Brown CS News

Celebrate With Andy: 50 Years Of CS At Brown


    As the department approaches its silver anniversary, Brown CS is making a comprehensive effort to document our early history online and in print. Our themes are the intellectual daring of our academic home, the making of things never before imagined, and the spirit of community that’s guided the entire effort. Throughout, our storytellers will be the faculty and alumni whose examples go before us. This article is the first in that series.


    Andy would like to mention that in the writing of this piece, he remembered many other stories and many other key contributors that we were unable to include due to space limitations.


    The semicentennial festivities known as Celebrate With Andy, to be held on May 22, mark three golden anniversaries for the Brown CS family: fifty years of the UTA program, undergraduate involvement in research, and Andy van Dam at Brown.


    A thousand words don’t suffice to tell the history of those three “institutions”, but one picture from our archives evokes something of their spirit. Believed to be taken at Commencement 1977, it shows van Dam and five others (Heather Claflin ‘77, Peter Relson ‘77, Douglas Dixon ‘77, the late David Notkin ‘77, and Henri Bulterman ‘71 ‘77 ScM ), some in regalia and some not. Laughter and conversation are passing diagonally across the scene, and just for a moment, neither in the center of the composition nor at the Golden Mean, Andy has turned to the camera very casually: “Look at the great people we’ve got here!”


    It’s time for a toast. Romulus and Remus are fables: Brown CS doesn’t require a founding myth because the true history of these three “institutions” is compelling enough. This article, however, will give only an abbreviated version of that story and tell just a fraction of the anecdotes. (We’d rather you join us in person or via livestream on May 22.) Instead, in preparation for all the fun of Celebrate With Andy, let’s take a few moments now to go back in time to reflect, enjoy, and celebrate.


    A Random Sequence


    The story of CS at Brown begins, humbly enough, in a bathroom. In 1962, Andy’s wife, Debbie, was teaching high school French, and one of her National Education Association magazines had ended up as his bathroom reading material. “I read an article,” he remembers, “about teaching students at Bronx High School of Science to program and thought, ‘That’s ridiculous! I’m in grad school, just learning to do that!’ But it kept bugging me, and I figured that it wasn’t quantum physics, that they could certainly learn logical thinking.”  


    So in 1962 he began a summer program of his own, showing Philadelphia area high school students and their teachers how to program, and even managed to obtain his first NSF grant, to support this novel project. “Running this course is how I fell in love with teaching,” van Dam says, noting that the program continued on after his tenure. Alumni of the pioneering effort include Elliot Perlman, a prominent local ophthalmologist who today counts Andy and Tom Doeppner among his clients; our own Steve Reiss was a later graduate.


    “I was teaching programming,” Andy says, “but in my own first course in grad school, I didn’t actually get to use a computer because Penn’s sole computer was too scarce a resource. We wrote machine code on paper and had our programs hand-corrected, but at least I got to see the mainframe!” He describes as “pseudo-religious” the experience of standing not just in the the machine room of UNIVAC 1, but inside its main memory, a little room consisting of a thousand words of memory implemented with mercury-filled acoustic delay lines.


    “My arrival at Brown is based on a nearly random sequence of happenstance events,” says Andy. “Reading the magazine article was the second one. The first was an unplanned conversation that got me into computer science to begin with. I was a hard-core electronics engineer, with an offer from Bell Labs to go design transistor circuits. But I went to grad school to learn more about the field, and my officemate mentioned that there was a new course on computers: ‘We both have electives, so how about this new Computers and Automata course??’ Up until that point, I’d only worked on analog computers, not digital ones, and I fell in love and switched my field. Over the years, and even now, as I interview new TAs, hundreds of people have shared similar experiences with me: taking a CS class at random, falling in love, and realizing for the first time what they want to do with their professional lives. It’s many people’s story, not just mine.”


    The third random event was the one that brought Andy to Brown. “Reading that magazine,” he says, “led to teaching, and teaching led to a phone call three years later from a graduate of the summer program, James Castellan, who was then a student in Applied Math at Brown.” Castellan called Andy after van Dam had all but accepted his first academic job elsewhere, asking if Andy knew that they were recruiting a CS person in what students called “Apple Math” at the time. “I don’t know anything about Brown or Applied Math,” Andy replied, adding that he’d essentially already made his decision. Jim persisted, saying that Brown was the perfect place because of its emphasis on undergraduate teaching.


    “I spent one day here,” Andy says, “and knew it was the right place because of their early history using computers in Applied Math and Engineering and especially because of the undergraduate emphasis.” The clincher, he adds, was when the Chairman of Applied Mathematics excused himself in the middle of Andy’s candidate interview to go teach a first-year course.


    Thinking Of Themselves As Computer Scientists


    Understanding the genesis of  undergraduate teaching assistants and undergraduate research is impossible without seeing the context of specialization and the department’s formation. Future articles will address this issue in depth, and be narrated by others, but these new roles for undergraduates were were born in what van Dam calls those “workaholic, all-consuming, frantic” early days. “For some time,” Andy says, “there had already been LISP and FORTRAN programming courses at Brown, but I was brought to Brown to create a formal computer science track within the Division of Applied Mathematics. Applied Math didn’t see CS as something self-standing, but by the late ‘60s, after John Savage and Peter Wegner and several others came to Brown, undergraduates began thinking of themselves as computer scientists: your degree would say ‘Applied Mathematics’ on it, but effectively, you were a computer scientist.”


    Andy, John, and Peter initially proposed a Center for Computer Science, but --amazing in retrospect-- the University’s response to the proposal was negative, and so the three colleagues drafted a two-division program uniting CS-oriented faculty and courses from Applied Math and Engineering. In essence, it was a trial department, followed by the official establishment of the Brown CS Department. Andy was the Program Director and reported to both Division Chairs; in the final year of three, John served as Acting Director in his absence until Andy returned from his sabbatical at CERN and the University of Geneva in the summer of 1979 to become Chair of the new Department.


    When the time came to start a computer science degree program at all three levels (undergraduate, Master’s, and PhD), competing universities were anything but amenable to the idea of undergraduate CS education. “Here’s the thing,” van Dam explains. “Even if we didn’t call it that, Brown almost exclusively pioneered the idea of an undergraduate CS program, which our competitors said was premature specialization -- they thought students should wait to specialize in CS at the few grad schools that offered programs at that time.”


    “The genesis of undergraduate participation in teaching and research has to be contextualized by the fact that in 1965, teaching computer science as a degree program, not just a few programming courses, to undergrads at all was novel. Teaching these various topics in computer science that were being taught at the graduate level elsewhere to undergraduates, even beginning undergraduates, worked at Brown because of the high quality of the students who were willing to be part of this ‘total immersion’ style of learning.”


    “Offering teaching and research assistant opportunities to undergrads,” he says, “was even more unusual, indeed was viewed with everything from skepticism to outright hostility. Hardly anyone said, ‘What a fantastic idea!’ Everyone was used to four years of preparation as an undergraduate, then n years of graduate work before you could contribute to a science. But we’re different. CS was and is young, experimental, and open for undergrads to contribute. And undergraduate participation in research in all fields has become commonplace, especially in the last decade.”


    Born Of Necessity: The UTA Program


    In 1965, a single, intense full-year course could cover much of the breadth, if not the depth, of the systems-oriented portion of the discipline, not including theory, AI, numerical analysis, and a few other topics. Andy insisted that students couldn’t learn to be good programmers by solving small “toy” problems; they had to write significantly-sized programs, each taking multiple weeks.


    Not just checking for the right answer but giving useful feedback on structure, style, and efficiency required careful reading and one-on-one help with concepts and debugging. In a class with forty students, it was impossible for one graduate TA and a professor to provide this level of attention, no matter how little sleep they were getting, so van Dam asked for help from students who had taken a prior programming course. In that first cohort, he remembers Bill Adcock; Dan Bergeron, who also subsequently got his PhD with Andy and became Chairman of the CS Department at UNH and went with him and a group of six other of Andy’s students for his first sabbatical in 1971 at the University of Nijmegen in Andy’s country of origin; and Dennis Ruggles, among others.


    “The undergraduate teaching assistants,” Andy explains, “though they were initially called graders, didn’t just grade programs -- they not only provided one-on-one help to students but also became active participants in course design and in subsequent years read research papers and brought new ideas into the curriculum. In fact, they did everything graduate TAs did, becoming producers and not just consumers of education. We kept modifying the course as we went along, but the one constant was the highly-appreciated UTA system.”


    Few people appreciate it more than Ed Lazowska ‘72, who will lead the first (“Stone Age”) panel for Celebrate With Andy. He says, “I’m a faculty member precisely because of the UTA program. I went to grad school because Andy told me to. In some way, everything I do professionally today is due to him.”


    To provide feedback for the course, students wrote detailed, multi-page evaluations, something that was almost unheard of in 1965. As Bob Munck recalls, “Also after every class, the graders would sit around on the floor of Andy's office (later my office) and critique the lecture and him. I'd never seen anything like it.”


    On his commute home from work, Andy would listen to tape recordings of his lectures, filling the empty minutes with self-critiques: “Boy, was that a clumsy explanation! Get rid of the ‘um’s and the ‘you know’s.” Presentation skills are still something that van Dam is keenly interested in. “Today’s equivalent of ‘you know’ is ‘like’, which I try to stamp out in all students who work with me. I’ve given up on ‘awesome’.”    


    An interesting aspect of the UTA program is that the system has essentially never been challenged by students due to the built-in checks and balances. “By having rotating TAs and detailed rubrics,” Andy says, “you create fairness. It’s a system that’s at least as fair as having a single faculty member grading. Besides, a single faculty member, even assisted by a few graduate TAs, can’t begin to read that many programs at the required level of detail, and students recognize that. Part of the checks and balances is that faculty members are responsible for assigning the final grades, and I personally review all borderline grades, hoping to find evidence for promotion to the next grade bin.”


    Originally something made up as they went along, the UTA program matured over a period of decades. Iteration and gradual regularization brought cross-course norms and standards that are used today by almost all Brown CS courses. “In my opinion,” says Andy, “We have the most systemic TA program, and there’s a well-defined appeal system in place to address any grading errors.”


    In some classes, van Dam explains, even PhD students are in a course with undergraduate TAs, but it works: for a particular topic, in a particular course, the younger student knows more. He drops his voice an octave to imitate a disbelieving critic. “Undergrads grading grad students? How can that work?”


    His own warm bark of a laugh is flung out with the response: “Just fine -- in fifty years, I’ve never had a complaint from a grad student! They respect competence as much as I do.”


    Fearless, Ambitious: Undergraduate Research


    After undergraduates had successfully assisted with teaching, the logical next step was assisting with research; the logical choice of accommodation for a research team was a shared room. If these conclusions seem obvious, listen to Andy’s description of what happened: “Applied Math had never seen anything like it! The biggest room in the building as this nerve center, six hundred square feet for me and an admin and four or five student researchers, going in and out at all hours of the night. They saw us as these unwashed hippie kids, loud adolescents -- how could they possibly do research?”


    “We occupied the entire basement, too,” says Ed, “dozens of desks, cheek-to-jowl. What really drove the Applied Math faculty crazy is that we were constantly running from the basement to the third floor. We had weekly project meetings in Andy's office, lined with bookshelves and filing cabinets surrounding the desks, where several dozen students would cram in, with no room to breathe.”


    “The schleps, as we called them,” Andy continues, “were a group of more than a dozen undergraduates apprenticing in the group who contributed in every capacity, from getting lunch to reading research papers and explaining them to me to see if they were usable in class. They worked with a few Master’s and PhD students, including the late Charles Strauss, Dan Bergeron, and Jim Michener, among others. The bullpen was noisily chaotic, but we were young, and kids have powers of concentration that adults don’t. The best part was the selective eavesdropping and peripheral conversations, learning by accident from the people around you. My graphics group still works that way: we strongly encourage everyone to be in the graphics lab together multiple nights per week.”


    “Resulting in at least five geek-geek marriages,” notes Bob Munck, crediting the fact that women were fully equal contributors in Andy’s group from the beginning.


    These were the days in which the IBM /360 Model 50, which started with 256KB of memory and no disks and was upgraded to 512KB and a disk array, served the entire university. Normally, users keypunched their programs on decks of “IBM cards” and submitted jobs that were processed in batches, many hours later. Andy’s group, doing graphics research on their IBM 2250 display (courtesy of an IBM research grant), were allowed small chunks of time during third shift to debug their programs, where an occasional crash that brought down the mainframe was reluctantly tolerated by the operators.


    Stories of how Andy and company were at times dilatory with food and candy wrapper disposal, how they allowed dogs (and therefore, without putting too fine a point on it, the things dogs do) into the machine room, and how they bought far more Girl Scout cookies than were strictly necessary from “Big Grace”, the head operator, are better left for Celebrate With Andy.  


    At the time, these quarter-of-a-million-dollar displays (in 1967 dollars: close to two million in today’s money) were rare indeed, and letting undergraduates have access to them was even rarer. With Brown’s acquisition of the 360/67, Andy’s group became one of the earliest users of virtual memory and virtual machine-based time-sharing. “Among the many firsts,” Andy explains, “Bob Munck and other students built time-sharing systems to run in a partition of the OS on the 360/50. Even before his graduation, Bob took the highly unusual step of teaching portions of my courses, on assembly language and other systems topics.”


    “Steve Carmody was another student in my first course in 1965, and is still associated with CIS at Brown. He was a leader in the group project to design and implement the first hypertext system on commercial equipment in 1967, the Hypertext Editing System (HES). HES was co-designed with Ted Nelson, coiner of the term ‘hypertext’. HES was followed by FRESS (File Retrieval and Editing System), which was an active project for more than a decade, starting in 1968. Many undergraduates contributed to its design and implementation, including the late Bob Wallace, who was one of the seven original founders of Microsoft and the inventor of shareware with his utility, PC-Write.” 


    As part of the LSD (Language for Systems Development) project to define a systems programming language and create an optimizing compiler for it, led by Dan Bergeron as a PhD student, Andy recalls the thrill of having exclusive access to some highly sought-after IBM software. “We were using a proprietary systems dialect of the standard language, PL/I, called PL/S,” he says. “It was never used by anyone else outside the company, and even years later, I’d get calls from people within IBM, wondering if I could tell them about this mysterious language that they’d heard about but weren’t able to get a hold of.” 


    “We were also among the very first to do simple distributed computing by attaching graphics mini-computers to a mainframe. The late George Stabler and Rick Harrington,” Andy adds, “a PhD student and undergraduate, respectively, designed, implemented, and published the first remote procedure call (RPC) protocol to allow code modules on the graphics satellite and code modules on the mainframe to call each other, and even to let code migrate to do load balancing from one to the other, at least a decade before other organizations reinvented the idea of RPC. To make our microprogrammed multiprocessor graphics satellite even more real-time, undergraduate Hal Webber designed and built the first high-performance, microprogrammable 3D and 4D homogeneous coordinate transformation engine, SIMALE, now part of our Computer Museum. PhD student Jack Stankovic, who became the Department Chair at UVA, and I ran the first workshops on distributed computing in the 1970’s.”  


    “The whole idea about being a research assistant,” says Ed Lazowska, “was that Andy asked us to figure out how to do things that hadn’t been done before. It was the first time that someone had treated me as an intellectual peer and showed confidence that I could do the tasks that adults could do. The whole group was remarkable, and Andy and Charles had an extraordinary impact on me. They totally captured my imagination.”


    Even the briefest look at Strauss’s research gives a powerful sense of the time. For the first time, his work with a light pen and specially-designed stereoscope that fused left and right images on a split screen allowed the user to navigate a live, 3D stereo representation of the layout of pipes in an oil refinery, helping identify potential interference between pipes. For the computer user of today, the world in which mainframes rented for tens of thousands of dollars per month is scarcely thinkable. Looking that many decades into the past, we might be impressed to find graphics of any kind, even 2D. Yet in this case, 2D wasn’t sufficient, and neither was static 3D: Brown had to pioneer not just 3D graphics but interactive 3D stereo motion graphics to provide the functionality required. Working with Professor Tom Banchoff of the Math Department, Charles was the first to provide real-time manipulation and visualization of Möbius strips and 4D geometry: hypercubes and hyper tori. “Banchoff-Straus Productions” continued for decades and produced impactful movies of manipulations of 4D geometry, which were greatly aided by Hal Webber’s SIMALE.


    “Up until this point, computers were used almost entirely for crunching numbers, and computers with graphics were for oil companies and car and airplane manufacturers,” says Ed, whose Brown independent concentration was titled “Non-Numerical Computer Science”. “With HES and FRESS, we were working with text! Not just text, but WYSIWYG hypertext. It wasn’t until that point,” says Ed, “that you could actually put the word ‘personal’ in front of the word ‘computer,’ although our PC was a multi-million dollar mainframe. That’s all Andy.”


    Andy shrugs. “We were just fearless, we had ambitious ideas, didn’t really know what was possible and what was not. I had all those smart and highly motivated kids available, so we took wild, crazy ideas and ran with them.”


    Creative Expression


    What’s the common link between undergraduate teaching assistants and undergraduate researchers? “Creating knowledge,” says Ed, “not absorbing knowledge. Creativity. Teaching and research both need this in spades.”


    Janete Perez ‘06, who will lead the third (“Machine Age”) panel at Celebrate With Andy, says, “I wanted to be a UTA like high schoolers want to be on the varsity football team...To add to a class, make it more fun, be part of it all...Andy finds the kids that are really excited, not just the straight-A students. He teaches you to work hard and be disciplined, but really to be creative.”


    Interestingly, van Dam’s thoughts move in a similar line when asked about the theme of his life’s work: “From the time I saw Sutherland’s mind-blowing Sketchpad movie in ‘64, I’ve loved human-computer interaction and the visual; I really value creative expression in various media.”


    “Watching the Sketchpad movie,” he says, “was another random event that changed my life forever. At the time, computing was done by entering programs and data via punch cards or teletype tape. Programs were run in batches and dealt with numeric data.  Sutherland showed the world interactivity, humans working with computers in real-time, and he showed us communicating through drawing and manipulating images directly. It was beyond revolutionary, and like the other random events, it ended up completely determining my career. When advisees come to me obsessed with making a commitment to one research area, I tell them to relax: ‘You’re going to experience ideas that’ll change your point of view. Be open, experiment, try different things. You’ll change your mind a half-dozen times.’ It was true then and it’s a hundred times more true now.”


    But back then or now, what could be a taller order than trying to foster creativity? After hundreds of thousands of years of human history, our understanding of our creative powers is still incomplete. But let’s try. Ed shares a visual from the height of the late 60’s hippie era that we’ll return to later: driving in a Volkswagen bus in mid-winter to northern Virginia, where the group programmed a special version of FRESS for a 3-lettered federal agency while locked in a Faraday cage and monitored by armed guards when they went to the bathroom. “There was no working heater in the bus (of course),” adds Steve Carmody, “so for the entire trip people took turns sitting atop a multi-platter disk, trying to keep it warmer than the frigid air inside the bus.”


    Now let’s try to press Andy again for the secret of bringing out people’s creativity. Asked for heroes of the recent past, he divides them into those outside his field (Einstein, Feynman, Gandhi, King, Mandela) and those inside: Engelbart, Turing, von Neumann, Maurice Wilkes. “Douglas Engelbart was just amazing, but he unfortunately never had the direct impact he should have had...he had trouble explaining his vision, and sadly is only remembered for inventing the mouse, probably the least important of his huge number of inventions in the oNLine System, NLS, from which so many of our modern ideas of word- and idea-processors derive.”


    Hearing that, anyone who has known van Dam for five minutes is going to think the same thing: Andy doesn’t have a problem explaining or sharing his vision. Whether it was inspiring Janete and her peer UTAs to create elaborate skits for CS15 class with classically-trained musicians playing instruments, or getting a bunch of tie-dyed rebels to bring peace and love into the fortress of unsmiling agents, Andy van Dam fosters creativity through personal connection. “Students relate to him,” Janete says. “More than fifty years younger, they relate to him because he relates to who they are.”


    “Andy believes in the power of young students,” adds Ed. “He taught me that impact and excellence are a multidimensional quantity...we can’t hire or admit people or motivate them to do great work through just one lens.”  


    When we ask Andy about a common theme for both undergraduate teaching assistants and researchers, there’s a long pause. “The idea that you can do serious work before the PhD is almost unique to CS,” he says. “We did it fifty years ago, and it’s true to a fare-thee-well’s about skilled and creative analysis and synthesis; I try to create a productive, challenging but nurturing environment for creatives to make things that are, to use the Jobs phrase, insanely great.”


    Responsibility, Then And Now


    Norm Meyrowitz ‘81, who will lead the second (“Iron Age”) panel at Celebrate With Andy, also believes that giving responsibility to young students was (and is) key. “We were seventeen, eighteen, nineteen, creating all the assignments for an entire class. I was a junior, writing a windows manager program, a twenty-year-old with other twenty-year-olds presenting our research to the NSF.”


    Did this seem strange at the time? “It just felt like something we were supposed to do!” Norm replies. “Remember, the only PC then was the Apple II: there were no mice, no graphics, no hypertext in general use. There were thousands of areas that nobody had started researching yet, so it was exciting. So Andy, Bill Shipp [then Vice-Provost of Computing and a professor of biology], and I --at the ripe old age of 23!-- raised 17 million dollars to put hundreds of graphical workstations on campus. The workstation effort led to the development of the Intermedia hypermedia system --which looked like the Web and worked on a LAN in 1985-- by me along with many of Andy's former students. Andy gave us incredible responsibility as undergrads, and that bred confidence in future endeavors.”


    Their confidence was clearly justified: the third extant HTML message ever created, by Tim Berners-Lee for his own research team, was about Intermedia.


    Responsibility and confidence are both in full view at a January, 2015 meeting of Andy’s group in the CIT Library. Students and researchers sit at a long table with their laptops, tablets, and smartphones; their mentor is leaning back in a chair, dressed in familiar sweater-over-the-shoulders style with arms crossed. His laptop (a ThinkPad touch tablet) is closed, and his smartphone only emerges for a momentary calendar check.


    “The batch inputter is going pretty well,” says a student.


    Andy’s eyes narrow a little as he grins. “Did I hear ‘pretty well’ or ‘very well’ in there?”


    “We’re working on the user interface, but we haven’t caught the edge cases yet. We’ll be done by end of day, then test.”


    “No more guesswork,” says Andy on another feature. “We may have to take over these transactions ourselves.”


    The atmosphere in the room is immediately recognizable to anyone who has worked in a production environment, in the technology sector, or in any setting where efficiency matters. The students are obstacle-oriented, focused, going back and forth seamlessly. Andy lets them converse, answering each other’s questions and asking for clarification when necessary.


    The sophistication of the students could be envied by professionals twice their age, and it’s mirrored in the vocabulary that van Dam uses in his comments: bona fides, rubric, interregnum, kibbitz. Later, he’ll credit this to “growing up surrounded by people who use language well” and not being a native speaker of English, but the fact remains: this is the way that someone addresses responsible peers, not children. For the hearer, confidence follows naturally.


    After an impromptu test of a new feature on an Android tablet works better than expected, the table explodes in cheers: “Yay!...Wow!...Show that again!”


    “See you next week,” Andy says at the end. “Same bat time, same bat station.”


    Getting It Done


    “Fifty years later,” Ed says, “every generation tells the same stories.”


    “Andy always said that we’re here to make the future happen,” Norm adds. “Research is a byproduct of having a vision of the future and sharing it through teaching, instilling it in generation after generation. Research is just those people making the vision happen, getting it done...As teaching assistants or researchers, everyone from first-years to PhD students, Andy had us think of ourselves as a collaborative troupe spanning the generations. Every day, we put on a show and did our best. After you leave Brown, the troupe still exists for you in the community, the camaraderie.”


    Celebrate With Andy is only one night, but it carries a thank-you that doesn’t end to Andy and everybody else for our past half-century. The troupe goes on, making the future happen.