Oral-History:Jonathan Grudin

From ETHW

About Jonathan Grudin[edit | edit source]

Jonathan Grudin was trained in mathematics, physics, and cognitive psychology. For many years, he has participated actively in the human-computer interaction and computer supported cooperative work communities. He has worked as a software engineer at Wang Laboratories, a team leader at MCC, a professor of information and computer science at University of California Irvine, and as a researcher at Microsoft. As he states on his personal website: “For thirty years I have addressed methods and processes to develop useful technologies, and the design, adoption, and especially use of communication and collaboration tools. Over the past fifteen years I have also focused on understanding the publication cultures and broad histories of the fields with which I’ve engaged.”

About the Interview[edit | edit source]

This interview conducted in several sessions is part of an oral history project on the early history of the computing. It focuses on the discipline of human-computer interaction from the 1980s, as it developed into a core element in computer science curricula. Grudin tells about his family, early upbringing, formal education at Reed College and UC San Diego, and a postdoc in Cambridge. He discusses his early positions in technology companies Wang Laboratories and MCC, subsequent academic positions, and his appointment at Microsoft Research.

Copyright Statement[edit | edit source]

This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE Computer Society. No part of the manuscript may be quoted for publication without the written permission of the IEEE Computer Society.

It is recommended that this oral history be cited as follows:

Jonathan Grudin, an oral history conducted in 2021 by Clayton Lewis, IEEE Computer Society

Interview[edit | edit source]

Interviewee: Jonathan Grudin

Interviewer: Clayton Lewis

Date: June-December 2021

Location: Virtual

Growing up, first exposure to computers, early education and jobs[edit | edit source]

Lewis:

This is Jonathan Grudin, and it's June 30th 2021.

Grudin:

Okay. I was attracted to computers in my teens, and really attracted to humans in my 20s. I didn't get to human-computer interaction until my 30s. This is a winding road.

I'll start with my family. My parents were the first in their families to attend college. Both earned advanced degrees. My father became a mathematician, my mother a teacher who studied psychology. My path to human-computer interaction involved two degrees in mathematics and two in psychology, so my parents are relevant.

Grudin:

My mother was born Elizabeth or Betty Lohnes in Weymouth, Massachusetts. Her father, Arthur Lohnes, thought he was descended from a Hessian mercenary. Later, we found that his ancestor reached America before the revolution and left to sit out the war in Canada. I'm descended from a draft dodger. My maternal grandmother was Swedish, though we didn't know that. She and her sister were orphaned in western Massachusetts. They moved to Weymouth, changed their last name from Peterson to Pedersen, and claimed to be Norwegian. There was widespread prejudice against Swedes. Norwegians were exotic. When she died, my grandfather learned about the deception and was outraged. My mother noted that he wouldn't have married her if he knew she was Swedish.

Lewis:

Oh my goodness. (laughs)

Grudin:

My grandfather came from a line of blacksmiths in Massachusetts. He built houses, he raised and raced trotters, and he owned a feed and grain store that got them through the Great Depression. Hobos came through town, my mother saw her mother quietly give them sandwiches. After her younger brother was born, my grandfather had zero interest in his daughter. He made fun of her avid reading. She left for Colby College. Her subsequent loss of her extended family always haunted her. But she and her brother and father wrote frequently. My father grew up in—

Lewis:

If I could take a minute there. The loss of her extended family?

Grudin:

A lot of the family lived in Massachusetts, around Weymouth, and still do. She described going around on Sundays to visit relatives. Often they'd just sit in the parlor and not say a whole lot. But there was that sense of the extended family. As happens with the first kids who go off to college, in some way they leave the family, the family no longer sees them as part.

Lewis:

All right, that's... Yeah, thanks.

Grudin:

My father grew up in Bronx and Manhattan, as Arnold Grudinsky. His father Joseph was born in Philadelphia, shortly after his parents arrived from Austria-Hungary. They moved to New York. My grandfather left school after second grade to work as a messenger boy. He had diverse jobs through his life. The most significant was building super, or maintenance man, for a five-floor Metropolitan Life building that had a different business on each floor, driven by power from a central shaft. He did the overall maintenance.

Grudin:

My grandmother pushed my aunt to become a successful dancer in New York, and pushed my father into college. He cut classes to play chess, dropped out. He washed out of a few jobs that his father arranged in the Metropolitan Life building, and joined the army in 1939. One day, near the end of his enlistment, he came out of the barracks to find a soldier crying. It was Pearl Harbor day, December 7th 1941, and they wouldn't get out until World War II ended. He served in the US and Trinidad. He helped with weather forecasting, which along with looking for submarines was one of the main things that, I guess, they did around there. He became interested in meteorology.

Grudin:

A turning point in his life came when he was a sergeant. He was assigned to put a special platoon through basic training. They were all engineers taking new tools and weapons that they had designed to the front, for testing and training others in their use. My father had to make these guys shine their boots and so forth. But he was more interested in talking about their work. As the platoon was shipping out to Europe, one guy handed my father a book, A Survey of Modern Algebra by Birkhoff and Mac Lane, published in 1941, and still in use when I studied algebra 30 years later. He read it.

Grudin:

After the war, he quickly got degrees in mathematics at NYU and Columbia, where my mother was getting a masters in education. They met at intense political gatherings. My father in his youth was an anarcho-syndicalist. My mother and college friends had previously traveled from Maine to New York, to try to join the Communist Party. (Laughter.) The USSR was our ally at the time, and they didn't see a problem with this. As I recall, she said that a burly union organizer listened to them, put down a cigar, and said, "Girls, go back to college."

(Laughter.)

They did, but my mother later became a well-known political organizer—not a communist—who spent her life in Ohio raising kids, teaching school, working to help people, taking care of a man who disappeared into mathematics. She also became my best friend.

Grudin:

I grew up seeing my relatives infrequently as they spread from Massachusetts to New York, Florida, Nevada, and other states. My father began his PhD studies with a mathematician who was investigated by the House Un-American Activities Committee and fired. He found a new topic and advisor at the University of Colorado. Finishing his degree with stressful. He never ceased being passionate about reading and exploring mathematics. He was creative, but he never got over publication anxiety.

Grudin:

He took a job at Denison University, which is in the small Ohio village of Granville. He was very content there. He was invested in students, he ordered books for the math library, he read every one, and taught every math course in the curriculum. He had no interest in salary level. He was thrilled to be paid to be doing math. My mother had to worry about the family finances.

Grudin:

While in Boulder, he took the abbreviated name Grudin that his parents had adopted. I was born, followed by a brother. We moved to Ohio when I was three. I reportedly was not interested in a few of my mother's efforts to teach me to read. She concluded I would be a late reader, until one day they found me reading a book aloud to my younger brother.

(Laughter.)

I don't remember that, but I do remember that when I was about four, I grilled my mother endlessly about “how Santa Claus did it.” She fended me off, she thought I would quit when Christmas was past. But when I resumed, she gave up and leveled with me, on condition that I wouldn't tell my brother. I just said, “Oh.” I may not have consciously thought it through, but a lot could have been absorbed about the value of logic, and persistence, and good people lying for a reason.

Grudin:

Years later, I saw that central Ohio is beautiful, positioned on the rising hills at the southern edge of the terminal moraine. Granville had an interesting history that we were never taught. All I knew at the time was that I didn't fit. The area was low-socioeconomic, it was officially Appalachia, we had two cent milk in school. It was extremely conservative. It was dry—no alcohol. Granville sponsored, and still sponsors, a Fourth of July carnival and parade that drew people from the area. In my high school years, the award for the best float in the parade went to the John Birch Society. The John Birch Society considered Republican President Dwight Eisenhower to be a tool of the international communist conspiracy.

Grudin:

My parents could hang out with a handful of liberal faculty families. My mother organized Democrats in the county and state while teaching in the small diverse city of Newark, 10 miles away. But I was in an unchanging school class of about 100 in each grade. None were Black, Latino, or Asian American. It was a public school, but each day began with the Lord's Prayer over the intercom-slash-surveillance system. We had religious assemblies. Many topics could not be discussed.

Grudin:

I had two great years that opened my eyes to the world, which with no television, much less internet, we were cut off from. When I was in sixth grade, my father was on sabbatical at MIT. I learned about the theory of evolution, which I kept quiet about when I returned to Granville. (Laughs.) I once was lured into talking about it, and it ended badly. (Laughs.) In 10th grade, my father was on leave, to teach at Reed College. Over the year, I was stunned to twice hear a high school student say that he didn't believe in God. I'd come to that belief myself, probably through a Santa Claus process that I don't recall. But I never told anyone including my parents. They took us to Quaker meetings for years, to fit into the community, but we never discussed religion in the house.

Grudin:

Granville schools did have a positive: Limited employment opportunities for women in the area meant that talented Denison faculty wives were among the teachers. So the education was probably better than in the surrounding schools.

I was always impressed by kids who were experts on any topic, chess or bridge or coin collecting. One student, Gary Warner, from a poor family, he was a poor student, too, but he had an encyclopedic knowledge of guns by fourth grade. For several lunches, I listened to him lecture, which was perhaps one of his best school experiences. I've always assumed that everyone knows things that I don't. Everyone is worth talking to and listening carefully to. This sounds like a good trait, but it's partly based on a lack of confidence, feeling that I don't know things I should. Although I was a very strong student, I always also felt a slight inferiority complex, a sense of being aware of how much I didn't know, and I worried about it.

Grudin:

Granville resembled the Pennsylvania town in John Updike's novel, Rabbit, Run. It focused on school athletics, not academics. There were seven sports, so each of the 150 high school boys was pressed to participate. For me, it was wrestling and tennis. I was in the chess club and organized a political science club. I represented the school in math and English in annual state competitions. In 1968, my senior year of high school, student rebellion reached Granville. When my role in an underground newspaper was exposed, all my academic awards were taken away. I was threatened by the principal with letters to colleges to get my acceptances rescinded.

Grudin:

Okay, what about computers? In the early '60s, my junior high school club took a field trip to see a huge vacuum tube computer. It had been donated by the Battelle Institute in Columbus, Ohio to a local college. It filled a large room. We were told it could compute missile trajectories and play blackjack. Communication was by teletype and keyboard, there was no display. You could see some of the vacuum tubes were blown out. We each had a chance to play blackjack. When I played it lost, but it announced it won.

(Laughter.)

That's where my fascination with computers began, and also told me that they aren't infallible, which was a good lesson.

Grudin:

A few years later, Denison leased an IBM computer, I've been trying to remember the model, maybe 1130, and placed it in a glass-walled room for passersby to admire. It was the size of a very large desk and accompanied by other desk-sized devices in the room: a card reader, a few card punch machines, an impressive mechanical card sorter. After a couple hours’ instruction on Fortran, I was on my own. My first program was to find twin primes. Another dealt random bridge hands that I provided to a local duplicate bridge club. My one program for a high school class calculated and printed results from a political science survey of students. The report looked quite impressive. It led to the administration banning student surveys. (Laughs.)

Grudin:

In high school, when asked what I wanted to be, I said “physicist,” even before I knew what physics was. It was considered the most challenging science, so that's what I thought I should take on. Secretly though, I hoped to be a writer. At high school graduation, I had a copy of Isaac Asimov's Foundation trilogy under my gown to read when waiting. I wrote science fiction stories that were long on science and very short on character development.

Grudin:

My senior year, my parents arranged for me to take a career aptitude test at Denison, the most pleasant test I'd ever taken. It asked, like, which of three activities I'd prefer. One section asked me to rate my attraction to a long list of professions, assuming no physical or financial barrier. I had to ask for definitions of physiologist, psychologist, and psychiatrist. They sent the bubble chart off for computer processing. A couple weeks later, I returned for the results. I was hoping for writer or physicist. Nope. The counselor said, "We rarely see someone so strongly matched to just one profession." Meaning none of the others are likely to work out, I suppose. "Please, promise me that you will take a psychology class in college."

Grudin:

In reflecting on growing up, few if any kids today have the degree of information scarcity and isolation that we grew up with. We had no television for 12 years, there was no movie theater in the village, a very small library. At some point we started getting the New York Times shipped by train, and Newsweek magazine by mail. We had a radio. At night, with the ionosphere bouncing signals from distant cities, I found baseball games, and I happened across the first Cassius Clay heavyweight fight—Muhammad Ali.

Grudin:

In the hall outside Denison’s student-run radio station, an AP teletype clattered incoming news. The building was unlocked and deserted at night. I went to stare at what was happening, minute by minute, in the world. My best friend's older brother wrote a letter about life in California. We printed it in our underground newspaper. It was a complete revelation.

Okay. Reed College.

Lewis:

If I could hold you up a little bit. You said a couple of times you had no TV, was that your family had no TV or there was no TV in your area?

Grudin:

My family.

Lewis:

Uh huh. Okay. I share that with you, I'll note.

Grudin:

Yeah, some families had televisions. We rented one, I think for the Cuban Missile Crisis...

(Laughter)

Grudin:

We got one when I was in junior high school.

Lewis:

All right. And then-

Grudin:

I lost a year or two trying to catch up.

Lewis:

You mentioned that your awards being taken away because of something you did in the student newspaper.

Grudin:

Yeah. It was an underground newspaper.

Lewis:

Okay. Would you care to say anything about what that was about?

Grudin:

Well, we printed a good bit of stuff, including the letter I mentioned. The main thing was that the principal of the school had a habit, when he got angry with students, he would call them into his office, scream at them, and then threaten them with severe repercussions if he ever heard about this from parents or other students, because he would know we had talked about it. We exposed that. Also, we wrote about the student body election. My senior year, the junior running for president was not somebody the administration really admired. Through a series of stages, they shut down the election and ended up canceling it. They first took him off the ballot but allowed a write-in and when it became very clear that he would win the write-in...

Grudin:

When it came time for an election, normally people took it very seriously, but my freshman year one student ran. He had this hilarious speech, his platform was like, three day school weeks, a whole range of things. Afterwards—I was a very serious, sober student—I walked into algebra class and Mrs. James said, "Wasn't Ron Kasson a hoot?" I looked at her and said, "Yeah.” I said, “I voted for him." She stared at me in astonishment. Then Doug Keenan said, "Yeah, I voted for him too." Another kid voted for him.

(Laughter.)

He was elected. He lasted about two weeks, the next fall, until he was caught playing hooky with his girlfriend. It was an interesting time.

Lewis:

Go ahead.

Grudin:

Although the underground newspaper had pretty severe repercussions for us, or at least, locally severe, the principal did not last another year. He was pushed out, moved to a school in Florida...

Lewis:

Yeah, so Reed College.

Grudin:

Reed College. In 1968, Reed had a lot of students from the East Coast, many from prep schools, and a lot of Californians. For me, it was exciting, but intimidating. I wandered around the Reed campus my sophomore year of high school when my father taught there. I saw discussion classes on the lawns. I knew it had a good reputation in the sciences, so it seemed idyllic. I didn't realize that the lawn discussions were all upper division humanities classes, not science classes.

Grudin:

A third of everyone's first year was a required humanities course that started with The Iliad and The Odyssey and concluded with Shakespeare. I was stunned, we could read Plato and Aristotle without weeks of preparation? I could have read it in high school? I liked the course and my discussion group leader, Diskin Clay, a classics professor. My first all-nighters were for the biweekly writing assignments. My other two courses were accelerated calculus and physics-chemistry.

Grudin:

The first semester was tough. The physics teacher had a strong French accent that I couldn't follow at all. But I loved the chemistry semester, which included learning to operate Reed’s small nuclear reactor. I caught up on physics sophomore year, where we went through all the same material, with calculus behind us. I became a math-physics major.

Grudin:

However, the culture shock was really too much. I went from being a Granville radical, to what I jokingly called a Reed fascist. Only the Vietnam War era draft deferment kept me there. The previous spring in Ohio, I’d managed a countywide presidential primary voting operation. In the fall, at Reed, I worked at one point 72 hours with no sleep on the Senate campaign of anti-war Senator Wayne Morse. But my classmates were hostile. “Don't vote, it only encourages them.” (Laughter)

Grudin:

Soon after my arrival, my freshmen and women cohort shut down the college for three days. They wanted to drop Greece-to-Shakespeare humanities and design our own curriculum, which would include Southeast Asia. I was happy reading what was assigned. I had no idea what we might replace it with. I liked Thucydides. Later in the semester, Black students occupied the administration building, demanding a Black Studies major. I had no problem with that. But I was shaken when they invaded the physics class and confronted me personally, calling me a racist for attending classes, or maybe just for being me. The only Black kid I'd ever known, named Paris, became a friend at a summer YMCA camp around third grade. We exchanged letters. My mother paid a Black housekeeper to come from Newark and paid her Social Security tax and when home we would eat lunch together. For a while my mother drove me to Columbus to tutor a young Black student in math. When we were in Massachusetts, a friend of hers was married to a Black man. I was a strong fan of Muhammad Ali. But I didn't just resent being called a racist, I was wounded and, and, just felt out of place.

Grudin:

However, I liked political debates. There was one fellow in the Students for Democratic Society, the SDS, who brought new recruits to my dorm to argue with me as part of their basic training. I wasn't the only fascist, which really meant traditional liberal. The tenured faculty and older students turned the tide. A religion major was created, not a Black Studies major. Some SDS members were expelled. Several young faculty were let go, including my humanities professor. I was shocked by the vindictiveness of my allies, who stopped being my allies at that point. My class had the highest attrition rate in Reed's history. Over the years, my views shifted steadily. I ended up loving most students and skeptical of the change-resistant faculty and administration.

Grudin:

The summer after my first year, my brother asked me to drive him to a music festival in New York. Woodstock was a revelation, not only for the hours of musicians jamming, but because the chaotic festival was planned and managed by people not much older than me. It was empowering. My generation could do anything. That optimism never left me, although we didn't produce the utopia that I was convinced we were capable of. I came away from my first year inspired to look around to see what was broken, to work to understand it, and not rely on authorities to take action. To look for patterns and apply logic to identify causes and effects.

Grudin:

My second summer taught me that technically flawless research might not be scientifically motivated. Tektronix was the leading oscilloscope manufacturer. It was founded by a Reed graduate. At one time, it was Oregon's largest private employer. I visited when I was in high school and saved babysitting money to buy Tektronix stock. I spent the summer measuring phosphorus decay characteristics using precisely timed photographs, one after another. I proposed, instead of a series of photographs, each on a different trial, that we use high speed video. That was introduced the following year, when I was not there. The problem was that no model suggested that any of these phosphors would be interesting. We were testing all possible compounds on the chance of finding something unexpected.

Grudin:

Reed was in a residential, not a commercial, area. Many students congregated in one activity hub on campus. A bridge- and hearts-playing crowd that I joined was a social leveler for someone who was shy and unsophisticated. Beer was a discovery, an inhibition suppressor that made dancing fun. I read avidly and fell in love with film. I took philosophy, literature and even music courses.

Grudin:

But I reneged on my promise to take psychology. Why was that? (Laughs.) My first year was full, but to get a head start, I looked for psychology. I read Freud's The Wolfman and Civilization and Its Discontents. I was appalled. “This isn't science!” I picked up the intro psych textbook. The Reed psych program at that time was Skinnerian. Each student was issued a rat to train. One author of the textbook was Skinner's collaborator. As long as the book was describing reinforcement and extinction and all that on animals, it was quite intriguing. But then, I think it was in chapter four, they wrote, "Okay, now let's apply these principles to human beings." And they described a new mental hospital in which they’d begun applying Skinnerian techniques.

Grudin:

Their example was a woman who had a very idealized version of marriage. After she got married, she became very depressed. She reached the point where she would only wear white clothes. She would only eat white food, and she wouldn't talk to her husband. She was committed to the institution. The first thing they did was take away all of her white clothes and give her something brown to wear, which she immediately tore up. Then they left her with a needle and thread so that she could fix it when she was ready to come out. She could get non-white foods freely, but she couldn't get white food unless she paid for it with tokens. She got tokens for talking to people or doing other pro-social things.

Grudin:

It's all based on a token economy with reinforcement. No discussion. They made a big point of never discussing her issues with her. I was totally appalled. They didn't even have an outcome yet. It was described as a work in progress. I shut the book. I just today checked and found that a second edition was published a few years later, so maybe there was an outcome, but there are no used copies of either edition on Amazon.

Grudin:

Although I dismissed what I'd seen of scientific approaches to psychology, Arthur Koestler's Act of Creation impressed me tremendously. And I also respected the psychological insights of novelists and philosophers, like Bertrand Russell, for example.

Grudin:

My senior year I wrote articles and a column and coedited the college newspaper. I saw journalism as a possible career path. But I couldn't plan. I faced probable conscription on graduation. My draft lottery number was 100, a low number, and the future seemed bleak. Our marches and protests were not deterring Nixon. We all talked about it a lot, about the bleakness of the future, including preferred suicide options. I seriously gave myself no chance of living past 30.

Grudin:

But I liked math, to which I'd shifted from physics, doing a thesis on set theory. I was accepted for graduate school by Stanford and Purdue. Purdue had good set theorists and offered support. On one school break, I visited San Francisco and absolutely loved it. An argument for Purdue was that it would be pretty much just me and math, so I could decide. And before reaching Purdue, I spent the summer hitchhiking around Europe, deciding whether to return to face the draft. I didn't show up for my draft physical. But they stopped at number 95. I left the PhD program at Purdue after a year with a masters.

Grudin:

I really liked math. I felt that I was getting more attached to it. I especially liked teaching. I was given courses to handle without a professor, unusual for a graduate student. But I saw that full-time math would take me down my father's path, or it could: diminished engagement with people and politics. It was just a world of its own. Also, I roomed for a few months with a friend who was a phenomenally more talented mathematician. My mother's influence was also there. Watergate happened. I did what I needed to get A’s, but I watched all the hearings and discussed them endlessly with friends and family. I was writing for an underground newspaper in Indiana. And in the fall, I left for Boston to look for a newspaper job, crashing with friends in Cambridge.

Grudin:

How did that go? Suddenly journalism with Woodward and Bernstein seemed to be popular. My only interview was with the Christian Science Monitor, where one question was, "Are you a Christian Scientist?" Soon I was broke. Then I won one hundred dollars, one hundred 1973 dollars, in a chess tournament. I also flew to Granville to attend my best friend's funeral. Bob was super-sharp and creative. He wasn't driven by external approval. He’d been teaching high school in Oberlin, Ohio. He made it to 23 and then swallowed a bottle of sleeping pills, and when they had no effect, he bought a gun.

Grudin:

Upon returning to Boston, I bit the bullet. At Reed, many students approached graduation as “falling off a cliff, looking for tree branches to grab on the way down.” Law school and computer jobs were branches. The Boston Globe had classified ads for computer operators and programmers. I wasn't qualified for system analyst, the other computer profession. As you can remember, operators managed the machine, loading programs and printers, pushing buttons, pulling switches, bursting printouts and so on. Kind of menial, but as it happens the only job with direct human-computer interaction. Programs were written on paper coding sheets.

Grudin:

I kept getting turned down for jobs. Sears operator job. "Will you wear a tie every day?" I said yes, but he could see I was wearing a clip-on tie. (Laughs.) Most humiliating was the Harvard Physics Department. The programmer I would work with really liked me, but the professor grilled me on advanced physics I'd covered in one course three years earlier. It wasn't my nature to laugh and tell him I could easily handle it if given a few hours to review. The most frustrating, but astute interviewer was a woman at Draper Labs, who was hiring an operator. She said, "I could hire you, but I'll do you a favor and not, because you're overqualified and could do better."

(Laughter.)

I didn't consider it a favor, until I got a great programming job at a small startup tech company in Tewksbury, Massachusetts, called Wang Laboratories, which turned into a very large and successful minicomputer company some years later.

Grudin:

I got the job by ignoring the requirements. Located out in the country with no public transport, the ad said, "Must have a car." After I got the job, I hitchhiked to work until I got a paycheck and could buy a bicycle. It was a snowy January, but it worked out. I was ecstatic to be paid to have fun writing programs. For that reason, perhaps, I paid no attention to salary, as had my father, and as do some young people today, although perhaps not in the tech field.

Grudin:

At Wang, I wrote hardware and software diagnostics and I/O controllers, input/output controllers. We might call them drivers now. We saw ourselves as systems programmers, and considered application software a bit soft. At one point, though, I was writing a controller to handle up to four terminals with people typing, and I wondered, "How fast do I have to service each keystroke?" Each keystroke could alter several or a good number of screen characters and the 8-bit processors were slow. I asked one of the fastest secretaries how fast she could type. Rhonda said, "Oh, well normally I type so and so words per minute. But when I really get on a roll, I can get up to 300 WPM." I went back and designed the controller to handle characters coming in at that rate. I later discovered that she was about an order of magnitude too optimistic about how fast she or anyone really typed. (Laughter.)

Grudin:

The thought crossed my mind that there might be value in trying to understand actual behavior, rather than just asking people, but I didn't know anybody who did that. I guess not many people did in 1975. Probably some of your colleagues at IBM, which dominated the field at the time. Wang had built some of the first digital calculator—I'd used one at Reed—and was selling electronic typewriters, which were IBM Selectrics with a chip added. As a competitor, IBM gave us no volume discount on Selectrics, even though we were buying hundreds of thousands of them.

Grudin:

My group was working toward the word processor that ended up turning Dr. Wang into one of the wealthiest Americans. One day, I came in to hear my two favorite managers discussing the news that Xerox was jumping in to compete with IBM and Wang on office systems. "What does this mean for us?" asked Dave. "Nothing," said Harold, the more senior manager, a VP at the time, or soon to be a VP. He said, "Two elephants are fighting, and we're a mouse." He paused and then corrected himself. "An elephant is fighting a dog, and we're a mouse." (Laughs.) Even our very conservative finance officer was hoping that the government would break up IBM's monopoly, an effort that had started but never happened.

Grudin:

Visiting friends outside Los Angeles in November reminded me that there was more to the world than frozen winters. When I got back, I told my manager I'd be around until the next fall. When I left, I spent a few months visiting relatives and family friends around the US before driving south from Laredo, destination Rio de Janeiro in time for Carnival, which I had been very impressed with seeing in the movie, Black Orpheus. I didn't make it, but I had adventures ,and I saw extreme poverty. On a hike in Mexico or Central America, I came across acacias with thorns that had small holes. I tapped one, and faster than expected, having read about them in E.O. Wilson's Insect Societies, small ants swarmed out and bit me. I didn't realize how small they were when reading the book.

Grudin:

At Wang, I'd become interested in psychology. A friend loaned me A.R. Luria's short book, The Man with a Shattered World, about a soldier with a brain injury and a remarkable set of strange problems. Luria argued and showed that by collecting such accounts of brain injuries and identifying the affected brain structures along with the behaviors, it would enable a science of psychology. I was thrilled. This was the scientific psychology that I hadn't found before.

Grudin:

Another revelation came from Wilson. The ant that bit me wasn't thinking, "An animal is about to eat our home." That and much more complex ant behaviors are wired into their genes. My previous optimistic view was that children raised in a just and equitable culture would perpetuate it. But that might not hold up. We really need to understand human genetic predispositions. I ordered and read books on psychology and anthropology. While traveling, I visited California campuses, and applied to take psych courses, outside of a degree program, to decide whether grad school made sense.

Grudin:

The state-subsidized UC campuses set conditions on taking courses outside a degree program. Stanford said, "If you pay, do what you want." So in March, I concluded my travels with a ferry from Puerto Vallarta to Cabo San Lucas and drove up to Palo Alto.

Lewis:

Let me pause you there for a sec to jump back for a couple of things. One, you've mentioned chess a few times. Most recently, when you won $100 in the chess tournament, when you were broke in Cambridge. Care to offer any other thoughts about chess?

Grudin:

Well, my father was a chess player and taught us chess. I played it in high school, I was on the college chess team and continued to play. I really liked postal chess. I signed up for some postal chess tournaments. They gave you as much time as you wanted to analyze moves. Being a little bit on the conservative side in some respects, making sacrifices was something that could be difficult for me, unless I could be pretty convinced that it had a good chance of working out, which I could work through in postal chess.

Grudin:

However, when I started traveling, I turned over my postal chess matches to my brother who had taken it up. He was playing for me. He took it very seriously. He got very engaged and studied it intensely, ended up a better chess player than I was. That gets at my positive impression about people who really dig in and study something intensely, because I didn't do that too much in chess. I did not go through books of openings and books of endings. What I did do was pick a fairly unorthodox opening, Bird’s opening, and really learned that one. Very few of the people that I played were able to play against it if I was playing white. But I liked chess.

Lewis:

You still play?

Grudin:

I haven't played it in some time. I played some with our kids...

Lewis:

The other thing was, in your notes when you're talking about getting the job at Wang, you said you got that through luck. What did you mean by that? There wasn't any strange coincidence, an advertisement blowing down the street or something like that?

Grudin:

No. If there was luck involved, the luck was that you couldn't get there by public transportation. I had no money for a taxi or anything like that. I looked at the map. Saw how close I could get by public transportation, this is in December, and set out really early to give myself like an hour or so to walk to Wang Labs, maybe more. But there was no sidewalk, just frozen mud. I got there just on time. If I'd cut it any shorter, I wouldn't have made it.

Grudin:

The other luck was that they hired back then with the IBM test, just pattern recognition and so forth. They didn't give me a programming test. I think I would have done okay on a programming test too. But I was a good match.

Lewis:

Then one more thing before going to Stanford, or checking the time or whatever. So you didn't make it to Rio. How far did you make it? And what did you get and what made you turn back?

Grudin:

The original plan was, I would join a college friend who after getting a masters in physics at Montana State joined a seismographic services corporation. He was going to be sent to Venezuela. We both liked the movie. We were going to meet in Rio. Shortly before I left, or maybe a few months before, I got a postcard saying, “I was double-crossed,” from Libya. Instead of sending him to Venezuela.

(Laughter).

I was on my own. I had told everybody I was doing it. I considered backing out, but I didn't want to do that.

Grudin:

I didn't know Spanish or Portuguese. But I made it down through Mexico into Central America. And then when I was in Guatemala, actually my first night in Guatemala, there was a huge earthquake, one of the worst in Central America history. The roads were destroyed, buildings were destroyed. The next day, I saw more dead people than... Well, I'd seen almost none before. I saw a large number. I made it into Guatemala City, I gave a ride to some UN folks who needed to get in there. But we had to go overland because the main highway was down.

Grudin:

After a little bit in Guatemala City, I made it down to Salvador, and thought things over and decided it's time to head back. But I had gotten interested in Mayan ruins and I spent a long time getting back by traveling around to see archaeological sites, revisiting the Museum of Anthropology in Mexico City. And I think I was in a little bit of... I think, some of my intense interest was probably just denial of what I'd been seeing in the devastation and the death. I also was a little worried about how well my car was going to hold up after bouncing around.

Lewis:

What kind of car was it?

Grudin:

It was a Dodge Dart, one of the last ones with a real steel motor.

Lewis:

Oh, wow. I'm impressed.

Grudin:

Yeah. And when I got to Stanford, I was finally safe, back, and I had some stress-related stomach issues, which they eventually diagnosed and said, "We almost never see this in anybody under 40." But I think that was, again, it was displacement from the stress of what I'd been through.

Lewis:

Okay, so we're at the hour, Do you want to go for another segment or do you want to schedule one later? I'm easy for a while. So let's see, we left off, you're just arriving at Stanford.

Grudin:

We could talk about it. I could do Stanford and then we could take a break.

Lewis:

Okay, that sounds good.

Grudin:

Let's see. When I arrived, it was in March. I'd kind of narrowed things down to anthropology or developmental psychology to get at these fundamental issues, getting at the core of what's human. I’d chosen psychology because that was a more plausible entrance point than anthropology for somebody who had not studied either as an undergraduate. I'd been thinking about developmental psychology, not quite realizing how hard it is to do good developmental psychology studies to get at issues like this.

Grudin:

But when I arrived in spring quarter, there were no undergraduate developmental psych courses. Stanford would let me take as many courses as I wanted. I took every moment that I could schedule, I took enough courses to fill every possible moment. I've never worked as hard as I did at Stanford. There was one time slot where you could take either human memory or history of psychology. I thought history of psychology sounded more interesting. so I got the book and read it before classes started. Then said, "Well, since I've already read this, I might as well take the human memory class instead.”

Grudin:

I didn't know that cognitive psychology existed. None of the Scientific American psych books that I'd been reading talked about cognitive psychology. But people told me, "If you're a programmer, this is what you should study. This makes sense." I let myself gravitate to cognitive even though ultimately maybe developmental or ethnography might have appealed to me more. I was working hard. Most of the classes were undergraduate. I took a couple graduate, such as math psych. I impressed the TAs and some faculty and became friends with some of them. I’d saved enough money to get through two quarters, then got a job as a programmer and lab assistant on a joint MIT-Stanford project, and was there for another year still taking courses.

Grudin:

The Stanford students were incredibly inspirational. They were doing all kinds of fascinating things, looking at dreaming and split brains and hypnosis and creativity. For the latter, we interviewed James Watson who was visiting, who said things of the sort that led to his intellectual banishment. Unfortunately, the faculty didn't like what a lot of the students were doing. A number of the students, including myself, formed a discussion group that met in the evenings to which faculty were not invited. Even older graduate students weren't invited, which some resented.

Grudin:

There was no HCI field with a cognitive focus in '76 or '77. But many cognitive and developmental grad students were using minicomputers, the now extinct species, which were cabinet size but low-powered by today's measure. They were learning to program, but more importantly, they were using computers hands-on, observing good and bad interaction design. The faculty at Stanford weren't. Many of the students went into HCI later, which generally meant abandoning academia, at least for a while. Included in this discussion group was Tom Malone, George Furnas, Jeff Johnson, one of the Cathy Marshalls, several people that did not stay in HCI too long, as well as some brilliant students who didn't go into HCI. It was an extraordinary group. Most of them were at odds with advisors who were stuck in experimental paradigms that were wearing out.

Grudin:

One downside was that Stanford provided my first examples of significant academic gender discrimination. The most creative and deepest thinker among the students, in my opinion, was Justine Owens. She and George Furnas were TAs is in my human memory course, as it happened. While I was there, she devised two experiments that had a strong impact on the study of memory. The first showed that providing even a little preliminary context greatly affected story recall. Her advisor discouraged her research until he saw the results. He then helped write it up and asked her to add, as co-author, an older grad student who was on the job market. His talks culminated in her work, without attributing it to her. And that was just the beginning of Justine's time at Stanford.

Grudin:

Since then I worked with women who did well in academia, but three of the most impressive scholars that I encountered, encountered clear discrimination. This limited my admiration for academia, which is not to say that industry is better.

Lewis:

Wow.

Grudin:

While I was at Stanford, Don Norman and Dave Rumelhart came up from UCSD to speak. What they were doing was a lot more exciting to me than what the faculty were doing at Stanford, so although I had a lot of friends at Stanford, I entered the formal PhD program at UCSD in 1977.

Lewis:

Who taught your human memory class, do you remember?

Grudin:

Yeah, Ed Smith.

Lewis:

Oh, no kidding.

Grudin:

If we're sort of wrapping up for now, I'll just tell you another example of how chance can... I mean, you can make your own chances. I don't know if this is that remarkable, but at Stanford, there were two lunch places near Jordan Hall, where psychology was, in Tressider Union. One had custom-made sandwiches and very appealing decor. The other one was a basic cafeteria, you could get hotdogs. The students, all of us, went to the custom- made sandwiches and salads shop. However, Ed Smith, who was from New York, had grown up in New York, really liked hot dogs and similar food. He would go to the other one. No students would accompany him, even though normally he was besieged by students. He was very popular, he always had students around him.

Grudin:

As soon as I noticed this, I made a practice of going and sitting and talking to Ed. He became a mentor. He got me the job as lab assistant and programmer. He went off to Rockefeller for a year or so, so he was not around. If he had been around and taken me on as a grad student, I would have stayed, I think, I really liked him.

Grudin:

But he came back just before I was heading to UCSD. He knew I applied for... I got into MIT, I got into Oregon, I got into other places. He asked where I was going. I said I was going to UCSD, and he said, "A good choice!" He said “UCSD believes in psychology," I think he said. "Stanford believes in power." (Laughs.) Ed did not stay at Stanford much longer. He was tenured. The other thing he asked, "What are you going to do when you get there?" I said, "Well, I'm going to get through as soon as possible." He said, "That would be a big mistake. Wherever you go after UCSD, is not going to be as good as UCSD, so you should stay there as long as you can."

Lewis:

I don't think I ever met Ed Smith. But I’ve known the name forever. I think he did his PhD at Michigan, which is where I was a grad student. And so that was one of the--

Grudin:

He went back there

Lewis:

Did he go back later?

Grudin:

Yeah. He went to BBN, he was at BBN for a while and then went to Michigan.

Lewis:

Okay, well, that was after I was gone. But he was one of the people that I remember hearing about as a student as somebody who's doing interesting stuff. More than that I don't really remember.

Grudin:

Yeah. He was impressive. Ed, at one point asked me, "What do you make of Kahneman and Tversky?” He’d assigned a couple of Kahneman and Tversky papers. I said I was impressed. And he said, "Yeah, I was impressed." But he wasn't asking the question rhetorically, because it was just when they were first starting to publish.

Grudin:

However, the interesting thing about that class, there were about 100, 120 of us in the class. I thought it was a fantastic class. He went through the history of cognitive psych, the progression of cognitive psychology, through the box models. The slow development of box models representing speech processing and visual processing, and slowly all the experiments that were done that created more elaborate box models to explain the phenomena about how mental processes might be organized. He went through them all, developing the more and more elaborate box models and giving us the papers. And then he assigned the levels of processing paper, which said, “throw out box models.” A different approach to explain cognitive functioning.

Grudin:

Now, I thought it was fantastic. I thought it was such a brilliant thing. Some time later, when I was talking with him, he said, "That class turned out to be a disaster. I got the worst teaching reviews—"

Lewis:

Really.

Grudin:

—I’ve ever gotten. Students wrote, “You taught us all this stuff that was wrong.

[Laughter.]

You made us learn all this stuff that was wrong.”

Lewis:

Oh gosh.

Grudin:

Yeah, it was an interesting time. When the first quarter ended, the first thing I did was go out to a movie because I had not had a moment when I wasn't studying or in classes. I'd become a big movie buff and that was the first break I took.

Lewis:

What was the film?

Grudin:

Not a good film.

[Laughter.]

It was the only film that was nearby, it was The Omen. It was a really bad film.

Lewis:

I've never seen it. One other little stitch here. You mentioned that you got a job, sounds like a halftime job at this MIT-Stanford project. What was that about and what were you doing?

Grudin:

It was one of the exhausted paradigms. The programming was, some of it was statistical analysis. But the main thing was tachistoscope studies where... I ran thousands of people, or at least hundreds, probably a thousand. People would see two cards, two words or images, one might be an image of a carrot and the word carrot, or an image of a carrot and the word vegetable. And they would have to click Yes or No with different instructions, like, are these both in the same category, are these both the same thing? They might both be images, one might be an image and one might be a word, or both words. There were lots of variations of that.

Grudin:

I had to load in the two cards, then push the button to say go. Then I read off of the T-Scope display the four-digit number of milliseconds that were taken. I had to enter that.

Grudin:

Later, another student was doing a study on memory for digits and reproduction of digits. I was incredibly fast... Normally, the more digits, the slower you get. But for four digits, I had a dip. I don't know if he threw out my data. I had this dip for four digits, because I could speed the study up if I memorized it while I was taking out cards and putting in the next cards, and then write it down when I had a little bit of time. So I went through that memorization.

Grudin:

Two things happened that were of interest. Parapsychology had sort of had a revival. There were people who were interested in it when I was an undergraduate, they had the [Zener] cards. Physicists turn out to be unusually willing to believe parapsychology. Some of my physics friends at Reed were big on parapsychology. I wasn't so big on it. It was having a revival in the '70s. Doing these T-Scope studies… basically, studying experimental psychology convinced me that people could not see the future, they could not see across space, because student reaction times… these cards are coming in and they weren't able to predict what they were, they had to process... I felt that it was very strong evidence that there aren't people around who can foresee stuff.

Grudin:

However, I had two sort-of counterexamples. One guy started getting extremely low reaction times, and I was mystified. I thought about it, about the whole experimental process. I realized that I was holding the cards that were going to come up next, before I put them in, down low, and if he leaned down, went out of his way to lean down, he could see what they were before I put them in. So I stopped putting them down there.

[Laughter.]

Suddenly his numbers went completely slower than the normal. I kind of stopped the experiment a little bit early after a handful of these. He was kind of sweating a bit. I just thanked him for his participation and explained what it was all about and let him go.

Grudin:

The other exceptional person was one young woman who came in. She was regularly like a couple hundred milliseconds faster than almost anybody else, faster than anybody else. I didn't suspect anything there. Sometime later… I was interested in hypnosis, and Ernest Hilgard was down on the second floor doing hypnosis studies, which were very scientific hypnosis studies. He was the only faculty member we invited to our private discussion group on occasion. In fact, he invited us all to meet at his house.

Grudin:

I was down in the hypnosis lab and as I was walking in, the woman who had been in my experiment was walking out. I asked one of the grad students, "Do you know who she was?" He said, "She is the most hypnotizable person that we've ever had in the lab." She was doing studies for them. When I told Jerry Barzano, who was one of the grad students, what I was doing with the T-Scopes, he said, "What a waste of a mind," to have me doing that work. But it was okay.

Lewis:

Yeah, interesting. You mentioned Tom Malone as somebody in that group, and I was realizing that I was very likely at Stanford very briefly, like a day, sometime when you were there, looking at the years, and I met Tom Malone then. What happened was that one of my mentors at Michigan was Dave Krantz. Oh, actually, I realize this might have been before you got there, I'd have to do the calculation, but anyway... So Krantz had... Well, they had coauthored or were coauthoring or something. Anyway, he had a relationship with Pat Suppes. I was doing something that Krantz thought Suppes could tell me something about. I can't remember what... I was working on power law for a while. He suggested that I might want to visit Suppes, which I did. I don't actually remember talking with Suppes, I'm sure I did. But I do remember, Tom Malone was kind of my host.

Grudin:

Probably it was when I was there, because it couldn't have been too much earlier. Because he wasn't there too much earlier than I was.

Lewis:

Yeah, that's right. It was before I went to Austin, which would have been in '76. Could have been '76 when you were there, or maybe '75 when you weren't there yet, I guess. But anyway, I don't remember meeting you, so I guess it doesn't matter.

Grudin:

Well, when we get to the Applied Psychology Unit, our paths will cross. I think you probably know this, but it's definitely part of my story because you and Don Norman were both volunteers in an experiment that Alan McLean and I did, and you illustrated the opposite ends of the spectrum that we found. [Laughs.]

Lewis:

Oh, boy, I eagerly look forward to that. I don’t remember anything about that. I remember the MRC of course. But I don’t remember being in an experiment.

Grudin:

You were in the experiment. It wasn't a very elaborate experiment. It was just one session.

Lewis:

Okay, I look forward to that. Okay, well, is this a good place to stop and schedule the next session?

Graduate school and postdoc[edit | edit source]

Lewis:

I'm recording and we're good. Okay. You're on. [July 9, 2021]

Grudin:

Okay. I arrived in La Jolla at UCSD in September, 1977. Don Norman, Dave Rumelhart, and Jay McClelland managed a lab that was at the center of several new initiatives. The Sloan Foundation had begun funding postdocs and workshops to create a field called Cognitive Science, which blended cognitive psychology, AI, computer science and linguistics. The first international Cognitive Science conference was held a few hundred feet from our lab in La Jolla. Don Norman became editor of the new journal, Cognitive Science.

Grudin:

I believe he was president of the Cognitive Science Society. Later he formed the first Cognitive Science Department. Dave Rumelhart’s backpropagation methods spurred the growth of parallel distributed processing, which was a forerunner of machine learning and deep learning. The faculty and some postdocs produced influential books in this area. The lab was busy day and night. It became an early ARPANET site. Everyone could access the lab's minicomputer from terminals in their offices, which was unusual back then, outside of Computer Science departments maybe. We talked about our interaction with computers, but the academic focus was on theories of human cognition, not developing better interfaces.

Grudin:

Computers could be easily programmed to present experimental conditions and record data, so they were a great tool for looking at the psychology of a person who in this case happens to be using a computer. Don Norman's group, when I arrived, was investigating motor control. And my PhD studies led to a detailed model of highly skilled transcription typing. I was pleased with my intellectual achievement, but I didn't plan a career in motor control and I continually explored other topics while I was there. It was a stimulating but difficult environment for grad students, because the brilliant, combative Sloan postdocs competed for faculty attention.

Grudin:

This context led to defensiveness which I had to try to get over after I left. I collaborated less than I might have there, as I worried about credit for my ideas, partly due to having seen some ideas being appropriated at Stanford. In fact, I ended up interacting more with Don Norman, my advisor, after my PhD than while I was there. My first assignment was to present a current article from the journal Psych Review. It was a study of how people solve verbal analogies. I saw a more plausible model. I’d just solved a lot of verbal analogies under time pressure for graduate school applications, in the GRE and Miller Analogies Test.

Grudin:

I contacted the author of the paper, who generously shared his experimental materials, and ran studies that proved my case. A journal accepted the paper despite one hostile review, which was clearly from the author who had loaned me his materials. Correcting a senior researcher attracted some attention, but it's a double-edged sword. I later used the opposite approach. In my typing data, I identified an error in causal attribution that was a minor point in a seminal motor control paper. When I published it, my paper included the correct analysis and I cited the earlier work for other things, but didn't mention the error. Years later, the author invited me to a friendly conversation and praised my work. We never discussed the correction, but he was clearly aware of it.

Grudin:

Two events at UCSD outside the lab made lasting impressions on me. Tony Deutsch was a physiological psychologist, not known for pleasantries. At a social event, one afternoon, he prefaced some forgotten advice by starting, "When you are a professor..." I was stunned. No faculty member has suggested that, of course, I would succeed. Growing up, I heard my professor father sometimes castigate a colleague, "that idiot, Mitchell," so I didn't have an inflated view of professors, but almost all grad students have moments of doubt.

Grudin:

My experience with science at its best came at the opening evening social event of the first Cognitive Science Conference. I worked up the courage to mention to MIT neurologist Norman Geschwind that another student, Craig Will, and I had with some effort retrieved dolphin brains. Could hemispheric differences, found in birds and humans, correlate with communication skill? Dolphins communicate. Geschwind immediately abandoned his friends to walk a few hundred feet to our office and used a kitchen knife to explore, late into the night, a dolphin brain. It wasn't conclusive, but it was thrilling. Our dolphin brain collecting had mostly been regarded by others with somewhat respectful humor.

Lewis:

I can't resist. So, collect dolphin brains, how does one do that?

Grudin:

Yes. There was a marine mammal center nearby on the southern California coast. When tuna ships were out netting tuna, if they accidentally netted a dolphin, they returned it to this station where they would analyze the demographic characteristics of the dolphins that were showing up in nets. Their age, gender, condition, and so forth. They weren't interested in the brain, but they were interested in the skulls, they wanted to preserve the skulls. We had to go in and extract the brain from the skull.

Grudin:

It was a very unpleasant, I would say, experience. Craig, who had the idea, and I went at it. He started to have nightmares about dolphins coming at him. Whereas, I thought I was sort of squeamish, but I actually got right down to it and was able to carry through on getting out these brains and getting them into formalin and preserving them, which we had in bottles around our offices.

Lewis:

That's amazing. Anyway. Okay. Thanks.

Grudin:

This is an example of the things that were going on. I did other experiments, to satisfy my curiosity mostly. One was a series of experiments on visual after-images. Another, that I later regretted not extending, was emptying a bag of typewriter keys on a table and asking very skilled typists working in the department to arrange them into their keyboard positions, with no time limit. No one was completely accurate. Most didn't even get close. This didn't surprise me so much. I expected it. I moved on, but others found it more interesting and suggested I should have pursued it. For one graduate seminar, I looked at metaphors for mental processes in papers in the initial 1894 issue of Psych Review. After that, Dedre Gentner and I charted shifts in such metaphors over time.

Grudin:

The whole field seemed like a wonderful collaborative effort, as you may remember, from back then. I spent two summers at the MIT AI Lab. In what seems remarkable today, their PDP-10 was completely open. Anyone there, or anyone logging in as a guest via a modem, could read, edit, or delete any file. [Laughter.]

The system was backed up daily, but it's a reminder of how small and trusting the tech world was as recently as 1979. So where did HCI fit in? Well—

Lewis:

Okay. Before you leave that, I had highlighted that in your notes. How did it happen that you ended up doing that at MIT? There has to be a story there. It's not the kind of thing that happens every day.

Grudin:

Yeah. When I was at Stanford, deciding to go to grad school, I was working on the joint MIT-Stanford project. One faculty member, Molly Potter, at MIT, and Judy Kroll at Stanford. I had some interaction with the folks at MIT, I visited at least once in that period. When I was applying to grad schools, I applied to MIT as well as west coast universities. I was accepted. They were shocked that I turned them down. I turned them down because I was interested in these fringe areas of psychology, you might say, split brain studies and so forth. And the MIT Psychology Department was very... basically it was cognitive, physiological, and just a little bit of developmental. Didn't have the range of psychology that I wanted to explore.

Grudin:

When I turned them down, one of the faculty, Whitman Richards, called me up and said he wanted to come out to California and play a tennis match, and if he won then I would go to MIT. I think I would have beaten him. I was the captain and first singles on my college and high school teams. But instead, we settled on me doing two postdocs there, which were split between David Marr and the AI Lab… Whitman’s group and the AI Lab.

Grudin:

David Marr had cancer. I worked with Shimon Ullman primarily, and with Whitman, who was a major admirer of David Marr. So, that was that experience. It was good. We were looking at motion perception, the automatic detection of motion perception. Shimon Ullman had been a pilot in the Israeli Air Force and was very interested in automating motion detection through artificial intelligence.

Lewis:

Mm-hmm (affirmative).

Grudin:

Okay. Where was HCI in this? As I was finishing my dissertation, Don Norman and a group of postdocs and students began formally exploring human computer interaction for its own sake. I followed that work, and later published with people in the lab. But in general, the broad UCSD education paid off. Two of my favorite HCI papers, written later, resulted from an article assigned in a linguistics course and from a method that I learned from Tony Deutsch, in physiological psychology. I'll probably get to those at the point where I wrote them up. I also took two courses in UCSD's neuroscience program. I sat in on clinical exams at the San Diego Veterans Hospital, which reconnected me with Luria's approach that first drew me to psychology several years before. It also led to a postdoc.

Grudin:

Cognitive psychology swept through US psych departments, I think you were part of that, in the 1970s. But the academic job market suddenly tanked around 1980. I graduated in 1981. In a happy coincidence perhaps, the home computer and PC era was underway, and computer companies and Bell Telephone—AT&T hadn't been broken up yet, but Bell Telephone and AT&T saw this coming—wanted to sell things to the public. They needed designs that appealed to non-engineers. Many of my Stanford and UCSD friends were hired into research labs and product groups. This led to what I now think is a unique situation. A large cohort had confidently anticipated academic careers, and they took their solid research skills and motivation into industry. There, we created a field that a decade later was recognized by universities. Many of us were then drawn back to academia, including yourself. But I didn't go immediately into industry.

Grudin:

I took a postdoc at the Medical Research Council Applied Psychology Unit, as it was then called, in Cambridge, England, to study brain injuries that affected motor control, such as handwriting. It went slowly. A young psychologist from California was not considered a peer by the old British doctors who controlled patient data. They just wouldn't share it. Worse than that, Luria's appealing model of case studies of afflicted individuals wasn't practiced. The experimental approach was to form a hypothesis about the function of a brain region, run dozens of people with an injury there through a quick battery of tests, revise the hypothesis and repeat. Very impersonal. And it prevented understanding the overall effects of an injury

Grudin:

Meanwhile, Datamation published an article, The Truth about Unix, by my recent adviser. Unix was a great, great tool. In grad school, we all programmed in C on Unix. Don's paper said, "Yes, Unix is great, but it has a terrible user interface, bristling with opaque command names." Commands were a way to interact with computers before graphical user interfaces. You typed the command, the computer carried it out and it waited for the next command.

Grudin:

Don's critique drew a hostile reaction from the Unix community. Unix was developed at Bell Labs where Tom Landauer, a great guy who was later a good friend of both of us led an HCI group. He published a quick study defending Unix. Participants had an hour to learn a bunch of command names. Some names were relevant to the task, some were irrelevant words, and some were pronounceable, but not words, called pseudowords. People learned them all in an hour. Landauer concluded, names don't matter. This was wrong. In grad school, my office had been on the second floor, and the lab was on the first floor. One Unix command, I wanted to use every three to six months. It stepped through files in a folder and asked, "Do you want to delete this? Y or N."

Grudin:

Remember, no graphical user interface. It was very useful for cleaning up. The command was DSW. I could never remember it. Every time I had to traipse downstairs to look it up. Another Unix command to merge files was called CAT C-A-T. Unix command names were a mess. There was a three-person HCI group in the APU led by Phil Barnard. It was actually studying command names. Lots of people in HCI were studying command names. Phil's group was exploring, "Find distinctions between relevant names, some highly specific and some more general." I said to them, “You can do an experiment to resolve the Unix controversy. Names may not matter in the first hour, but they matter over time.” Phil said, "Sounds good, but we're busy. You can use our system to run an experiment if you wish."

Grudin:

This was my first formal HCI experiment. I tested people on day one, day two and a week later. It turned out as expected. I planned to go back to wrestling with British MDs, but something happened. The most interesting result was unexpected. This often happened in my research. I'd given participants five conditions. They had meaningful commands, abbreviations for the meaningful commands, unrelated words, pseudo words, and consonant strings, modeled on my memory of DSW. I expected people to do worst with consonant strings, which they did, and best with abbreviations, which require less typing. They did not do well with abbreviations.

Grudin:

Why? I'd assumed that the best abbreviation is to delete vowels and leave consonants, like CTRL for control. And that is good, if your job is to look at an abbreviation on a key and remember the command. In my task, they knew the command and had to remember the abbreviation. Truncation, taking the first couple letters, is a better strategy. I also realized, thinking about it, that in another context where you type commands thousands of times, a single letter shortcut could be best, even if the key isn't always the first letter of the command. We did more studies on command abbreviations. I never got back to motor control. In one study, I let people create their own commands or abbreviations once they knew the tasks, at the end of the first day. They loved being able to create their own interface, which was unheard of in that era. But the next day they did terribly with their own commands. [Laughter.]

Grudin:

You visited the APU. You remember, was it late 1982 or early 1983?

Lewis:

Yeah.

Grudin:

I was doing another study at that point with Allan MacLean. Graphical user interfaces weren't around, but Alan sometimes used reverse video to highlight a word on a display, turning black characters on a white background into white on black. It looked cool. I thought, "This might lead someone to prefer an interface, even if it was less efficient." It also gave me a chance to use a method that had impressed me from Tony Deutsch's lab at UCSD. He was studying electrical stimulation of the pleasure center in rat brains. The question that he was investigating was, what frequency of pulses, what amplitude pulses, and what duration of total stimulation was preferred by the rat. So many possible combinations. To test the combinations efficiently, Tony turned rats into expert collaborators.

Grudin:

He built a Y-shaped box with levers at each of the three endpoints. Pressing the lever at one branch of the Y delivered one combination of frequency, amplitude, and duration. The other had a different combination. And the lever at the base of the Y reset it, so a rat could run back, reset it and do it again. Rats learned how it worked. They would try both branches and settle on the one that they liked. They also learned that after a little while, the two stimuli would change. So once they picked one, occasionally they would go over to check the other one to see if it had improved. Tony could let a rat go, and home in on preferences quickly. Incredibly elegant.

Lewis:

Yeah.

Grudin:

No electrical stimulations. I tested performance versus preference on a routine, but reasonably realistic, repetitive data entry task. People had to enter a number by typing it in or stepping through a list with the current selection in cool reverse video until they reached it. For short numbers, typing was faster, for short lists, stepping through was faster. After trying both approaches, they picked one approach to do a last set, then were given a different digit length and list length. Over many sets, did their preference match performance time, when they chose the one to continue with?

We found individual differences, but most people preferred the cool UI on sets where their performance was a bit worse. On different days, while we were doing this, Don Norman visited and you visited. You both tried it, and were at opposite ends of the individual difference spectrum, one leaning toward memorizing long numbers and one preferring to scroll.

Grudin:

Okay. The result was that people sometimes preferred an interface that was less efficient. This was controversial. At that time, many of the HCI researchers who were not studying command names were building an AI-oriented cognitive model called GOMS to automatically predict task performance on a design. This could identify the best interface for performance without needing to test people. We had shown that the automated design tool might not identify the preferred interface.

Grudin:

A senior researcher at the APU told me, "Don't publish it." He said that their grand effort was more important than a minor paper. This was not the last time I got such advice, and I ignored it. However, the paper was published at a conference that went largely unnoticed. I didn't realize back then that word doesn't necessarily get out, if you don't continue to promote it, or promote it in the right places.

Grudin:

The main minicomputer at the APU was all keystroke-driven from terminals. There were no mice. Its operating system RSX crashed fairly frequently. It had an ingenious recovery mechanism. When you edited a document at your terminal, it dumped every keystroke into a temporary buffer. When you wrote the document out to the disc, it deleted the temporary buffer. When the system was rebooted after a crash, it went around looking for temporary buffers. When it found one, somebody had been in an editing session that was interrupted before a save. It went to the most recent version, if there was one, or opened a new file. It said, "We have a document on which your work was interrupted when the system went down, would you like to have it restored?" When you typed Y, it fed each keystroke from the temporary buffer back in.

Grudin:

You saw on your display the history of your last session reenacted about three times the speed you'd typed it, which was very cool. I thought, "My God, they're saving every keystroke. How can they do that? Computer memory is so expensive." Of course, memory prices were steadily dropping, but when I looked, the buffers weren't large. People don't type that many characters, even in an hour, especially if they're doing any thinking while they're typing. “If we can record everything that somebody does, why not use that information to figure out how they might be able to work more efficiently?” I was thrilled.

Grudin:

But for this, industry had better equipment and faster impact. I had two more years of postdoc, but I was worried about my health at that point. My apartment had no central heating and I'd come down with chronic bronchitis, which lasted longer each winter. Eventually, I contacted my friends back at Wang Labs about returning. One other remark about my stay at the APU: It was near the end of the pre-internet era, and it was changing academia profoundly. When I entered grad school, faculty relied on people around them for feedback and collaboration. They invested in explaining their ideas to other faculty and to grad students. They attended colloquia by speakers in other areas. When their colleagues knew and could comment on their work, leaving a university came with a cost.

Grudin:

In 1977, everything that Don, Dave, and Jay worked on had been discussed in the lab. Soon after I reached the APU, on different visits, I was startled when Don and Dave presented work I had not heard about. The ARPANET had slipped in while I was a student, and the faculty began working via email with people outside UCSD. They no longer depended on colleagues for feedback. Attendance at colloquia in different areas dropped off. Switching universities had less cost. The tribe was gone, the first of many disappearing tribes I saw over my career. It was a new world..

Grudin:

Also, the internet was just reaching Cambridge at that time. I could chat with people in La Jolla. There was no email. To reach friends there took several flaky steps on the system, and was brutally slow. But it was free. Long distance phone was very expensive then. I relied entirely on postal mail. But I had very long chat sessions late at night with Craig Will. After I left the APU, I was told that someone from the British Ministry of Defense showed up one day asking who Jonathan Grudin was. I'd used up one third of the ministry of defense telecommunications budget. But I had escaped.

Grudin:

I was back at Wang. In my eight-year absence, Wang had become a large computer company. Dr. Wang was one of the wealthiest Americans. Former programmer friends and tennis partners were now vice-presidents. My title this time was software engineer, not programmer. A month after my return in 1983, the first CHI conference was held in Boston. Computer Human Interaction.

Grudin:

About my “Suggestion” or “Active Help” system, to be based on analyzing saved typing patterns, they said, "Could you please first improve our normal Help system?" That was as far as I got on Help. But maybe it was just as well, because Alan Cypher, one of the UCSD post-docs, spent years on active help, and almost 40 years later, it remains an unrealized goal. I attended a workshop after I arrived, with the UCSD group. You were there. To discuss chapter drafts of a book, User Centered System Design, which I believe is still in print 35 years later—because it addressed unchanging psychology, as much as fleeting technology.

Grudin:

Then in 1984, the Mac arrived. I went down to the Harvard Book Store to see it. A kid standing next to me knew more about it than the salesman demoing it. All my software engineering colleagues hoped that this competitor would succeed, and it almost failed. Hard to imagine now, but it didn't have enough memory or processing power to run anything significant. It had a tiny display and it was very expensive. Apple almost went bankrupt. But in 1985, a new version had twice the memory and some companies published software for it. And in 1986, the Mac II had twice the processing speed, so graphical user interfaces were off and running, and there would be no more studies of command names. My friends studying command names were depressed at this realization, whereas my view was, "Let's quickly publish all of our studies before anyone notices that there won't be any more."

Grudin:

In my earliest stint at Wang, I learned some basic beginner lessons, such as “It's better to appear a little overconfident in industry than under-confident.” “Rules are not always expected to be followed, it might be just that if you don't follow them and carelessly cause problems, you're responsible.” This time I saw higher-level organizational dynamics. While I was there, Wang made two fateful decisions: a Wang PC that was not IBM-compatible, and their first GUI project, a workstation that I worked on, didn't adopt Unix. Wang had done well with proprietary operating systems in the past, but these now couldn't support popular applications that were proliferating. Although Wang would soon be on a downward spiral along with the other minicomputer companies, there was no end of things to learn there. The GUI project was a wonderful opportunity to explore the new interaction paradigm. These three years provided two great topics that shaped my research career, and also an approach that had mixed results.

Grudin:

One was that the waterfall model software development process—design it up front, don't test it until finished—was surprisingly inadequate for interactive software. At CHI 83, you and John Gould gave one of my all-time favorite papers, describing how a user-centered iterative design process should work. Nothing like that was practiced by my colleagues before or after CHI 83.

Grudin:

I thought that the command abbreviation papers might be my last publications, but what I was learning seemed useful for the CHI crowd, so I submitted a paper on design challenges to CHI’86. A lot of the CHI papers were still experimental. This wasn't formal experimentation. It wasn't basic research in psychology. Most of the early CHI participants were psychologists, but I considered this basic research in HCI.

Grudin:

Second [topic]. By chance, I worked on half a dozen applications or features that were designed to support group activities, such as shared calendars or voice applications or project management. Group support was new. The spreadsheet and word processor were individual productivity killer apps. But there was little beyond email to support the millions of groups in the world. We tried, on these projects. We overcame the technical problems, but none of them succeeded in the marketplace. And trying to understand why that was, before starting on the seventh, seemed important. [Laughter]

Grudin:

The third discovery in this period, in which I probably invested too much time over the years to come, was interdisciplinary possibilities. The Human Factors Society co-sponsored the first two CHI conferences and then left. One impediment to collaboration there, which you colorfully described at IBM, was that human factors engineers were trained in behavioral psychology and dismissed cognitive psychology, whereas most founders of CHI were cognitive psychologists who dismissed human factors. But human factors did rigorous experiments. I believed everyone knew things that could help. I attended human factors meetings and called myself a human factors engineer, solely so cognitive colleagues might be more accepting.

Grudin:

This was the first of about a dozen efforts over the years to bring fields together, and none really succeeded. My closest colleague at Wang was Susan or Carrie Ehrlich. She was a cognitive psychologist who also had a background in management. She introduced me to the Management Information Systems literature. I approached Wang's competitive analysis group, but they focused on price and functionality, not on usability. Then, as we began working on the GUI, I convinced a designer who worked downstairs on hardware boxes that software design might be interesting. It took almost a year to get him accepted by other software engineers.

Grudin:

I encountered two other disciplines in this time. I always enjoyed reading anthropology of exotic places, but HCI, software engineering, didn't seem exotic. Then I read a short article in Transactions on Office Information Systems by Lucy Suchman, an anthropologist at Xerox PARC. She described a rational, methodical three-step purchasing process at Xerox, followed by a transcription of a chaotic session in Finance, where it broke down completely and required problem-solving and exception handling. She concluded, "This is what work is really like," or words to that effect. I was dismayed. She took a simple, elegant process and generalized from one dysfunctional case. “This isn't science.”

Grudin:

I went upstairs and found people in our purchasing department and asked them to describe the process. One described the rational process. I said, "Right. That's how it works." She hesitated. Then she said, "That's how it's supposed to work, but something always goes wrong."

[Laughter]

I handed her the paper. I said, "Would you read this paper, and tomorrow tell me what you think of it?" The next day she said, "Exactly. If anything, it's worse than that. Some exceptions are so common that we have exceptions to the exception." Anthropology is a science with different rules. You describe representative examples, not aberrations. I later got some ethnographic training from Ed Hutchins, who had been one of the Sloan postdocs at UCSD. Later, reading papers by ethnographers describing workplaces like mine, usually I think, "Yeah, they pretty much got it right." But sometimes they reveal important aspects of my environment that I hadn't noticed.

Grudin:

Finally, artificial intelligence. AI has had summers and winters. A 197O Life Magazine article quoted professors at MIT and elsewhere who said that by 1980 or 1985, computers would be more intelligent than people. They were wrong and their funding went away, for a while. But interest was picking up again. AI was included in Cognitive Science.

Grudin:

Then, in 1984, Japan launched the Fifth Generation Project. The US and Europe responded. The US effort had military spending from DARPA and a civilian consortium, MCC, founded by major corporations and headed by Bobby Ray Inman. If this time AI was going to take off, I didn't want to miss it. In 1986, I moved to Austin to lead a group in MCC's Human Interface Lab. I soon saw that the Singularity wasn't imminent, but MCC was a great place for research. We were encouraged to publish, because the companies in the consortium wanted peer reviewing to show that our work was solid before investing to apply it. Months after I arrived, the first Computer-Supported Cooperative Work conference was held in the MCC auditorium. I hadn't heard about it.

Grudin:

CSCW and CHI would be my community for the next 30 years. CSCW was initially a collaboration of CHI and Management Information Systems researchers, but MIS soon withdrew. Twenty years later, we tried a CHI and MIS collaboration. It also fell apart, as did efforts to bring CHI and Design research communities together in the 1990s and 2000s. The CHI and CSCW communities are centered on their large annual conferences. Most HCI research in the United States is reported in proceedings of these highly selective conferences, not journal articles. This was one of several barriers to collaborating with the other fields. They valued journal publication and mostly used conferences for work in progress. Through the 1980s and early 1990s, most CHI and CSCW participants were well-informed industry researchers.

Grudin:

A conference workshop could lead to a book, back when books were relied on more. This changed by the 2000s. HCI had earned academic respect and conferences shifted to become rigorous training grounds for PhD students. Much more of an academic focus. The new field was exciting. Conferences were completely run by volunteers. CHI 89 was in Austin, when I was there. I was sponsorship chair and computer support chair, which meant borrowing computers and transporting them in my car to and from the conference. Apple built and brought an extremely ambitious photo booth and information kiosk system to support interaction among attendees, and included information about the program and the city of Austin. They successfully deployed it live. It was working on a high wire with no net. It was very impressive.

Grudin:

Another influential group was the HCI Consortium: prominent industry and academic research labs that met annually for a week, during which faculty and senior researchers presented, followed by prepared responses and extended general discussion.

Grudin:

MCC provided an opportunity to study interface development practices in our shareholder companies. I could determine which of my observations at Wang generalized. I also wrote up material that I'd collected at Wang. For example, the HCI field promoted design consistency. Over and over, I had run into cases where it wasn't so simple. It was task-dependent, like the consistent vowel deletion command abbreviations that didn't work. I collected examples in a paper titled “The Case Against User Interface Consistency.” Again, senior people asked me not to publish it, lest it confuse people into believing that unmotivated inconsistency is okay, which it isn't. The only paper I wrote and never submitted was also at MCC, an account of why natural language understanding systems were funded through the 1980s for organizational reasons, despite no prospects of success. I dropped that because it would upset too many people, including friends in MCC’s two natural language groups.

Grudin:

My subsequent career was determined by a presentation at CSCW 88 in Portland, Oregon. Examples from my Wang experience illustrated three key challenges in building successful group support applications. I'd written the first draft of that paper after leaving Wang, before starting at MCC. When presented, it was praised by several subsequent speakers and was widely cited. Also attending CSCW 88 were several Scandinavians who described Participatory Design: engaging users throughout the development process. This was certainly not done by MCC shareholders. Wang marketing and salespeople did not want developers to talk to customers at all. Our shaggy hair, or our tales of wondrous future products, might discourage them from buying what was available.

Grudin:

Participatory Design seemed logical. The Nordic researchers liked my paper and invited me to take a leave of absence to spend a year at Aarhus University. As I prepared to teach an HCI course on developing interactive software products, I asked students where they expected to work. Hospitals, government agencies, financial institutions. There were no large computer or software companies in Denmark. They would write software for use in just one company. Participatory Design suddenly made perfect sense. They knew exactly who would use it. It was nevertheless hard for them to convince companies, because companies had bought into the waterfall model: Design it first and evaluate at the end. Thanks to you and John Gould, I now saw evaluation as an integral part of iterative design, but my Danish Participatory Design colleagues saw it as a bad step, the final step in the waterfall model, always taken too late to change much.

Grudin:

This meant redesigning my course. Aarhus gave me a light teaching load and time to prepare. This generosity contributed to my two years there being my most productive period. One of my favorite papers, published in IEEE Computer, contrasted the nature of developing interactive software for commercial products, developing for internal use by an organization, and developing through a contract process. As I researched this in the computer science library near my office, I dusted off articles that had probably been unread for years. It was an interesting feeling.

Grudin:

It turned out that the waterfall model was designed for contract development of non-interactive military systems, but it was then used everywhere as the first available fully-specified development process. Could Participatory Design be transferred to commercial product development? My Danish colleagues socialized with their future users, played football (soccer) with them. They worked with them. They understood them. Commercial products hoped to have millions of future users.

Grudin:

Many years later, I saw a solution: Designing for personas, fictional people, but based closely on data from categories of users. Fictional characters can become real to us. Personas were especially good when a product development cycle lasted a few years so the developers could become familiar with these personas and be updated with new data about their behaviors. In teaching, I knew I didn't understand the perspectives of Danish students, so I designed courses around their reactions to readings, which proved effective, and I continued using that approach. My Aarhus visit extended to two years when I, and most of my HCI colleagues, were laid off by MCC, while I was in Denmark. MCC was forced to downsize by the rapidly evolving tech world. Many of its shareholders were aging mainframe and minicomputer companies, such as Burroughs and Digital Equipment Corporation. They were struggling in the PC era, and couldn't afford the annual fee.

Grudin:

Also, the software MCC produced was written in LISP, an AI programming language that was not used by the shareholders. AI was under-delivering. The Reagan-backed DARPA military initiatives, nicknamed Star Wars, all failed. An icy AI winter set in. The 1990s were terrible for AI, but the best decade for HCI, which I don't think is a coincidence. AI and HCI competed for government funding, for faculty and industry positions, and for good students. AI researchers at MCC heard me describe what I hoped to do and said, "Why waste time on these user interface issues? In several years, ultra-intelligent computers will be able to do all of that for us before breakfast." Or words to that effect. The only way I could continue to support a group that wasn't working on AI was by putting some of my time into the major AI CYC project at MCC. But with AI in disgrace, computer science departments adopted HCI in their curricula, and sometimes CSCW. They hired people away from industry, including me. Good students followed and flooded into our conferences.

I think we're at the end of the hour here.

Lewis:

Okay. If you have another minute--

Grudin:

Yeah, definitely.

Lewis:

The first place I had a question, you're finishing up at the MRC and you went to Wang Labs. What was your decision process there? It's clear from the background, you knew people at Wang and all that, but there must have been some process you went through that ended up with your deciding to go back there as opposed to doing anything else?

Grudin:

Yeah. Well, the two factors which I touched on. Seriously, I was really worried about my health, you know?

Lewis:

Yeah.

Grudin:

The mathematician Ramanujan had moved from India to Cambridge and died shortly thereafter.

Lewis:

Yeah.

Grudin:

Not that I was putting myself in that category. I had a four-year postdoc, I could have stayed at least that long. The doctor said, "Oh, in the old days, we would have sent you down to the south coast to take in the sea breezes." Well, it wasn't the old days, nobody was suggesting I go down to the south coast.

Grudin:

Another serious consideration was that I really was keen on this idea of building a Suggestion system or an Active Help system, because I could see lots of cases where you could take a pattern in somebody's typing and detect ways that they could be told about an improvement. I was pretty interested in that. I mentioned that in my application, when I was talking to my friends, in the informal part of the application. Why has that not turned out to be successful? I think part of the reason is because we keep adding features and capabilities to the software... If we took the software that existed in 1983, I think we could do a really good job of that, but we keep adding features, we keep pushing the boundaries.

Grudin:

If anything, people are more interested in the newer features that they haven't already learned and need more help with those. Alan Cypher did a lot of work on that, and it really never was picked up. At Microsoft, I've seen people try that. I've tried some, for Excel or other applications, but they don't really take hold.

Lewis:

Okay. So the move to Wang, as opposed to somewhere else, was driven in part, by the notion that you had this thing you wanted to do, and you correctly detected that they would be interested in that?

Grudin:

And that it would be easier to do in industry. You would have the computing capability and you could also have an impact if you did it.

Lewis:

Okay. Yeah. Then the next transition, you went to MCC and you mentioned that you went there as head of a group. Is there more you might say about how you thought about that and how you put the group together and things like that?

Grudin:

Yeah, that was a bumpy process. When they interviewed me... They had this sort of general concept. What they wanted was an ‘intelligent user interface management system.’ MCC was an interesting experiment. While I was there, a group of us formed an Organizational Science Discussion Group where we read papers in organizational science. One paper stood out to all of us. All of us agreed that this was the most interesting paper. It was by Henry Mintzberg. It described a typology of organizations.

Grudin:

There are five different parts of an organization. The strategic apex or senior management, middle management, the individual contributors, and on the side, you had a support group and you had the technostructure. The technostructure had no technology necessarily involved. It was the people who defined all the work processes in the company. The support staff was everybody else, including IT and kitchen help and librarians and so forth. His argument was that, in different organizations, these always vie for influence. And in different organizations, different of these become dominant.

Grudin:

In a startup it’s strategic apex, the founders are in complete control. For a divisionalized company, middle management handles different divisions and has a lot of influence. Manufacturing, where all the processes are incredibly important and need to be defined, in that division or in that company, the technostructure could be the dominant group. The support staff is the dominant group in what he calls ad hoc groups, like a film company where different experts come together. Or maybe a chemical, or pharmaceutical company that doesn't actually produce anything, but creates the designs as others create the components. The individual contributors tend to dominate in what he called professional bureaucracies, such as universities, where the faculty have a lot of power. So that was his thesis. We all liked it. I don't think we completely realized why we liked it. Afterwards, I realized, how would you test the hypothesis that you have these five competing groups, they vie for power, and then one becomes dominant? The way you might test it is to create an organization, like about 500 people, out of nothing, with only the vaguest of charters, and put them down and let them loose and see what happens. And that was MCC.

Grudin:

That is exactly what happened. Each of those groups… the chief scientific officer thought that he should be able to control everything. The individual contributors felt that we should be in charge of things. The support staff often came from the shareholder companies and they looked at what we were doing and said, "I know better than these people, what will be useful back at Honeywell or NCR." So they thought they should have a lot more influence. The people who dealt with the work processes, we had these tricky legal restrictions because MCC had several programs and different shareholders contributed to different ones. What could you share between the programs that would be available to groups that didn't subscribe to that program? They had a lot of these sorts of work processes, and thought that they should basically be running the show, whereas we tended to want to collaborate with people independent of who happened to be funding one group or another. Then the middle managers, all of the people who were project leads, wanted to build their little empire and not be constrained from above. In the Human Interface Lab, they said, "We have several different projects going on, but let's put them all together and create this unified user interface management system that will draw from each of them. And let’s let Jonathan handle that." Well, none of the other project leaders were really enthused about this idea at all.

Lewis:

Yeah.

Grudin:

The natural language understanding system, as is its wont, always needed more resources. It's like it was a black hole. It always needed more because it wasn't approaching the goal. The user interface management system went away. I came in, and I did end up with a small preexisting group that worked on the things that I was interested in. We did studies, as I mentioned, of how interactive systems were developed in our shareholder companies. We sent large surveys around to all them and they were happy to participate. We went out and spent months as participant observers on development projects in different shareholder companies.

Lewis:

A comment on that. I think about MCC, because you know I never worked there, but I visited it a couple of times and I often think about them. And most frequently, it's because of one paper that I realize fit the pattern you just described, although it was not done in your group, it was done in the software engineering group. And that's Curtis, Krasner and Iscoe. Do you know that paper?

Grudin:

Yeah.

Lewis:

I'm constantly telling people about that, in part, because, I can't remember whether this was their top bullet or not, but they used this wonderful phrase, “the thin spread of application domain knowledge.” That is, to this day, a huge limit in so many software projects. People just don't know what they're doing really. I mean, they think they do, but they don't because they don't know the context. And as software people, they frequently don't know that they don't know. Or maybe worse, they assume that what they don't know couldn't be important.

Grudin:

I can tell you one other—

Lewis:

Okay. Yeah, go ahead.

Grudin:

... another thing that might be of interest here. But first of all, I agree, I was very impressed with the paper and Bill Curtis. I hoped to collaborate with Bill Curtis, but he'd had a bumpy time at MCC, just about when I arrived. He was original head of the Human Interface Lab.

Lewis:

Oh really?

Grudin:

Then for some reason, which I didn't completely get, he was moved out of that position. There were personalities. Possibly it was a Honeywell vice-president, who was in sales. He hadn't really done HCI. He told me this, that he had one more career and he thought that leading this Human Interface Lab was what that career should be. Ray Allard. He came down, he took the position of the head of the Human Interface Lab while I was there. Bill moved to head up a software engineering group within a larger program. When MCC began, it was located in a couple of buildings, while a building was built by the University of Texas for MCC.

Lewis:

Yeah.

Grudin:

When that building was built, we all moved into the new building, except the Software Engineering Program. It was a space issue. Later, we had a little bit of downsizing and it became conceivable for Software Engineering to move into our building. They didn't. They went through a thought process, "What would the advantages be? Disadvantages? Maybe some loss of lab space?" I later found out from Bill what the decision really hinged on. In our building, all the MCC executives were on the fourth floor, I was on the third floor and so forth. MCC was under some pressure, and one way that that manifested itself is that the shareholders were asking, "We have to subscribe to an entire large program," of which there were just four. "We would like to be able to fund specific projects within a program."

Lewis:

Mm-hmm (affirmative).

Grudin:

MCC didn't want to do this, because it would mean less money and more complication. But they convinced our senior management that we should be moving in that direction. Software realized that if they moved into the same building, the senior management could hammer on them very easily. They would be forced to do that. Whereas, if they were off miles away in another building, they could probably resist it. And it worked perfectly. They were the one program that did not get subdivided that way.

Professor of Information and Computer Science, UC Irvine[edit | edit source]

Lewis:

This is Clayton Lewis and it's July 16th, 2021.

Grudin:

Okay. When I was preparing to leave Aarhus, my uncertainty about my future kicked in. I applied for three positions. One in industry: Bell Labs; one in government: National Science Foundation, and a university position at UC Irvine. On the one hand, my recent work—innovative industry projects at Wang, and MCC, a crossroads for a chunk of the field, and then the European perspective—gave me a lot of confidence. On the other hand, my prospects seemed highly uncertain because we were struggling for respect from engineers in industry, and from academic computer scientists. I'd never had a job that lasted even three years. Two recent jobs were in prominent organizations that when I left were sliding into bankruptcy. I didn't guess that the nineties would be a period of stability, both in the field and for myself. I was very uncertain, and my personal life was also unsettled, probably for the same reason. My father had reached his mandatory retirement age. My mother saw to it immediately that they left the Ohio village that she'd never liked, and settled in a Quaker-backed retirement community, a new community in State College, Pennsylvania whose founders included calculus textbook author, George Thomas--

Lewis:

I can't resist. He was my advisor. But anyway, go ahead.

Grudin:

Yeah. That's where he lived out his life into his late nineties, I think, in State College. There were some other engineers and mathematicians that he'd recruited in. They didn't overwhelm the community, but they were there for my father to talk with. My mother had accepted living for 30 years in Ohio because it was perfect for my father, this small, friendly mathematics immersion. She was very active. She was a prominent Ohio political organizer, known to the Democratic governors of the state. She loved teaching kindergarten, until near the end, when a focus on reading was forced into kindergarten. She felt that some kids are ready to read in kindergarten and some basically would only like reading if they were given another year to mature, and not forced through a failure experience in kindergarten. The years in Massachusetts and Oregon showed how much she could accomplish elsewhere.

Grudin:

I saw all that, and I knew that I'd never constrain somebody I loved the way that my father had, without it necessarily having surfaced as an issue. I would never constrain someone that way. And my insecurity about my own future prevented me from even imagining a stable future. So my relationships also tended to last less than three years. My younger brother was an influence on me. He was less academic. He was smart and serious, and he had social skills that I think I envied, and I pushed him away in our early teens. But later I admired him and we became close again. He periodically updated my music consumption—I mentioned dragging me off to Woodstock—and later introducing me to jazz. In the other direction, I was responsible for him taking chess seriously, as I mentioned. After working primarily as a cook and chef, he earned degrees in education and taught school in New York city, working as a tennis pro in the summers, until his second marriage to a tennis playing Japanese businesswoman led to quasi retirement outside Tokyo, where he's been for a decade.

Grudin:

My mother remained the strong influence in my life. We had the same love of typed correspondence, same sensitivity to injustices, same sense of humor. She and I were two people who would be just sitting thinking about something and burst out laughing when there wasn't anybody else around. We were both voracious readers, we liked politics, and we also liked friendly gossip about friends and relatives. We collaborated at one point collecting and reading a couple hundred books on Watergate and Nixon, which became a library exhibit. We also exchanged book recommendations and read books that the other recommended. I have a bookcase behind me that reveals some influences that I haven't mentioned from books. Early on, I read biographies of physicists, and I read GH Hardy's A mathematician's apology. Have you read that?

Lewis:

Yeah.

Grudin:

Yeah. Well, Hardy very elegantly reflected my father's view of mathematics, both theoretical mathematics and applied mathematics. My father had taught every course at Denison in the math department. He did out of principle teach the computer science courses. He learned the computer science when he taught it, but he never thought it was real math. Hardy's book was also a source of some anxiety for me, because he insisted that your only creative years are your very young years.

Lewis:

Yeah.

Grudin:

That was something I reflected on... I think that pushed me. Quickly, up or out as far as working in mathematics is concerned, if it didn't take really fast. I was in graduate school for one year, and exited with a master's when I decided that maybe my distraction while focusing on Watergate and so forth, was robbing me of these potentially creative mathematical years. I read a lot of science fiction and a lot of stuff in the very small Granville library when I was in high school. But my serious reading began in college.

Grudin:

The strongest influence on me by far was Arthur Koestler. I first encountered him with Darkness at Noon, his political novel, which explored psychology and politics, personal psychology. Even though I wasn't at that point interested in Freud or Skinner, I was interested in politics and psychology, talking with my parents about how people worked. Koestler's longer book, The Act of Creation, changed my outlook and my approach. My feeling of the legitimacy of my approach to science, or to studying. It was unlike anything that I'd read. There are now books more like that, but if you haven't read it, it had a wealth of—

Lewis:

I haven't.

Grudin:

Okay. Well, it's a big book and it's not a new book, but it had a huge wealth of compelling insights into creativity, which he came at from three perspectives: humor, art, and science. It wasn't based on experimentation. He was at the Stanford Center for Advanced Study in the Behavioral Sciences for a year, and did an experiment. He wrote about it, he was quite pleased to have actually done an experiment. The Act of Creation argued for a view of the world that was orthogonal to my math, physics, chemistry, biology perspective of science. I wrestled with it. It was science, but it wasn't how I had thought about science. I wrote to Koestler and said, "Okay, what should I read next?"

Grudin:

He wrote back and suggested works in General Systems Theory, which does have this lateral orientation, but none of those books communicated it nearly as well to me as Koestler did indirectly through the examples in Act of Creation. His method in writing this and his other books, which I have a collection of, was a wealth of observations, looking for patterns, finding logical causal explanations for the patterns, then working relentlessly to find alternative causal explanations for those patterns, and for looking for and finding disconfirming cases, because confirmation bias is the enemy when you get into non-experimental and even in experimental science.

Grudin:

This may sound straightforward, but what I've seen, and I'm sure you've seen, is that researchers are about as prone as anyone to settle too quickly on a causal explanation for correlational data. Many strong researchers, once they're publicly associated with a phenomenon, resist acknowledging where it doesn't apply. Once at Stanford, I walked by a psychologist standing in the hall, unaware that anybody was nearby, holding a printout of experimental data against the wall and slamming his hand into the wall, shouting, "I. Hate. People!"

[Laughter]

My breakthrough came when I reached the point when finding evidence that contradicted my treasured position was genuinely a source of joy, that it was like the first step to reaching a better or more comprehensive position. An example of this arose when I visited Microsoft and Sun Microsystems, and found employees were using automatic meeting scheduling features in online calendars. This was an example that I'd used for why group support applications failed in the marketplace, in my well-received paper. I had written that managers saw the benefit of this automatic meeting scheduler, managers and their admins, or secretaries back then. But individual contributors didn't want to put their relatively few meetings into an online calendar. They saw no benefit from it.

Grudin:

Some people when I pointed this out predicted that managers would get around to mandating calendar maintenance, in order to get these benefits. But I didn't see that happening, and I wasn't sure that it would happen, because of the lower cost of the computing and the fact that this didn't seem like something that they would necessarily mandate. So now it was a little embarrassing that my paper was wrong on that point, and that would be noticed. But the question is, what had happened? Were managers at Sun and Microsoft mandating that everyone maintain calendars? My student Leysia Palen and I looked into it. A new generation of calendars had a feature that individual contributors loved. They didn't have many meetings. Their problem was when they didn't notice that it was time for a meeting and they missed it. Calendars that would beep or signal in some way a meeting time, were an incentive to put in their meetings. Then that enabled scheduling.

Grudin:

Okay. So on the bookshelf behind me are other authors whose influence was less direct. Bertrand Russell. My mother suggested some of his writing. Norman Mailer, Aldous Huxley, Thomas McGuane. One Hundred Years of Solitude was a favorite that pulled me out of a completely realistic... to magical realism. For understanding the human side of human computer interaction, I thought some fiction writers provide great insights. I've known people who would not read anything written on psychology... would not read Koestler, because he didn't have a PhD in psychology. And similarly, of course, would not look at what fiction writers might say.

Grudin:

My second strongest influence overall was Thomas Pynchon, who trained as a chemist at Cornell and was briefly a technical writer at Boeing before focusing on fiction. From his first novel on, he has focused on the impacts of technology on society, a major focus, one of the currents through his work.

Grudin:

Now I’ll move to when I arrived at UC Irvine. Rob Kling contacted me when I was at Aarhus and said, “As soon as you're done in Aarhus, why don't you come to Irvine?” I considered the National Science Foundation Program Manager and Bell Labs positions, but joined the UCI group in late 91, officially starting January 92. However, I said that if they held the position open for me for two years, I'd first do a two year rotation at NSF. And they said, no.

Grudin:

The group was called CORPS, C-O-R-P-S for Computers, Organizations, Policy, and Society. A month or two after I got there, Rob burst into my office with a worried look, and said, "Is it too late for you to do the NSF rotation? Larry Rosenberg just rejected my grant proposal."

[Laughter]

But it was too late. The CORPS group focused on social theory and system use and organizations. The next year, 1993, we hired Mark Ackerman, another CSCW researcher. It was a well-respected group. Rob Kling was an area editor of Communications of the ACM, which was more of a journal at that time. Rob had coined the term social informatics. He bequeathed me a course, social analysis of computing, which was a requirement for information and computer science majors, covering the impacts and responsibilities and technology risks. Students really appreciated it. John King had a joint position in Irvine's Graduate School of Management, with which I got involved and saw more of the management information systems side of things. He later became Dean at the University of Michigan School of information.

Grudin:

Rob and John had no interest in CHI or individual interaction with computers, but CSCW had an organizational element, or it did at first. The MIS contingent left at just about this time, which left CSCW to CHI researchers whose focus was on technology to support smaller groups, primarily, such as collaborative writing and video conferencing. These were of interest to computer companies and software companies, and the regional telecommunications companies that are now gone, but at that time there were major research centers after the AT&T monopoly was broken. To balance this, in 1989, ECSCW or European CSCW formed, focused on organizational support. They were not being driven by computer companies and software companies. I served on both program committees and submitted papers to both CSCW and ECSCW.

Grudin:

Rob Kling captured the different viewpoints in a keynote address at CSCW, asking, “Why do we use the term cooperative, when much work is competitive, conflictual or even coercive?” This came from the organizational perspective. My CHI colleagues, who created CSCW, focused on products for small groups, and if small groups that cooperated used a product, that was good enough. You didn’t need to build a collaborative writing tool for co-authors who didn’t get along. The Europeans also saw cooperation as an ideal, or as cooperation amongst workers who might at times be in conflict with management.

Grudin:

I taught Irvine's first HCI courses. They weren't official CORPS courses, because HCI wasn't social enough. But HCI was becoming respected within computer science, the department wanted it taught. I also taught CSCW, I believe the first course at UCI with that name. On a sabbatical, I co-taught the first HCI course at the University of Washington. Acceptance by computer science led to an expansion of HCI that if anything is accelerating. In the eighties, we struggled for acceptance by engineers and computer scientists. We declared ourselves engineers—usability engineers. Formal experimentation remained a core skill, even as we added other approaches. Our nervous focus on being seen as objective meant that we avoided subjective social, political aspects of computing. We were engineers. How technology was used was for others to decide.

Grudin:

So I thought there was resistance to us taking positions, as a group, on political or social issues, until HCI programs and departments began forming in the nineties. They had more autonomy, people could do more of what they wanted to do. Things like digital divides were more visible and could be discussed. Today, nothing is out of bounds, which I think is a good thing. Our impact can be greater, and we really need to be anticipating consequences of technology. Looking back to see what wasn't discussed can be as valuable in some respects as seeing what was discussed, it reveals pressures that shape research. It also explains why researchers didn't anticipate, or prepare better for, some of the consequences that we’re now facing.

Grudin:

Anyway, while at UCI I combined sabbaticals and unpaid leaves to spend six months teaching at Keio University, at the University of Oslo, and at the University of Washington. I also participated with you in that ACM sponsored lecture series in former Soviet-bloc countries. Japan was very interesting. I have a few things to say about some of those visits. Travel always fascinated me. I think I'm a pretty good traveler. I'm not very imaginative. I'm not naturally theoretical or top-down. So I look around for things that surprise me, or for patterns in what I'm seeing. In 1981, I had more or less parachuted into England where there was no internet or fax, and long distance was very expensive. I'd saved a little time and money at the end of that postdoc in 1983 to visit Greece, which I'd wanted to see since reading Thucydides at Reed.

Grudin:

I haven't yet been there. Not long before I was headed there, I was in a pub with two colleagues and one of their former boyfriends, Paul. He left England and was teaching at the University of Jos in Nigeria. He'd just ridden his motorcycle solo across the Sahara desert to reach Europe. He invited the three of us to visit Nigeria, which was not open to tourism, but he could get us a letter of invitation. He clearly hoped that his ex-girlfriend would come. All three of us said we would. I was the only one who actually did.

Grudin:

Paul never got my postal mail indicating when I would arrive, and there weren't any working telephones. I arrived, and when he wasn't at the airport as expected, I gave some thought to getting back on the plane, heading back to England. The airport I'd arrived at was quite a distance from Jos, and I had not done any research on how to get around. But I thought it'd be too embarrassing to do that. As a result, I spent a couple fascinating days where I didn't interact with any white people. After that, I absolutely loved the ex-pat community in Jos, and might've stayed in Africa if I thought that I had skills that would be useful there. At the end of my visit, I gave a carefully prepared lecture at the university, and was stunned by the forceful arguments and intelligent questions from the students. This was echoed in a way in 1989, when I was heading to Aarhus. I planned to come three months early to study Danish. When I asked them to find me a class to take, they wrote, “No! Please! Your Danish will never be as good as our English, and we don't want to put up with it!”

[Laughter]

Grudin:

So I tossed out the language tapes that I was listening to, and contacted a high school classmate who was living in Nairobi and the ethnomusicologist sister of a colleague of mine, who on a visit to Austin invited people to visit her in Burundi. She was out of Burundi because there was warfare between the Tutsi and the Hutu. But now things were peaceful. So I spent a few months between the two places. An informal focus was learning how technologies like radio, television, computers if there were any, or even money, had come into the places I visited. Being a tennis player got me invited to play the US ambassador in Bujumbura, and to a party where I talked with an earnest graduate of the national university, relatively meritocratic in the Swiss model Burundi followed.

Grudin:

He'd been a computer science major. He was very sharp, he was very personable, and he wasn't looking for anything from me other than information about the computer field, which was not always my experience when traveling. Many people were interested in seeing if they could find a way to get to the United States. He'd just taken a government position. Eventually in the conversation, I said, "This sounds like a great job, but you don't sound very happy." He paused. Then he said, "I was a computer science major. I hoped I would get a job that would let me use a computer for the first time."

Grudin:

The first thought that crossed my mind was, “If people this smart and motivated could compete with me, I wouldn't stand a chance.”

[Laughter]

I could see that globalization was going to be complicated. Back to the sabbatical leaves. I didn't become fluent in Japanese in six months at Keio University, but I made a strong effort to immerse in the history and the culture, reading at night, taking notes, exploring. Many mornings I came in with questions for my colleagues, until finally one said, "I focused on engineering starting when I was 15. You know more about our history than I do. Let's talk about engineering." I had great experiences with colleagues in Japan, and students, and going around Japan, but I never felt I could completely understand it. I was convinced that it was because Japan had been largely cut off from the west, was never colonized, and it's fundamentally different from other places that I've been. I should discuss this with my brother. He's visiting later this summer, he's been living there for a decade. He's highly integrated into a community in Japan, which I visited.

Grudin:

I'll give a relevant example, which may result from Japan having been an oral culture more recently than European cultures have been. UCI had a heavy Asian-American contingent, many of whom were first-generation Americans. One of the most unpleasant professional duties in the Information and Computer Science department was dealing with chronic plagiarism. At the beginning of classes where students would have an essay or a report to do at the end, I would hand out materials on plagiarism and talk about plagiarism. Toward the end, I would sometimes ask, "Does everybody understand that?" People would nod. I would say, "Now who in the class is planning to plagiarize something? Because it will make it a whole lot faster for all of us if you just raised your hand now."

Grudin:

That was probably not a very sensible thing to say, but it describes the situation. Some of my colleagues, if somebody did it, you could get them thrown out of the university, you could get them thrown out of the class. I would just fail them on the particular assignment that they had plagiarized. Okay. At Keio, I quickly discovered that uncited quotation, which we call plagiarism, was perfectly acceptable.

Lewis:

I have a story about that for you sometime.

Grudin:

Okay, well, I'll finish my story, then I'd be happy to hear yours.

Lewis:

Okay.

Grudin:

I knew that students would plagiarize from the class readings that I gave in their essays. I didn't try to stop them. I didn't bring up the subject. They even plagiarized me. I asked one of them, “Why not cite me? Put in quotation marks, is that hard?” He looked puzzled. He explained: “If I cited you in my paper, I would be implying that you didn't realize that you had said that.” At the end of my stay there--this was one of the leading engineering labs in Japan--they had an offsite. It was a big deal for the lab. The lab consisted of 10 PhD students, 20 masters students, and 10 fourth-year undergraduates, most of whom would probably become a master's student the following year. They met for two and a half days at a resort on the coast. It was a really brilliant mix of work and relaxation for the people in the lab. There was some discussion amongst them as to whether I should be invited, because I really hadn't done the hard work that merited lab members doing this, but they did invite me.

Grudin:

One of the exercises was that the fourth year undergraduates had to present a major project. A graduate student had given them a paper to read, and they presented that paper to the faculty and the graduate students. The other undergraduates could relax then. They presented it in the first person singular, as their work, explaining exactly what they had done, how they did it. The effect of that was that if somebody asked a question, like “Why did you do this analysis?” Or “Why did you do it that way?” They couldn't say, “I don't know.” It was like method acting, that they had to completely immerse in that role and be a researcher. Okay. When I returned to UCI, I saw things differently. I regretted, especially, one past action against a student whose parents may well have guided him in how you write reports.

Grudin:

Another cultural discovery was when I nicely told a pleasant but annoying student that his incessant fighting over his grades had built a faculty view that might hurt him when it came to writing job references. He literally staggered toward the floor in terror, when I said that. He explained that in his Middle East culture, haggling even for a minor gain was standard procedure. He said he haggled more in my course than some as a show of respect, even a minor upward adjustment in that course would mean a lot. After graduating, he occasionally wrote to me to tell me about his career progress.

Grudin:

When I think about these things, I wonder if today, with more globalization and more interaction, there still are the same level of cultural differences. I'm not sure. An example from my time in Norway touches on, and was maybe my first clearest embrace, of what I think is the single greatest challenge and risk for most of us in technology innovation, which is unintended consequences. Norway has a population in the far north, apparently put there as a first line of defense when the Soviet Union rolled across the north to invade Europe. Norway had to support those folks. So they were pioneering in telemedicine as one way to support them. When I first saw it, I thought telemedicine was the single most wonderful technology I'd ever seen.

Grudin:

In Burundi, when I was serving as photographer for the ethnomusicologist, she was recording a unique form of whispering music. It was only played by old men in villages at this point. I had a discussion in one of those villages with a young doctor who was near the end of a two-year post there. It was required for graduates of the university medical school that they spend two years where they're assigned, after which they can go where they like. Most of them returned to the capital, to Bujumbura. I asked him what he was planning to do, and he was torn. In addition to medical care, as he walked around the area, he saw sanitation, he saw poor diet, and other things that could be fixed. He worked with kids and encouraged the most talented to consider going to university, which no one else would have done. He was a tremendously respected member of that community, of course, but he missed the cultural and social life in Bujumbura.

Grudin:

In thinking about telemedicine in Norway, probably it would lead to fewer visits by the doctors who had to go up and travel around the northern communities. I was only there for six months and I don't think my suggestion to look at indirect effects of telemedicine in the north was acted on, as far as I could determine.

Lewis:

That's really interesting.

Grudin:

To sum up my time at UCI, I entered as an assistant professor, and within a few years, was promoted to full professor. I did the usual things, although maybe at less volume than professors now are expected to do, as such as obtaining funding from NSF and industry. I became the second editor-in-chief of ACM Transactions on Human Computer Interaction and associate editor for human computer interaction for Computing Surveys, ACM’s most widely subscribed journal. I converted TOCHI’s completely paper-based submission and review process to digital. It was the era where that shift was happening. In retrospect, my own graduate student experience, where I worked largely alone and worried excessively about being original, I think it hindered my collaboration with my students who perhaps ended up working alone more than they anticipated and collaborated with me less than might've been the case.

Grudin:

I held off on publishing things to make sure that I didn’t intrude on their ability to reach their own discoveries and conclusions and see them as original. However, the two who finished while I was at UCI, Becky Grinter and Leysia Palen, did great work and became professors at Georgia Tech and Colorado. My largest project in the six years there had a mixed return on the effort invested. I'd urged Ron Becker and Bill Buxton to do a second edition of their Readings in Human Computer Interaction. Ron agreed, contingent on me participating. Bill dropped out, although he came in at the end, and we added Saul Greenberg toward the end. The book was organized with extensive surveys for each section of the book, which was 900 large-scale dense pages. It was a major effort to do those surveys. I thought the analyses in some of those overviews would have been the basis for journal articles. The book was widely used at first, but primarily to assign the important readings we collected, not to discuss our overviews. We got very little feedback or comment on the overviews.

Lewis:

I assigned only the overviews.

[Laughter]

Grudin:

I should talk more with you. Already I'm feeling better. That was great. Ben Shneiderman also did mention that he appreciated the overviews. The book came out as the web was ending the value of expensive collections of articles, many of which became freely available through university digital library membership. But that said, there was one unexpected positive outcome of the book. Ron asked me to revise and update his essay in the original edition that covered HCI history.

Grudin:

As I did this, questions arose. Why had the GOMS model approach failed to attract new researchers, and declined? It was considered by many to be our most important effort in HCI. It was featured in the readings book, and in HCI courses. Why had no manager of the National Science Foundation Human Interaction Program come from the CHI community? Or even attend any of our conferences? Why did Human Factors and CHI part ways? Why did MIS and CSCW part ways? How did the visionary writings from the 1960s that were featured in the readings books fit in? They were by authors who weren't familiar to me, or to other cognitive psychologists who were the primary founders of CHI. And so on. I wasn't able to resolve all these mysteries for the readings book under the time schedule, but I continued in the following years to interview people and read about it and think about it and came up with answers to virtually all the questions, or answers that satisfied me, and published a series of HCI history articles and handbook chapters, leading to my book in 2017 titled From Tool to Partner, The Evolution of Human Computer Interaction.

Grudin:

Since 2017, the field has continued to evolve dramatically. Not only have things happened since then, but early developments came into focus with significance I hadn't appreciated at the time. I’ll work through the big one. A few major events occurred in 1995. Commercial activity was officially permitted on the internet, and the Mosaic and Netscape web browsers brought the use of the internet to millions and then billions of people. The small, trusting, homogeneous tech world of 1980 was gone. Prior to 1995, commercial activity was prohibited. The internet was managed by the US Defense Department until 1985 and the National Science Foundation until 1995. Anonymity was almost non-existent. It was easy to identify an email author. In 1993, someone in Finland created a remailer. You could send me email, routing it through the server, which substituted a random address, and I wouldn't know who sent the email. When that came out the community strongly opposed it and the server was forced to shut down. Like, “Who would want to hide in our trusting community?”

Grudin:

Also, malware was not an issue really before 1995. No one had to spend money making software secure. Software was wonderful stuff, who would mess with it? It was possible, we knew it was possible. Primarily it was used for jokes. One worm that got loose, reportedly accidentally, the guy who wrote it ended up a professor at MIT. It wasn't seen as a major issue then. But that changed dramatically. Anonymous, bad actors swarmed in. Today, trillions of dollars are spent every year, combating bad actors. The costs are spiraling out of control, as it did half a century ago in the nuclear arms race. But back then, all the nuclear powers could meet in a room. Today, anywhere in the world an anonymous bad actor at a desk can wreak havoc across the globe in seconds. My HCI history book considered technology as a partner that was anticipated by the visionaries back in the 1960s. But it can also be a foe. That's becoming more evident, most recently, with the ransomware attacks.

Grudin:

After 1995 came the internet boom and the internet bubble. I'll say one thing about the late nineties at Irvine, then wrap up for now. Applications to be a computer science major skyrocketed, and the growth of the department was limited by the upper division lab space. We had to decide how to address this. We decided to admit more students, and use lower division courses as filters, by significantly ramping up the workload in those courses, creating painful experiences for the students and the lecturers who taught them, in order to filter the applicants. One result was a drop in female computer science majors. When we looked into it, we found that they did as well as the men, but the workloads that they knew were unnecessary turned them to other things that they'd enjoy more. Whereas the guys were more or less willing to go through that grueling experience to stay in the program.

Grudin:

One other major event, of personal significance, occurred in 1995, at an Aarhus conference called “Computers in Context: Joining Forces in Design.” Gayna Williams from Microsoft attended. It was mostly a small Scandinavian event. Why was somebody from Microsoft at a far-away Participatory Design conference? It turned out to be a visa issue. She needed to leave and reenter the US to get her passport stamped. She convinced her manager to let her do it this way, by promising to stay in a hostel to cut costs. Two years later, we were married, and I had a thousand-mile commute between Seattle and Irvine. I was looking for a reason to be in Seattle. I did a sabbatical at the University of Washington. I consulted one summer at Boeing. My NSF grant enabled me to spend another summer at Boeing. In that second summer, I was invited to visit Microsoft Research, and on the spot asked if I'd like to join. I agreed to consult one day a week. The antitrust charge hadn't seemed accurate to me, because I'd been using Unix boxes and Macs and other extinct systems, and never a PC, as a computer professional. I liked the fact that Microsoft greatly reduced the consumer price for office applications back in the eighties. But until I spent the summer consulting there and discovering that it wasn't the monolithic force that we imagined down in California, I hadn’t been ready to join. Then in 1998 I did join, and moved from Southern California to Seattle.

Lewis:

Cool. Is that a wrap for now?

Grudin:

Yeah. I think that sounds good.

Microsoft Research[edit | edit source]

August 5th, 2021

Grudin:

Okay. My key reason for joining Microsoft was getting to the same city as Gayna, who couldn't find a comparable position in Orange County, California. But the Information Computer Science department at UC Irvine had tensions and at that point. Rob Kling had left for Indiana, and John King and Mark Ackerman, the other faculty in the group, would soon leave for Michigan. The department was unique at UCI in not being in a School. We had no Dean to mediate disputes. Our CORPS group, the Computer, Organizations, Policy, and Society group, had an implacable foe, the Computer Systems group. They voted against all of our hires, all of our promotions, and schemed against us. The Systems group almost all came from former Soviet bloc countries, and while they had disliked the deceit and backstabbing there, that was the politics they knew.

Grudin:

Prior to the internet bubble, the department was under pressure to join a school. Physical Sciences wouldn't consider Computer Science on a par with physics or math. The Engineering School dean was a big problem. We seriously considered joining the Arts and Humanities School. People made a case for it. The Systems group struck a secret deal with the Engineering dean, who said he would take the department if the CORPS group was let go. The department ended up making no decision. After I left, the department became a school, as Computer Science built up. The CORPS faculty who had left in the meantime were replaced by people aligned with CSCW and HCI, less of the Organizations, Policy and Society side, but good people.

Grudin:

Throughout all of this, we were supported by the Software Engineering group, and we were accepted by AI and Theory. But it was stressful, and relied on Rob Kling, who had forceful, brilliant defenses against attacks. I was on good personal terms with the groups. I worked with the Software folks. I played bridge with some of the AI faculty, and I played tennis with the leader of the Systems group, whom I disarmed early on by calling his shots “in” if they were within six inches of the line. Finally, he said after one, "I think that was out." I said, "I played it." But the fact that we played tennis didn't mean that he wouldn't vote against my promotions.

Grudin:

Some people thought that I went to Microsoft for the money, but in my career, I never negotiated effectively for salary. For my first job in 1973, I was told when I was given an offer, “The only problem is your requested salary.” I nodded. Then, “It isn't high enough.” Today, it's easier to research salaries than it was back then. My second stay at Wang, I was given a personality test, and immediately saw that it was intended to identify whether I was motivated by money, security, working on interesting problems, and so on. I figured I could make sure I found interesting problems to work on, and filled it out to reveal a financial motivation, which worked. A few years later, my manager, viewing my file, looked puzzled and said, "You can think you understand what motivates someone, but you really don't."

Lewis:

Oh, that's fabulous… what I like about being right for the wrong reason.

Grudin:

Yeah. Microsoft asked me what my full professor compensation was, and they met it. I maximized it by assuming full consulting potential, but I didn't include what the university would put into a pension, which I didn't know. It didn't occur to me to think of it. But I was happy because I continued working on what interested me. My personal net worth over the years oscillated around zero until I was 35. Now I knew I'd be okay.

What was my view of Microsoft coming in? Well, of course I'd heard about it from the press and from my wife, who was on her way to managing a group. I liked the group I consulted with over the summer. I had also become more interested in Microsoft through my teaching. In the 1980s, money was in hardware. IBM, Digital, Apple. In the '70s, at Wang, they didn't even track the cost of producing software, which was just thrown in to get people to pay for hardware.

Grudin:

Although Bill Gates saw the potential for software, it was Apple, building on what they saw at Xerox PARC next door, that pioneered interaction design, my area. However, despite not using a PC and only recently having adopted Word and PowerPoint on the Mac, I'd seen Microsoft evolve, through the first assignment I gave HCI students in a course I taught through the '90s. In the first class, I gave them a survey asking about their favorite book, favorite movie and so forth, and I asked them to identify an application that they wanted to learn and had access to, but hadn't tried yet. Then I assigned them to use it daily for a couple of weeks and record their experiences. When computer science students get to liking software, they forget the pain of learning to use it.

Grudin:

Every year some of the students chose to use a Microsoft application. Over the years, those reviews improved steadily, with comments such as “To my surprise, the help system actually helped.” By the late '90s, Microsoft was even with Apple. When Steve Jobs returned to Apple in 1997, after his 12-year absence, he fired all the HCI people, including my former advisor Don Norman, reportedly saying, "Why do we need them when we have me?"

Grudin:

In contrast, when I joined Microsoft, it had, in my informed but possibly biased opinion, the best user experience team in the world. They pored over the research literature and explored every new method that came along, to see how it could be adapted for use. Apple had its ups and downs, it eventually triumphed with perfectly timed products that had superb visual design, although some of their products missed. I concluded that Jobs wasn't entirely wrong. He had a strong sense of where to rein in high-end designs, to keep them simple and usable.

Grudin:

I told Microsoft, I expected to stay for five years. It’s been 23. Some of that is the inertia or stability. When you get a family and have kids in school, you don't want to move. But also, Microsoft is involved in almost every facet of digital technology, which is alluring to someone who is stimulated by their environment. It would be tough on a theoretician, constantly being interrupted. At Microsoft I've ended up co-authoring research papers with developers as often as other researchers. If I'd stayed at Irvine or gone elsewhere, I might've worked at greater depth on fewer topics, which has something to be said for it.

Grudin:

There was a disparity between Microsoft’s public image and reality. Soon after I arrived, America Online had the dominant instant messenger application. It only allowed its users to communicate with each other. The competitors wanted it to open up. One day people woke up to find Microsoft users could message AOL users. This made headlines. AOL sprang into action and wrote patches to block Microsoft user access. However, Microsoft engineers had figured out in advance what AOL was going to do and had already prepared another set of patches to immediately restore access. It was war. Newspaper articles analyzed Bill Gates's strategy. Why he was doing this? Well, Gates had no strategy. This was concocted by probably 20-something programmer pranksters, and when it came to Gates’s attention, he told them to stop.

Grudin:

Microsoft was actually a decentralized, chaotic place, and really more optimistic than arrogant, not as we had seen it from California. The company was open. Earlier, when my student Leysia Palen was an intern, she wrote to ask people about setting up a meeting, and the response was often, "Don't ask, find some time on my calendar and schedule it." In a way that's impossible to imagine today, I could take my personal Mac into the building, plug it into the ethernet and get some work done without any security concerns. Bill Gates was also open. Twice a year, he scheduled a Think Week where he went off with this family and read and commented on anything that employees sent, things they'd written or read.

Another example, NetMeeting was a conferencing system like Skype or Zoom. Do you remember NetMeeting?

Lewis:

In a way, yeah.

Grudin:

It was released in 1996. In 1998, when I was consulting at Boeing, it was used extremely heavily. They had a large staff to deploy it, train people on it and manage it. When I got to Microsoft, no one used it at all. In fact, Microsoft had begun trying to terminate it. For a long time it couldn't, because major customers like Boeing and Hewlett-Packard lobbied for it. At one point, I sent Bill Gates a chart from Boeing showing its use skyrocketing and said, "We shouldn't assume our customers work like we do." Gates immediately replied, copying it to the vice president in charge of NetMeeting, missing my point, telling the VP, “I said that if it had a decent user interface, it would have succeeded." Later I sent the VP a sad report that Boeing has decided to switch to WebEx. He replied, "This is great. If Boeing is abandoning our free product to pay for WebEx, and the huge cost of switching, it proves money can be made in this area." But Microsoft did drop NetMeeting, although it provided secret ways to enable customers to access it as late as Windows 7.

Grudin:

Another illustration of how things were changing: In 1993, Microsoft let a writer, Fred Moody, join a team for over a year. His book, which was not completely flattering, I Sing the Body Electronic, described visual designers agonizing over every pixel when designing interfaces. This concern was rapidly melting away, as memory and processing costs dropped dramatically. That empowerment of visual design and interaction was a major and very visible shift.

Grudin:

Most Microsoft researchers worked on systems and other issues, but the head of Microsoft Research, Rick Rashid, in an annual MSR meeting, highlighted the number of CHI and SIGGRAPH papers published. When the cost of expensive technology drops, whether it's computers or medical scanners, research centers lose their advantage. By spending money, Xerox PARC and Bell Labs and maybe IBM Research stayed decades ahead of universities and companies. Their researchers never left. But by the mid '90s, when Microsoft Research came along, anyone with a new high-end PC was on par with most Microsoft researchers. Other things changed. In 1990, we knew by name the handful of researchers in the world who were working on our topic. By 2000, thousands of entrepreneurs and students were competing in the same areas. Installing new software went from being expensive to simple and often free. I remember the day in 1999 when our prototype software automatically updated itself, without any installation process. It was magic. Now of course, we complain about these automatic updates.

Grudin:

I'll mention a couple of broad areas that drew me in over the years, and a couple unpublished projects. At Microsoft, I initially worked mostly on streaming media prototypes for higher education. We deployed them in internal training courses. They were used in some courses at the University of Washington and MIT. They were a little ahead of their time, because although most students had computer access at home or in labs, wireless hadn't arrived, and our software wouldn't run on low-end PCs of the time. We published a lot of papers based on that work. Last summer, I published an article that summarized some of our findings, that were relevant 20 years later to those adopting remote learning via Zoom and other tools during the pandemic.

Grudin:

As I said earlier, I also worked on personas and HCI history. In 1985, when I arrived the second time at Wang Labs, they sold an email system, but internally email was called “a way people waste time.” Managers disliked email. They dictated memos. I was told when I arrived, “keyboard use is secretarial, avoid using a keyboard.” I didn't avoid keyboards, and I didn't become a manager either. An illuminating paper came out a few years later, a study of email use by the ethnographer Constance Perrin, describing use at similar Hewlett-Packard sites.

Grudin:

As a developer and user of email, I saw two roles associated with email: ‘sender’ and ‘receiver.’ Perrin also observed two critical roles, but they were ‘individual contributor’ and ‘manager.’ Individual contributors were mostly interrupt-driven, they were focused on rapid information-sharing, which email was perfect for. Managers had heavily scheduled time, and preferred formal communication. Email was a pain. It also supported rapid dissemination of rumors that they'd have to deal with, and announcements that they'd prefer to have time to shape for their group. Some managers in her article, and also at Wang, received email, printed it out and put it in a folder to read prior to their next meeting with its sender.

Grudin:

Having picked up keyboard and email use earlier, I knew that they'd be useful anywhere. What was next? I kept asking younger people who came along, what were they using? They had begun when younger, but they were using the same technology: email, word processing and so forth. Then around 2005, I heard two ethnographers describe how students were using technology. A revolution that I'd stopped expecting was occurring. Messaging, blogs, wikis, social media. Again, organizations distrusted these as time-wasters. Gartner published many reports urging organizations to keep instant messaging out, to keep these other technologies out. But I knew it would change. I spent several years studying early workplace uses of these technologies to see what might be coming.

Grudin:

I published that work. Now I'll describe a couple unpublished projects. One was an email system. Someone said, "We call it email retention but really it's email deletion." We were told that starting the next April, all email a year old would be automatically deleted. IBM had such a system and some of our customers wanted it. I contacted friends at IBM who described it as a nightmare. One avoided it by not upgrading her email system. She went on medical leave and a friend upgraded it for her. When she returned all of her email had vanished. A manager told an IT guy to set up her machine to not delete email. He said, "That would be against policy." She replied, "Set it up to not delete."

Grudin:

Why did we think it would be a good idea to use it internally at Microsoft? Some guessed storage costs, but those were dropping daily. Well, companies might have bodies that they'd like to remain buried, conversations that they would prefer not to surface. But you can't legally destroy inculpatory evidence, and an embarrassing remark that makes headlines generally has little weight in court where they look for patterns of behavior over time.

The real reason turned out to be discovery costs. Microsoft and many companies are involved in far more legal proceedings than you read about. They have to pay attorneys to read all subpoenaed emails. It reportedly came to about $30 million a year. A team of about 10 people were managing the email deletion project. Some had given up other jobs to work on it, because they loved this idea. Most had information management backgrounds. They believed that only records with business value should be kept. Seeing big email folders “makes my skin crawl,” one remarked. This view came from an era of paper documents and Rolodexes when filing and finding documents was manual. It was really difficult. It was expensive. Whereas for me and others, email is a Rolodex as well as a source of a lot of information whose future value we don't know.

Grudin:

I learned that 1000 Microsoft employees were testing the software, a process referred here to as eating dogfood. I asked how it was going for these folks. An information manager beamed and said, "It's working!” The software is a little tricky, because some email, such as patent-related email or attorney-client email, can't be deleted or should be excluded from sharing. As I worked on the project, Microsoft Research was given a waiver: our email wouldn't be deleted. The most important exception to email deletion is that email and documents of any employee who is subpoenaed, or is likely to be subpoenaed, has to be retained and turned over to the court.

Grudin:

I asked, "What do the employees using it, think about it?" This surprised the team. It never occurred to them to ask. They were sure that the employees would see the value of email deletion for the company. They were really curious. They did realize that a survey and interview might uncover gripes, but they wanted to find out. So I trained them on interview basics. “Each informant represents a large group and is not to be argued with.” “Protect informants. For example, if one says they copied email to a private account—“ An information manager turned white and hissed, “A firing offence!” She didn't participate in any of the interviews, which of course did find ingenious and time-consuming ways that people were dodging deletion. One interview I couldn't participate in, due to a scheduling conflict. I asked for the notes afterwards, and was told sheepishly that there were none. "We converted him." My instruction worked to a point, but I had to be there to manage the interviews. It was worth it, to get them to appreciate this technique, and the idea that learning from your users might be worthwhile.

Grudin:

So what did we find? Well, the cost to the company, in lost time and effort from email deletion, would easily exceed $30 million annually. And we'd have to spend almost all of that $30 million anyway, because most executives and many managers are perpetually under mandatory retention for litigation. Their email was collected and read by attorneys. It was the 70% of the employees whose email would never be of interest to anyone who would pay the price for using the software. The decision about our April deployment was made at a remarkably efficient meeting. Brad Smith, the company president and top attorney, came in, went around the group of us in the room and asked questions that had two levels that I could see, and probably a third level that I couldn't see. He asked me, "We exempted Microsoft Research. Why are you here?" I replied, "I'm not here for Microsoft Research. I'm here for Microsoft." Well, that turned out to be the right answer. He moved on to the next person. The deployment was canceled.

Grudin:

There were other positive outcomes. We found user-interface fixes in the course of this that would make the system a lot less painful for customers who acquired it. I connected people in an e-discovery software firm that built tools to help attorneys go through email with people in Microsoft Exchange to find ways to improve their tools. A partner in a San Francisco law firm heard about my findings and called up. He said that some companies would use email deletion software, whatever the cost. He explained, "Phillip Morris is in the business of addicting people to something that will kill them. They'll pay what they need to as long as the business is profitable. Once it stops being profitable, they'll stop."

Grudin:

I'd seen a problem coming. I began to promote a new role for the information managers. They could work with people who were being put on retention hold for litigation, typically managers or executives, to help them organize their email, to identify project code names and any personal or company attorneys that they corresponded with, which could save a lot of time for the attorneys who would have to go through their email and documents. But it was too late. In the end, I didn't protect my informants. The information managers who had welcomed me into the group were all laid off, no longer needed to manage a deployment that wasn't going to happen. After some deliberation they invited me to their farewell party, which was a potluck since there's no morale budget for people who are laid off. I had some misgivings about how that all turned out, which was probably a reason I didn't end up writing it up.

Grudin:

One more project. I was invited to a brainstorming exercise for project called Mauna Kea, which was an expertise location system. I think the brainstorming session was run by IDEO. My advice to them was going to be “Don't bother,” because I've been involved in four or five expertise location projects. I managed one at Wang. I contributed to one at MCC and participated on two Microsoft. None succeeded. None ended up being used after being built. Mauna Kea was being run by Dana Zimmerman and the Enterprise Strategy group, which serves Microsoft’s top consultants.

Grudin:

These consultants typically had one customer. They often work at the customer’s location. Their job is less to sell them software than to enable them to make the best use of multimillion-dollar licenses that they already have. The goal is to become the best friend of the CEO or CIO if possible, to help them align their plans with our plans and answer any questions they have. Some of the questions are routine and some aren't. Mauna Kea was for the latter. The idea behind it was that with 100,000 employees, somewhere we have the answer to any question. We just have to find it.

Grudin:

Dana won me over with the first step in his proposed process. To determine the feasibility, and to build the system, we offered some customers who came to us with a problem a free consulting report, if they'd let us interview all the stakeholders, film all the work activities, collect documents, get an understanding of the problem that they thought they had, and the real problem if it differed. We did two, one with the Los Angeles County Fire Department and one with Children's Healthcare of Atlanta. In each case, the problem statement we ended up with was not the one presented. We worked with the client and found the real problem, so we wouldn't set out to answer the wrong question.

Grudin:

You might ask, what do Microsoft employees know about fighting wildfires? Well, we had a distribution list of volunteer fire fighters in the company. Other people who saw the problem said things like, "I was in the Marines and I know about communication breakdowns amidst chaos and I have some thoughts about that." The question is, how to reach them and engage them. To do this, we developed an asynchronous game. An employee could play it when they had time. In it they heard from each stakeholder in their own words. They could access information sources to get an understanding of the context and the problem. Briefly, in the wildfire scenario, the game would start off in typical game fashion. It had points that you could earn along the way

Grudin:

“You are Christine, you have gone to meet Chief Mitchell of the Los Angeles County Fire Department about a possible system license. When you get there he comes up and says, ‘Sorry, we're going to have to postpone this meeting because believe it or not, 100 yards down the road, a wildfire has broken out.’ He heads off in that direction. And Christine, you, say, ‘Wait a minute, can I stick around and see what's going on here, see how you work?’ He looks at you and says, "Don't tell anybody I said this, yes, go ahead.’ You look around and see, well, there's smoke coming up from the canyon where he's headed. Over there, there's some guy setting up a card table.” And now you as a player have to decide. You can go over to the card table, where they're putting out maps and other information. You talk with them and move on.

Grudin:

As you go through the scenario, at different steps you encounter people who explain what their job is. It's excerpts from the tapes that we recorded. You end up getting a perspective from each of the stakeholders. If you want to, you can look at some of those maps and documents, or you can just move on. People will take different approaches. Some will go over to the map table, others will say, "Hey," and head down into the canyon to see where the action is. At the end, you've sort of gone through. It's set up so that you are exposed to a lot of the different players and their perspectives, and you can read or not read the documents.

Grudin:

You can take notes as you go through. At the end you run into Chief Mitchell again. He says, "Hey, looks like we've got this under control. Now, here are the two problems that I want to talk to you about." He lays out the problems, such as organizing communication in canyons where often cell phones aren't working. How technology might help. As the game player, you are presented with the problems and asked your thoughts on what might help. You can look at other things that players have entered in before, and comment on them, agree with them or say, "Hey yeah, that's the problem and here's a technology I can think of."

Grudin:

It’s asynchronous. Employees could play it when they want. We built and tested three versions. We worked with Steve Benford's research group at Nottingham. Not me, but one of us was flying around in a helicopter during a wildfire. I spent my time with the hospital in Atlanta. At one point, I took Dana, the project lead, to try to recruit someone I knew who had worked on the Microsoft multiplayer game Asheron's Call. He ended up recruiting Dana to work on his project, which left me to manage Mauna Kea with a business manager from consulting, and a largely skunkworks team of volunteers. Many who played the first version of the game signed up to help on the project.

Grudin:

The game was well received. We presented it to a gathering of Microsoft consultants at the Seattle Convention Center. However, the head of Microsoft sales organization reportedly didn't want to tell major customers that Microsoft employees would help them by playing games, so it wasn't put into production. It did have one last chance. Microsoft had a competition to come up with ideas for building employees’ understanding of customers. One of my thoughts from the start was that one advantage to this game was that even after we finished our report for the customer, the software was there and people could play these games and learn a lot about different customers, different organizations, the different stakeholders, and how they thought and interacted and worked. It won the contest. I got an interview with the head of HR, who really liked it, but the barrier was the cost... the price of setting up a team to produce it, which wasn't in her area.

Grudin:

This reminds me of a study at the company developing e-discovery software that I mentioned. Attorneys could use the software to organize messages and documents on multiple axes. An attorney could navigate through and select groups of items that could be safely ignored or examined. They did a study and found that using a game controller instead of a keyboard enabled the attorneys to work a lot faster at this task. But there too, the management didn't want customers to see attorneys they were paying high hourly fees apparently having fun at their desks playing games. It did not move forward. Perhaps during COVID with remote work, if they kept the software around, they could have made use of it.

Grudin:

Okay, while engaged in these Microsoft projects, I became an affiliate professor at the University of Washington Information School when it formed, and later co-chaired the annual Information School Conference with the Dean. I wrote the school's first NSF grant proposal, a joint study of collaborative information retrieval and decision-making in Microsoft and Boeing design projects.

Grudin:

The Boeing participant was Steve Poltrock, my closest and longest collaborator, since 1986. Steve and I had remarkably parallel careers. He was an engineering major at Caltech. He worked in industry then returned to get a PhD in cognitive psychology at the University of Washington. He then taught for a while in Denver, joined Bell Labs and then went to MCC in Austin before I did. When I got there, we worked closely together and became good friends. When I left MCC he returned to Seattle to work as a researcher at Boeing. We had a discussion to see if we could ensure that we would continue to collaborate, a continent apart. We decided to propose a tutorial on CSCW for CHI and other HCI conferences, which also became an introductory tutorial for CSCW conferences. We gave that tutorial for almost 20 years, until the web rendered conference tutorials less necessary for getting background on a topic. In fact, as Tutorial Co-chairs we shut down the CSCW series in the early 2000s.

Grudin:

I'm bringing this up because the tutorial was one of my most valuable intellectual drivers over those decades. Of course it pushed us to stay on top of changes in a field that was moving quickly, but it was also a place to try out observations and ideas about where the uses of technology by groups, organizations and society were headed. We could try out the thoughts and get feedback, and ultimately many of these led to more polished analysis that I published, on where the field was going, in areas such as privacy and community.

Grudin:

Steve chaired three CSCW conferences over the years. We co-chaired CSCW '98, and in 2012 I led the program committee. At that point, I had become uneasy about the effect on our conferences of the high submission rejection rates in the field. Our conferences set acceptances at 20% to 25% to maintain a claim to quality. But this meant that the fate of papers is often due to chance assignments of reviewers, which I saw sitting in program committee meetings. It meant that we were saying that 80% of the work submitted by professionals in our field was not of high enough quality to share. I had misgivings about the effect of this on the field.

Grudin:

The 2012 review process had been lengthened a little to avoid the deadlines of other conferences. I proposed that we introduce a revision cycle that would enable us to increase the paper quality, as well as the validity of the review process. My other goal, not mentioned, was to increase the acceptance rate, since people would have a chance to revise and resubmit. That year we accepted 40%. Although these papers were later heavily cited, considerably more than papers from the conferences just before and even afterwards, there was concern that the high acceptance rate would be interpreted by tenure committees as indicating that this was not a strong conference. So in subsequent years, they kept the revision cycle but steadily lowered the acceptance rates back to 25%. The revision cycle idea spread to other ACM conferences. This year CHI announced that next year they will adopt it. Okay. That's what I have to say about the conference experience.

Approaching retirement[edit | edit source]

August 13th, 2021

Lewis:

All right, you're on.

Grudin:

I anticipate retiring by the end of the year. Technology and education could be my final area of research and development focus. To put my cards on the table, technology comes fourth in my prioritized list for education, after pedagogy, ongoing professional development for educators, and infrastructure. Based on my high school experience, I've long felt that every school should have a few computers for interested students, like I was. But they were expensive, not capable of much, and few teachers liked them, so not more than a few. Then, in 2014, Anoop Gupta, the former Stanford professor who hired me 16 years earlier and was now a point person for education at Microsoft, asked me to "Find out if students need a keyboard or a digital pen." iPads were popular and had neither. Windows machines had keyboards and Microsoft had pioneered digital ink.

Grudin:

I looked and found the world had changed. Tablets and laptops were capable, not very expensive, and many teachers were fine with them. Alan Kay's and Nicholas Negroponte's one laptop per child visions were premature, but computers could finally be used for more than just teaching programming or computer science. So I spent seven years in education. Microsoft Research’s interest was limited to computer science education and women in computing. In 2017, I joined a new education engineering group.

Grudin:

I looked at historical trajectories in primary and secondary schools, and different branches of higher education: community colleges, four-year colleges, research universities and online education. I published a little of this, but focused on using Microsoft's influence through product direction and design. My influence diminished when I left Microsoft Research, though. Before the move I helped save Microsoft's investment in digital pens and ink. Microsoft had just come out with touch screen support, with Windows 8, and executives were asking, "Who needs a finger and a digital pen?" None of them used digital pens and all had fingers. But from what I had observed in schools, despite the growing use of voice, in many contexts a good digital pen is powerful, but only for kids who have a device that they can use anytime, at school or at home. This has given wealthy private school kids a huge advantage, but with a pandemic and remote learning, device per student has spread dramatically, and pens are getting better and cheaper.

Grudin:

But for education as a whole, I found that change will be slower than I expected. My daughter's high school curriculum in a Seattle suburb was extraordinarily similar to mine in an Ohio village almost a half century earlier. The curriculum emerged around the Civil War. It made more sense then. For example, trigonometry was useful for new jobs in much demand, like surveying. My kids couldn't anticipate any job they might have that would require trigonometry. Experimenting with curriculum is possible in earlier grades, and later in community colleges and four-year degree programs, but high schools in the US are constrained to enable students to get into slowly-changing elite universities, whose primary goal is preparing them for graduate school or medical or business or law schools, and not the changing job scene that will be important to everyone who doesn't proceed to post-graduate education. The elite schools are risk-averse, the system works well for them. So the high school system curriculum isn't changing very quickly.

Grudin:

Microsoft's education group stopped growing. They had less need for foundational research. I've worked on software applications that I like. These would have been super-exciting at the beginning of my career, but now the opportunity costs weigh on me. If I still thought I'd live forever, as I did after I made it to 30, spending a decade on this would seem great. With maybe a decade left, it's a different calculus.

Grudin:

I've saved the discussion of Gayna William's influence on me. When we met, she was an entry-level usability engineer just out of Minnesota. When I read her undergraduate thesis from Loughborough University, a study of situations governing the effectiveness of video conferencing, and in particular of video in conferencing, I said, "You could publish this." And she did. But overall, her focus has been intellectual but not academic.

Grudin:

She became a manager of user research, and director of user experience, and then took a handful of Microsoft management jobs in different customer-facing divisions, before leaving Microsoft to consult and focus on career coaching, which she'd trained for while at Microsoft, primarily focusing on careers for women in IT. She launched a nonprofit and taught at Cornish College of Arts, and recently took a UX management position at Amazon Web Services where she's immersed in the new world of cloud computing, machine learning and data science. Through the years, we only collaborated on one paper, which covered pioneers of user centered design: Lillian Gilbreath and Grace Hopper. Grace Hopper isn't recognized for that role. However, Gayna has been a constant source of information and insights, which she notices showing up in my work. One example: it was maybe the second or third time that she raised the use of personas that I paid close attention. I then went on to write papers with John Pruitt, who was Microsoft’s personas expert and worked for Gayna.

Grudin:

My daughters Eleanor and Isabel have continually revealed the rapidly changing role or nature of interaction with technology. That started when Eleanor, when she was six, asking, "What is a typewriter?" I had to think before replying, "It's like a computer." This was ironic, because as you know, 20 years earlier we were teaching people that a word processing computer was like a typewriter. I’ll finish this look back by describing an experience that my kids set in motion, which transformed my view of human computer interaction and also my view of the impact of technology on society, by bringing into sharper focus the human side. Several years ago, Eleanor and Isobel wanted to spend time helping animals. We volunteered at the Riverside Wildlife Center in South Africa, where volunteers worked with orphaned young baboons, vervets, and samango monkeys.

Grudin:

We had a lot of tasks, but our job was to form troops that could be reintroduced into the wild, because an individual primate released into the wild will die. It needs a troop, to cooperate on foraging and identifying predators. Observing their genetically programmed troop behaviors, similar to those of the human volunteers as they drifted in and progressed quickly from novice to responsible group leader, I saw answers to my old questions about human behavior.

Grudin:

One example: the very young baboons were in a spacious enclosure with a two-gate entrance that prevented their escape. They could spend time sitting and looking around at the world outside. When one managed to slip out, leap into the trees, it might take us two days to lure it back with food or friendly attention. However, most days, some of us alpha humans showed up, opened the door, and all of us, humans and baboons, walked along a path to a watering hole a quarter mile through the jungle. The humans sat around and talked for maybe 45 minutes, with the baboons racing around, jumping in the mud and water, digging in the grass, climbing trees, clambering over us, trying to grab our sunglasses off our heads, often successfully. After a while, we stood up and started back. They scrambled to join us. And back we went. Into the enclosure. Not once did one of them take off into the trees. This was troop behavior. Genetically encoded, it enables them to survive. When they sit around thinking, they come up with ideas. Such as escaping.

Grudin:

Now, we aren't baboons, but at Riverside, I recognized distant relatives. I saw it play out as volunteers who rotated in and out, mostly students taking a gap year or month, organized themselves into troops to accomplish the jobs that we needed to do. Or, looking at a first-grade class and seeing kids jockeying for position around an alpha teacher, very similar. One day a small vervet monkey attacked my daughter Isobel out of the blue. Perhaps the top vervet noticed that she was the smallest human and thought it could move up to her position. It let out a shriek as it landed on her and instantly every vervet in the troop screamed and mobbed Isobel. Mob is a technical term to describe this behavior.

Lewis:

Yeah.

Grudin:

None stopped to think, "She's five times my size and could crush me." This is instinctual, ‘us against them.’ Isobel was the ‘them.’ She retreated from the enclosure with some scratches and bites, not really a big deal. These were small monkeys. The point is that I spent decades studying human group behavior with technology, including social media, and assumed we had rational underpinnings to our behaviors. Our ancestors for millions of years lived in groups, including extended families. Communication was face-to-face, real time. From infancy, they learned tasks they would do their entire life. They were interdependent, respected, each had roles, foraging, avoiding predators.

Grudin:

We've avidly adopted technologies and created a world that cannot be more different from that. Fewer and fewer interactions are real-time and face-to-face. And where is the community, the troop, going? We order goods and foods for delivery, we don't know clerks, bank tellers or many of our neighbors. Kids are told: “The jobs available when you get out of school, don't exist now.” In short, your parents can't help you prepare, and neither can your teachers. You'll have several jobs over your career, they're told. Forget about lifetime employment, which your ancestors through millions of years, up until your grandparents, counted on. Kids form cohorts in school, but on graduation they head in different directions, their troop has dissolved. They are primates released alone into the wild, and their genes scream, "We're doomed."

(Laughter.)

Or maybe they go online and find a group that says, "Believe what we believe and join our troop. It's easy.”

Grudin:

I don't want to go on at length about this, but it isn't confirmation bias to say that just about wherever I look, I see communities disintegrating. SIGCHI and all the ACM professional communities have steadily declined in membership since 1990, despite the number of professionals in most areas growing by one or two orders of magnitude over that time. When I arrived, Microsoft had an annual meeting in a stadium, a summer picnic for families at a large farm outside Seattle, a holiday party for each division in December, and many admins, who like secretaries in my past jobs, were the social glue of groups, tracking birthdays and circulating get-well cards. Underpaid but respected community members. All these activities and people are gone, as is the company's investment in training courses. Now, it's “Need a skill? Hire it.”

Grudin:

I started out with the Woodstock Festival euphoria that we could do anything and make the world a better place. This continued well into the social media era. It was unbelievably wonderful to reconnect with classmates and colleagues from the distant past. And I saw that we can adjust. Flaming was a major issue with early email, it was a communication that felt real time, but allowed no rapid tone adjustments. But we adapted. Flaming is not seen as being the problem now that it was then. However, more serious social media problems have surfaced, and optimism is more difficult to come by.

Grudin:

How does this come about? Well, in the book, Sapiens, Yuval Harari paints a chilling picture of terrible consequences for most of our species from the introduction of agriculture, and after reciting a litany of disastrous effects, he asked, "Why did people do this to themselves?" And answers, "For the same reason people throughout history miscalculated, they were unable to fathom the full consequences of their decisions." Unanticipated consequences.

Grudin:

I saw some of these challenges early, but I was resolutely optimistic. My tutorials 20 years ago included a section in which I contrast oral, literate, and digital cultures on several dimensions. Oral cultures were aggregative, literate were analytic, with digital cultures becoming synthetic. The last dimension was: oral cultures are empathic, literate cultures are objective. What about digital cultures? I would ask tutorial participants, “As technology lets us see people as they really are, not as we assumed they were, or as we expected them to be, will we become more tolerant of these behaviors that we hadn't recognized, or more controlling?” Most said, and I agreed, “More tolerant.”

Grudin:

In some ways, many of us are more tolerant now. But now, when I ask this, as I did at a tutorial last year, most people say, “Controlling.” The tolerance of Brave New World giving away to the intolerance of 1984 in our projections of the future. What astonishes me is that in 2016, when I subtitled my book on HCI history “From tool to partner,” I didn't see one unanticipated consequence of empowering users through computing: the empowerment of bad actors, from criminal and malicious, through greedy and heedless.

Grudin:

Even now, computer science departments, research labs, and development groups are usually small and trusting. In the past, our community was homogeneous. Now we see positive efforts to reduce the homogeneity, but within the labs, there's not attention paid to these external developments. In 2019, my email-deletion interrogator Brad Smith, along with Carol Ann Brown, published the book, Tools and Weapons, which lays out the dangers that technology has unleashed. Despite his position at Microsoft, it isn't on balance an optimistic view of our situation.

Grudin:

I did mention the huge costs of combating malware, which few outside the tech field are aware of and few inside the tech field think about. Some people worry about surveillance. We have Hollywood's evil AI scenarios, encouraged by forecasts of AI experts that we will reach the singularity that AI researchers first claimed was on the doorstep half a century ago. But ironically, the conversational agent that can pass as human, which was the original goal of AI research, as in films like Her, may never materialize due to an alliance of malicious trolls and progressive human rights defenders.

Grudin:

One last recent-research detour, an analysis of chatbots that I undertook after an educational chatbot that I worked on failed. In the course of understanding why it failed, I helped analyze the multi-year development and deployment of Zo , a conversational agent that engaged in open-ended conversations on any topic. A very distant descendant of Eliza, and also a descendant of Alice, a conversational agent around the turn of the century that inspired Spike Jonze to write the movie Her.

Grudin:

A paper on Zo was my only paper that was blocked from publication, although an earlier manager once asked whether I really thought publishing one was a good idea, and a third manager punished me after I published a paper. All of these were on AI, where I felt I was being objective. It ran against the current.

Grudin:

Open-ended conversational agents are rapidly approaching extinction. For two reasons. The endless effort to thwart organized bands of malicious trolls, who try to get them to say lewd or hateful things, is expensive. The second reason is that keeping a record of everything that people say to one of these conversational agents in hour-long discussions could be legally risky, and organizing all that information for timely, useful recall by the agent would be intractably difficult. As a result, after a three-hour conversation, someone who resumes another conversation with that agent will find it doesn't recall anything that was said in the previous session. Discussions with an amnesiac can be interesting for a while, and maybe the fodder for some Hollywood movies, but they get boring. We found that those who engage in the longest conversations have fewer sessions than other participants.

Grudin:

It's the small successes of AI that worry me most. Marketers can use sophisticated tools to discover exactly what pitches will appeal to people, convincing them to buy things that they can't afford and don't need. The result that we're seeing is a massively indebted public, from student debt to consumer and business debt, which leaves people increasingly unable to support the infrastructure and social safety net that we need. These marketers may be the descendants of those that sold agriculture to hunter-gatherers. Many of them believe their products serve a positive purpose, and they may, for those who can afford them.

Grudin:

How it happens, again, is unanticipated consequences. After a presentation last year by a researcher at a leading university who argued that it was not difficult to do more to anticipate consequences of our work, I read a past paper of theirs and saw a likely consequence so alarming that I might have decided not to publish it, if it was my work. Whether the researcher was aware of that, and making amends in this talk, I'm not sure.

Grudin:

I'll wrap up by saying where has my journey brought me. In the beginning I spent a few years on my father's path, math and physics. I didn't use them afterwards, but they did render me immune to the physics-envy of many psychologists and left me unintimidated by math models, which was beneficial.

Grudin:

I then took my mother's direction: improving the world. Whether I succeeded there is questionable, but I still see tremendous potential for technology to improve the world. The limits to what technology could do, are limits that we create. Whether our ancestors prepared us to operate effectively in a world that they couldn't imagine, I don't know. We avoided nuclear war, but we didn't bring ourselves to get rid of the things. The next generation, as they push alone into unknown worlds, will have to do far better than their parents did, possibly with fewer resources than their parents had. Perhaps this year of fires and floods may move us from ‘us-them’ to ‘we-it’, where ‘it’ is an existential threat.

Grudin:

Two approaches to social organization have triumphed on planet Earth: small primate bands collaborating intelligently, every individual having a valued role. And the social insects, with swarms that number millions and each individual expendable. With enough time, in our global network of millions that has appeared in the last two decades, natural selection may favor people who are comfortable with uncertainty, who are not hung up on security or the respect of others. But we don't have much time for natural selection to work. In the meantime, we're tied to our primate heritage. More about me and my work, and some of these thoughts can be found on my website, jonathangrudin.com. There are peer-reviewed papers, and blog essays into which I put a lot of effort. Most of those are records of how far I got on issues that I long wrestled with, including some that we've covered. With the intent that someone who's interested can get a better start than I did, and take them further than I did.

If you have questions…

Lewis:

I've got a couple. One thing you mentioned in passing was that when you moved from research into the education group, that your influence diminished, anything to say in elaborating that thought?

Grudin:

Yeah, Microsoft Research had a huge amount of respect, both internally and outside. For example, we have Executive Briefing sessions where significant people from universities and schools and businesses come to Microsoft, and get a day or sometimes two of presentations and tours. The people managing the EBCs, as they're called, like to have somebody from Microsoft Research give a talk, because people coming in will be curious. When I was in Microsoft Research in the first few years working on education, I had regular invitations to speak at these EBCs, where I'm being heard by the visitors, and being heard by internal people. Usually high-level people, sales-people and marketing, as well as others. When I was outside of Microsoft Research, they would find somebody else from Microsoft Research to do that.

Lewis:

Clear enough, clear enough.

Grudin:

Externally, not so many people noticed that I'd left Microsoft Research.

Lewis:

Yeah. Okay. And that's something you mentioned, which just surprised me. I mean, I shouldn't be surprised because I don't follow this kind of thing, but you mentioned that a decline of the SIGs (ACM Special Interest Groups). If I think about why that was surprising to me, maybe I wasn't surprised to hear that the totality of SIGs had declined, but I thought SIGCHI had continued to grow. That's not the case?

Grudin:

It's not the case. The membership is now smaller than it was in 1990.

Lewis:

Is that right?

Grudin:

Yes, and it’s been steady. I can send you the data. The same with SIGGRAPH, which is the largest .

Lewis:

No kidding.

Grudin:

However, it is true that SIGCHI has declined less than most. Quite a bit less. Part of that may be the fact that if you were a member, the conference discount was the same as the membership cost. There was no reason not to join. It was free to join whereas most SIGs--

Lewis:

If you went to a conference.

Grudin:

If you were going to the conference. The conferences brought in huge numbers of graduate students, primarily, each year. They would all join, but they wouldn't necessarily renew the next year.

Lewis:

Okay. That explains another thing that I had missed. My impression was just based on going to the conference and thinking that it was actually bigger than I remembered it, but I've attended only sporadically for many years. I don't even know when the last year of my regular attendance was.

Grudin:

The conference size has waxed and waned a bit, for CHI and CSCW and others. But the conference size has gone up primarily because of the huge increase in students, which more than made up for the departure of professionals. When I came into Microsoft, for the usability engineers or user researchers—now they’re ‘design researchers,’ usually reporting to design managers—CHI was like the prom. It was this huge event. Now none of them go, because it's not seen as relevant to them, it's only relevant to the students.

Lewis:

They think it is, or whatever. That's another subject. I can sort it all out now, so I'm not wrong in my impression that the conference is, well, it's bigger than I knew it to be, but I was mistaken in thinking that that mapped out a SIG membership in an ongoing way.

Grudin:

When I talked to the UX people, the younger ones who've come in, many of them attended one CHI, and they have a very fond, very fond memory of it.

Lewis:

Oh, really?

Grudin:

Oh yes. But they don't attend anymore.

Lewis:

That's something at some point, I guess we might discuss more. I think we have talked about it the past, but we don't need to bring it into this. The final thing is just a sort of an odd jolt that you gave me. I realize you have updated the meaning of ‘turn of the century,’ and I have not.

(Laughter)

I forget what it was that you referred to, but the wheels were going around in my head, “come on, this wasn't in the 1900s.”

Grudin:

It's been 20 years, Clayton. It's only because I've been jolted a couple of times.

Lewis:

I probably would just think of some other way to refer to that period, not because ‘the turn of the century’ wouldn't be a perfectly good way to do it, but ‘turn of the century’ so clearly has this other meaning that I would just never use it that way.

Grudin:

I used to say millennium. I used to bring in ‘the new millennium.’ But I have heard a few other people say “turn of the century.” So I've decided, okay, time to adopt it.

Lewis:

Well, and I guess they're transitional forms like ‘turn of the present century’ and stuff like that. Okay. All right. Well, I'll shut this off.

Final reflections[edit | edit source]

December 10th, 2021

Lewis:

Okay, so you're on.

Grudin:

Our discussion has resembled my career in moving from topic to topic without structured planning on my part, following my interests. My attention over time has been drawn by unexpected results, mysterious behaviors, patterns that seemed to appear, problems identified by people around me. A positive outcome of being driven by my environment was that I don't think much of what I did was a waste of time, it was motivated by the things I encountered.

Grudin:

Once I understood the unexpected, or solved the mystery, or I could explain the pattern, or help solve the problem, I turned to something else. So I didn't work to add to theory or create a large research program. I didn't encourage students to work on my projects, I assumed they’d find their own mysteries to solve. This was probably not always true. At Microsoft Research, I directed and published with summer interns.

Grudin:

Many friends developed sustained research programs. Some focused on theory. We certainly need large programs to address our existential threats, but there can be risks. At times, Moore's law diverted the flow of technology, stranding some of our colleagues on deserted beaches, including some of the most impressive expeditions.

Grudin:

Do researchers choose an approach, or does our nature decide for us? One summer I lay in the sun, reading Sartre's Being and Nothingness. After the first chapter, I showed a fellow student my notes and said something like, “I found six things I disagree with and 11 in need of supporting evidence,” and he replied, “You're going about it wrong. You don't read philosophy that way.”

Grudin:

Theory in math and physics were fine, but behavioral science seemed to lack a necessary factual foundation for theory. I rationalized that in math and physics, the earliest theoreticians were astrologers, in chemistry, they were alchemists. Those fields became theoretical after centuries of descriptive science. So what can we do? Today, we don't have centuries to figure out the causes and effects of what we're doing. You're interested in empirical work and theory, how do you fit them together?

Lewis:

Ah, well, as by chance, I happened to be talking with a colleague about this. I've always been skeptical about the prospect for a psychological theory anytime soon. Two responses to that. One, when I was doing more theoretical work, I tried to hedge my bets and work on issues that seem to be both of theoretical interest and practical interest. Over time, the practical interest came to dominate. Until very recently I’ve sort of not done any theory at all. As it happens, I'm doing more now in retirement. Mostly I just switched to more practical stuff.

Grudin:

That makes sense.

Grudin:

Another topic: today, technology enables us to travel without leaving our home or office. Travel had a major effect on my view of the world, and the possibilities for technology. Growing up, the years I spent away from the village—near Boston, and in Portland—were crucial. My postdoc in England set my research direction for the second time. While I was there, I saw the Internet arrive. During my two years in Aarhus, and I spent time in other Nordic countries, where I found different and exciting approaches to life, work and technology use. Teaching at Keio University was a major culture contrast. My earlier observations from traveling in Africa and Latin America arose because they were salient, and could have been extended.

Grudin:

I may have been especially affected because my focus is on what is around me. Travel did become easier for our generation, and many of us were changed by it. Now, we can explore everywhere through film and the web.

Grudin:

Total immersion is valuable, but could decline. Globalization is diminishing cultural differences, homogenizing the planet. It can make travel more comfortable. Food, for example. Food in England in the early 1980s was difficult, apart from ploughman's lunch and bitter, and Indian restaurants. In Scandinavia, breakfasts were good, but the stoic approach to food in the film Babette’s Feast prevailed when I was there. Nevertheless, those who appreciate novelty may now more easily find a lot of it on Amazon, without having to leave the house.

Grudin:

Following from this, how will the Web effect intellectual exploration? We can rapidly find ideas and knowledge. We love this, but it could be a double-edged sword. Even before the web, colleagues who put tremendous energy into developing an idea that they came up with, sometimes lost interest when they discovered it was less original than they’d assumed. With very little that is new under the sun, and everything easily discoverable, how will motivation and pride be affected?

Grudin:

Topics I haven't discussed, despite being a product of the 16th, include drugs sex and rock and roll, apart from my Woodstock epiphany. But rather than taking those up, I'll describe significant trends that have stood out, or grown over time.

Grudin:

First, consequences of greater visibility of our online activity. It's astounding how rapidly activity is moving online. I mentioned that I once saw this as a net positive, believing that if we discover how people really are, we’ll develop greater tolerance for diversity. This hasn't been borne out. Respect for historical and political figures has plummeted, transgressions get more attention than achievements. We really need capable people in public service. How many will subject themselves to such scrutiny?

Grudin:

I discovered early that rules and organizations are often not there to be followed. But now when they're not followed it can be very visible. In my first job, several of us were working on a large piece of code that one person at a time could edit, by manually entering assembly language instructions using switches on the computer. A change could cause mysterious errors for the next person. A rule was imposed by management that after testing an edit, we were to restore the code to its previous state. This was very onerous. I eventually realized that the real message was, “If you make a change and don't undo it, if it causes trouble for others, you're accountable.” It was an easy way to ensure that we were more careful.

Grudin:

Similarly, a team I was on met, once with a company lawyer, once with a Vice President, and were told emphatically that we could not download unauthorized applications to our computer. “Do you understand?” Everyone nodded. I waited. I was astonished that no one spoke up, because every person in the room was running unauthorized software to get our work done. I didn't speak up.

Grudin:

Later in a different context, my manager remarked, “I'm a rule-follower.” I asked, “What about those meetings where we were told not to use unauthorized software?” He replied, “I have to stop going to meetings like that.” (Laughter.)

Grudin:

Of course, now technology can detect whether such rules are being followed. This came up when I was consulting at Boeing. One poor person caught in a layoff wrote to the CEO, “I've worked here for some number of years and always worked through my shift. I've been laid off. There are some people here who post personal email during work hours. Why aren't they laid off?”

The CEO said, “I agree with this guy,” and asked an IT guy, “Can I get a list of everybody who has posted email during work hours?” The IT guy said, “Yeah.” So they got the list. They didn’t get fired. A reprimand was sent out to people who had posted during office hours.

Grudin:

This caused a firestorm. One of them was our friend Steve Poltrock’s officemate. Steve, who is pretty much a rule follower, wrote a message to the CEO saying that this was an extreme reaction. There was a discussion about it. Finally, Boeing came up with a solution. On the factory floor are U.S. postal mailboxes where people who write a letter during their lunch break or coffee break could post the mail. It is a company facility for supporting non-work-related activity. They could assume that maybe these emails were sent out during lunch breaks or coffee breaks. (Laughs) They didn’t ask the IT guy to find the time—

Lewis:

Ah ha.

Grudin:

Again, the point is that the visibility of rule-following can create problems that didn't exist before. Another example is that rules were once passed down by upper management to middle management to serve as a model that could be adjusted to fit local conditions.

Grudin:

Once such announcements were sent in email, it was too easy for groups to see inconsistencies and multiple versions. A manager would decide, “This rule is fine for sales, but isn't so good for engineering, so we won't enforce it here.” With email dissemination, that could get into the hands of people in sales, who would protest the uneven accountability for following the rule. This is a type of challenge that major religions have confronted. When deviations from the official orthodoxy are tolerated in some countries, it works smoothly until the inconsistency becomes widely visible.

Grudin:

We discussed the effect of digital communication tools in reducing reliance on, and engagement with, local and professional communities. This has been dramatic. It extends from families, where kids get more information online than from their parents, to people shopping and working from home, which Covid has accelerated. Conferences are cancelled or have low virtual attendance.

Workers with experience may see advantages and want to continue working remotely, but new hires don't get integrated. They don't pick up tacit organizational knowledge that is often communicated in informal discussions. We're starting to see examples of where management is worried about where this could be heading.

Grudin:

Within a group, individuals take on different roles and earn respect. Secretaries, and later more skilled admins, were valued members of teams I was on. They mingled and socialized with other team members. In tech companies, and perhaps others, there are far fewer in this role. They interact far less with other team members. In a university department, each faculty member had their area of expertise that other faculty would rely on if they had questions in that area. Now, it's easy to bypass the local expert and track down a highly prominent expert elsewhere, or a Ted Talk on the web.

The revered family doctor in my hometown has been replaced by revolving clinic physicians and specialists. People working in grocery stores, banks and other shops interacted with customers. Those fulfilling orders for Amazon don't.

Grudin:

We don't mix with the same people. This came up in New York, when we went to buy a ticket in advance at the box office for the Book of Mormon, the musical. The guy behind the counter window was so pleased to have a real person to talk with. He said “This never happens anymore, everybody's buying their tickets online.”

Lewis:

Huh.

Grudin:

He absolutely loved it, he talked with us. When we were leaving, he said, “Check the date on your ticket.” He had given us a date in January (instead of the next day, in November), so he had to go through the process again—because was so engaged talking with us. He said repeatedly that it was so nice to talk to a person again.

Lewis:

Huh.

Grudin:

So that's disappearing. Where is this leading? In 1995 (before the commercial internet and web), I was impressed by a book by Barbara Gutek called—It’s a long title—The dynamics of service: Reflections on the changing nature of customer/provider interactions. It’s an examination of a service transaction from a social, psychological and management perspective. She analyzed two forms of interaction: relationships versus encounters.

Grudin:

Encounters have some advantages. They can lead to less gender or racial bias, discrimination that can be prevalent where business requires building individual relationships. A lot of technology is in support of encounters. Relationships are declining. Colleges and universities are shifting from intentional communities, with faculty residing and working together, to transactional service institutions. Students can be a continent away. The challenge is that respect comes from relationships.

Grudin:

Human beings long depended on the respect of family and group members for survival. Now some turn to online groups, some of which are positive and many are not. Others turn to money. Athletes, doctors and professors who no longer have relationships with local fans, or patients, students, or colleagues, whose interactions have become more transactional, look to salary as the measure of respect. It may help, but it may not be enough.

Grudin:

I will conclude by emphasizing how much my journey relied on relationships. I described them along the way, but I would like to pull it together. My family, of course. Until recently, I usually had lunch with teammates or colleagues, and socialized with them after work. I shared ideas and paper drafts, and attended conferences with people.

My advisor, Don Norman, and Bob Glushko from UCSD. Phil Barnard, whom I worked with on my postdoc at Cambridge, later my Best Man. Steve Poltrock from MCC through the present. I've interacted with Scandinavian colleagues through to the present. You, Clayton, beginning with our discussions at the HCI Consortium in the 1980s. John King from UCI.

Grudin:

Not only has my work been improved immensely by the consideration given by these and others, our interactions have motivated me to do much of the work I’ve done.

That's what I had to say.