. Capitalizing on those early efforts, Savage, van Dam, and Wegner wanted to create a new framework for the study of CS at Brown. An initial attempt to obtain permission for a Department of Computer Science failed, but with the aid of Maurice Glicksman (first Dean of the Graduate School and later Provost), the three created a cross-Division Program in Computer Science in 1975.
“After we proved through being a Program that we were a survivable entity,” Andy remembers, “we asked again about becoming a Department, and that's when the antibodies came flooding out. There was fear and even anger that we were proposing to remove valued parts of both Applied Math and Engineering, and Maurice had to make assurances that new positions would be made available to the Divisions. Throughout the whole effort, John and I had complementary views that CS was worthy of becoming a unit of some kind, and we each represented part of the entire discipline. He and I worked together to advance the ball, so I know how much effort he put into it. He was definitely skilled at building the argumentation and delivering it effectively.”
Four years after the Program was established, John assembled a committee that included such luminaries as Dick Karp (University of California, Berkeley), Peter Elias (MIT), Juris Hartmanis (Cornell), and Alan Newell (Carnegie Mellon), and the trio’s years of effort finally paid off with the creation of the Department of Computer Science in 1979. “John saw beyond that era’s conceptions of what computer science was,” says Tony Medeiros. “And then he ran with it!” But even the inaugural symposium, which John organized, wasn’t without its challenges: a student, perhaps overawed by the honor of carrying slides for the world-renowned Donald Knuth, who gave the keynote address, promptly spilled the entire carousel on the floor.
“Our department has worked together so well for decades,” John says, “and to some degree, I take it as a measure of success of our launch. We established practices that are still followed today, and we balanced theory and practice at a time when our competitors were only doing one or the other. One of the first things that our new faculty members say is that our faculty gets along. By no means is that true everywhere!”
In The Chair
The 1980s were a time of rapid growth for Brown CS. Early in the decade, grants from the National Science Foundation and Exxon funded the first electronic workstation classroom and the first large group of networked Apollo workstations anywhere. John followed Andy as Department Chair, and by his second year in the role, new workstations were needed. John led the effort of creating a request for proposal to all vendors, which were eventually narrowed down to Sun, DEC, and NeXT. The goal was what was called a “3M” machine: one million bytes of memory, cycles per second, and pixels on a screen. A highly contentious “bake-off” (Steve Jobs of NeXT was in his argumentative prime) followed, with Tom Doeppner making the case for Sun, Steve Reiss for DEC, and Andy van Dam for NeXT. Sun was the winner, resulting in a multi-decade relationship.
“There was always a lot of back and forth when I was Chair,” John remembers. “Between students and faculty, among faculty members, with the administration. I started weekly faculty meetings over lunch to help us solve problems collaboratively, and they continue today.”
Andy notes, “As Chair, John maintained his strong interest in growing the department. He stayed the course, which for that particular time period was much more challenging than it sounds, given that we were still a very small department trying to compete for faculty, students, and research grants with much better-known and much larger CS departments. Absolutely, he was a great Chair, always active, always helping to build in a variety of ways.”
Another of John’s major contributions was the Industry Partners Program (IPP), which he founded in 1989 in conjunction with Brown’s Development Office and Roy Bonner of IBM. The program creates closer connections between Brown CS and industry, and member companies are encouraged to recruit students, participate in the selection of topics for IPP symposia, and advise on the employment and research needs of corporations. Income from the program has proved crucial to the department’s growth, supporting everything from equipment to faculty searches to distinguished lectures.
The most visible reminder of John’s days as Department Chair is the Thomas J. Watson, Senior Center for Information Technology (the CIT). Andy explains that the building was finished on John’s watch, and he oversaw its execution: “You could count on John. If he said he’d do something, he got it done, on time, and with high quality.” Professor Stan Zdonik adds that John’s work as Department Chair was one of his biggest contributions to Brown CS. “He took the job extremely seriously and did great things,” says Stan. “He was part of a small group of people who deserve huge credit for basically getting the CIT built. That took an incredible amount of work with the university, getting people to agree.”
An Awful Lot To Learn
As the 1980s turned into the 1990s, John’s research interests turned toward scientific computing, working with a student named Jose Castellanos who later joined the IBM Blue Gene team. In addition to their work on implementing a finite element method in distributed computing and mesh adaptation to 2-space and 3-space tetrahedral to simulate the Navier-Stokes equations used to model fluid moving around obstacles, again with an eye toward distributed processing.
It was more evidence for the polymathy that’s been a hallmark of his career. “When I met John in 1976,” says Tony Medeiros, “personal computing was in its early days, and I was a bit of a techie, which meant stepping outside my field, but I was nothing compared to John, who was full of wonderful ideas in my area of interest, microbial antibiotic resistance, and everything else. You couldn’t bring up a topic that he wasn’t intrigued about.”
“Ideas excite me,” John says when asked to explain his wide-ranging interests, telling the story of how he took charge of his college education after being poorly prepared in high school, and was motivated to delve into computational nanotechnology by reading through a rising stack of nanotechnology papers on his desk, plunging in even when he didn’t fully understand them. “I can’t really help being curious about things, and I have other interests that I haven’t fully explored, like cryptography. There’s still an awful lot to learn.”
John’s work with nanotechnology began in 2001, with a one-year starter grant. That led to a large four-year NSF grant that he shared with Charles Lieber of Harvard and André DeHon of Caltech. He says, “When I discovered nanotechnology, I thought it would have the impact that the VLSI revolution did in the 1970s. It hasn’t turned out that way yet, but it might eventually.” Meanwhile, he was learning by attacking the problem in several different directions: giving lots of talks and developing an introductory course on nanotechnology that included quantum computing, nanowire-based technology, and synthetic computing. His research continued until 2011, and he earned his last patent in the area in 2016. Most recently, John’s interest in applied theory of computation has led to publications on input/output complexity for multi-core chips with shared memory.
According to Barrett Hazeltine, the pleasure that John gets out of new and varied research finds synergy with his interpersonal skill: “When you talk to John, you see that he’s enjoying what he’s doing, and he expects people to do the same. ‘Wouldn’t it be fun if’ is how he starts a conversation. Maybe because of his many interests, he gets people to approach their research in another way, and he’s gracious, he makes you think you’re the most important person in the room. He’s a very strong colleague who’s had the courage to change the university for the better in many ways.”
Service Near And Far
Much of this change has occurred through John’s decades of service to Brown, often during times of protest (in response to the effect that the proposed 1975-1976 budget would have on minorities, students occupied University Hall) and strained relations between the faculty and administration. Savage has been Chair, Vice Chair, and Past Chair of the Faculty, Chair of the Task Force on Faculty Governance, Chair of the Nominations Committee, Chair of the Search Committee for Vice President for Public Affairs and University Relations, President of the Faculty Club Board of Managers, and a chair or member of numerous other committees. Among many other achievements, he formed the Academic Priorities Committee, made significant changes to policies concerning budgets and tenure, and created a massive Handbook for the Task Force on Faculty Governance. In recognition of all these efforts, John received the President’s Award for Excellence in Faculty Governance in 2009.
“A lot of people see service as a chore,” Barrett adds, “but John is both a leader and a doer. He answered the question of how faculty can usefully contribute to how Brown runs. He’s one of the people responsible for all the benefit we get from being in a community where the best ideas can come forth and be made known.”
And although it might not be service in the traditional sense, Tony Medeiros points out the Savage family’s incredible allegiance to Brown: “I don’t know if anyone else has said it, so I want to mention that John may have set some kind of a record: all four of his children went to Brown and married Brown alums. He’s a great father and family man, a wonderful friend.”
John’s service has also taken him far beyond the Van Wickle Gates. His contributions to professional societies have been many, including being chosen as a member of the NSF Review Panel on Emerging Technologies and the Program Committee for the IEEE/ACM International Symposium On Nanoscale Architectures. But it’s John’s work in cybersecurity and Internet governance that’s had a truly international scale: he’s served as a Jefferson Science Fellow for the US Department of State, a panelist for the Global Futures Forum in Singapore, an honored guest of the government of Vietnam, and a member of the Scientific and Technical Intelligence Committee.
Judith Strotz, Director of the Office of Cyber Affairs at the US Department of State, worked extensively with John when he was asked to help policymakers in her newly-formed Office better understand science and follow rapidly-emerging trends. “To me,” she says, “the main thing is that he trained us all. It doesn’t sound momentous, but it really is. It’s really hard for non-scientists to address all these issues, and we made huge strides thanks to his knowledge, persistence, and patience. He’s a true expert and a great colleague who helped us sort through problems and find a way forward. I miss working with him.”
John is also a Professorial Fellow for the EastWest Institute, a non-profit organization with the goal of reducing conflict worldwide. Bruce McConnell is the organization’s Global Vice-President, and one of Savage’s frequent collaborators, including a recent paper on Internet governance. He says, “John’s influence and impact are immeasurable…He’s hardworking, and when he says he’ll do something, he gets it done. But then he goes beyond that by bringing his enthusiasm, curiosity, and creativity. He always opens up options.”
“It’s been great fun,” John says of his time in Washington, remembering days when he had the maximum level of security clearance, unable to bring any technology into his office and required to store his top-secret disk in a safe every night. Always the educator, he makes an interesting point about autodidacticism: “90 to 95 percent of the information used in intelligence analysis is open source. A deep understanding of a region or country is something that can be taught, and avid readers who devote themselves to an area can teach themselves.”
When asked about John’s contributions, Judith Strotz begins with his instructional skill, saying, “John is really humble. He taught our team of policy makers in simple terms at first, but without ever being condescending. He showed respect and understanding of where everyone was at in order to help them learn. To me, that’s great teaching.”
Over the years, John has taught more than 15 courses at Brown, covering topics from computational complexity to cybersecurity. This spring, he’s teaching CSCI 1800 Cybersecurity and International Relations for the seventh time, bringing the subject material to life with a series of world-class speakers, including an expert in cyber economics and the National Intelligence Officer for Cyber Issues.
He’s also fielded multiple winning teams of students at the CYFY, CyberSeed, and Cyber 9/12 competitions. “Teaching is how I get firsthand experience,” he says, citing a class (CSCI 1951-E Computer Systems Security: Principles and Practice) that functions as an enriched version of a course developed by Professor Roberto Tamassia. “The best way to fill gaps in my knowledge is to study something, then teach it.”
The theme of co-evolution was John’s carefully-considered choice for the upcoming celebration in May. It reflects not just his multifaceted career and the growth of our department but the evolving face of computer science and the continuing challenge of developing the technology needed for a changing world.
“Start by thinking about how difficult it is to build an operating system,” John says. “Very roughly, you’re looking at 80 million lines of code that have to be designed, written, tested, documented. If someone can write 20 lines of code per day, at 250 working days per year, you’re looking at 16,000 person-years of work. It’s an enormous effort.”
At the same time as John and his colleagues were crafting policy and procedure for an evolving Brown CS, they were creating an equivalent framework for computer science itself: “Over the years, we’ve had to build the entire infrastructure to support our own discipline: we had to invent computer graphics so we could have graphical user interfaces, we had to create operating systems that scheduled jobs correctly, we had to define protocols for sending packets so the Internet can work. If this had been commonly understood in the 1970s, we would’ve been given a department the first time we asked!”
And yet there’s no sign of cynicism in the remark. Bruce McConnell sees John’s positivity as an unmistakable part of his secret to success: “He always sees another angle and brings in ideas from other domains. It’s intellect plus optimism. For John, the glass is always more than half full!”
The next set of challenges, John explains, demands rigor. As scientists, responsibility has been placed directly in front of us. “Most people can learn to write software for simple tasks,” he says, “but the difficult problems need scientists. There are new challenges in artificial intelligence, cybersecurity, nanotechnology, data science. There are very specific problems, like the role of social media in promoting fake news or the ethical questions of self-driving cars, which demonstrate that we’re in an era where the technology we invented has unintended consequences. Computer science has become a critically important field. These are technical challenges that demand expertise and will play out on a social, economic, and geopolitical scale. We need to play a central role.”
That role is also an interdisciplinary and intersectional one: “Computer science will become important to more and more people. I’d love to see more women in the field and more historically underrepresented groups: we need different perspectives because of the impact that we’re having on international relations, engineering, and other areas. I want to see people be tolerant and generous with each other -- we’re all human beings. A hundred thousand years of thought have gone into our survival, and we need to cooperate and pass on knowledge. My message to the next generation of students is to stay positive: when you encounter problems, get the best advice, work them out, keep plowing through, and don’t give up.”
John Savage’s track record shows a more than ample willingness to take up that gauntlet with energy and purpose. “He has this quality,” says Andy van Dam, “of positivity and non-superficiality, of digging in, of having a grounded position and willingness to keep pushing at something.”
“Whatever the challenge is,” says Tony Medeiros, “he always has ideas. It’s trivial to say that John is incredibly smart, but his enthusiasm has always taken him everywhere, and it still does.”
For more information, click the link that follows to contact Brown CS Communication Outreach Specialist Jesse C. Polhemus.