Issue
Article
Vol.28 No.1, January 1996
Article
Issue
Issue |
Article |
Vol.28 No.1, January 1996 |
Article |
Issue |
Given the commemorative nature of this special issue of SIGCHI Bulletin, the student editors set out to crawl into the minds and times of two leading HCI figures, Dr. James Foley (of Georgia Tech's Graphics, Visualization, & Usability (GVU) Center) and Dr. Stuart Card (of Xerox PARC's User Interface Research Lab). Below we present the recollections of these world-renowned researchers on their student experiences, observations of current HCI students, and prescriptions for the future, based on an electronic interview with them.
Let's start with your beginnings, what led you to HCI?
Foley: It's a long story. When I was a kid, my grandparents gave me a Lionel train set. I kept on expanding it, and needed to learn more about electricity so I could do the wiring with blocks for controlling multiple trains. This led to a decision to study Electrical Engineering, which I did at Lehigh University. During my senior year there I became interested in computers, because our circuit synthesis professor had just come back from sabbatical at IBM and taught us programming (WIZ on a GE-225). Thus when I went to Michigan for graduate school, I transitioned into computer engineering, which in turn led me to Bert Herzog's computer graphics course. I worked for a while with the technical aspects of computer graphics, but increasingly came to see it as a tool for human-computer interaction. Working with Fred Brooks and Vic Wallace and Bill Wright at the University of North Carolina (UNC) helped with that transition.
Card: As a student, there wasn't any HCI program to be led to, so it's just sort of something I made up to do. As an undergraduate I was a physics major. Then I took a year off to run the campus computer center at Oberlin College. I came to CMU to study with Newell and Simon in a program called Systems and Communications Sciences. That program ended just as soon as I arrived as a new graduate student. I was appalled that they were starting to think of me as a plain psychology graduate student. So I plopped myself down in Newell's office and said, "This isn't what I want to be." Now Newell was this great bubbling, irresistible force for academic organizational re-engineering. To the psychology department's utter amazement, we rearranged their program, throwing out half their qualifiers and courses and substituting courses and qualifiers in computer science (harder qualifiers and courses, by the way), and we invented a joint program. I even eventually got a second office, in the computer science department, because the psychology offices were too far from the printer (there was only one for the campus) and mainly because their offices were air conditioned and the psychology department's weren't.
In my first job, I didn't exactly set out to do HCI per se (there not being any per se to exactly do). What we set out to do was applied psychology with some degree of engineering. Instead of HCI, the closest things were applied psychology (in the tradition of Welford and Broadbent), human factors (like Fitts), industrial engineering (like Gilbreth), and man-machine systems (like Liklider). I was always impatient with the notion that you couldn't treat behavioral sciences just like any other sciences. One test of whether there was anything real in them, and a brutal test it is, is whether you could use them to do anything practical. The analysis of users' behavior with computers (we actually intended to study programming) was good because the behavior was bafflingly complex, had people who spent thousands of hours doing it, and seemed within the range of the symbolic sorts of theories we had in mind. So HCI was mainly a means to test and learn about some ideas for how to do behavioral theory (and about the nature of interactive systems). The practical things -- influencing products, getting patents, inventing new paradigms -- had to be done because they are the coin of the realm for demonstrating that you could do anything. Later on, we invented the name Human-Computer Interaction in order to have a title for a book and that is the first published use I know of, although, as I said there were various versions of this stuff going on. Man-Machine Systems was already taken for something slightly different, and it wasn't sex-neutral enough. Cognitive Engineering seemed too trendy. So then, we could finally be said to be doing HCI.
How many of your graduate peers were interested in HCI issues?
Foley: Are you kidding? As a graduate student in Computer, Information, and Control Engineering from 1964 to 1969 (yes, Ph.D.'s once upon a time were completed, along with an MS, in five years or less), there weren't many. In particular, there was just one, Vic Wallace, who was building an interactive system on a DEC 338 display for creating, editing, and analyzing queuing networks. Vic and I both went to UNC, where we wrote together "The Art of Natural Graphic Man-Machine Conversation"(1) We laid out a set of design principles for graphical user interfaces.
Card: On the one hand, out of the eight in my class, probably no one else. But looking at it from the point of the relevant theories there were several. The smartest of all the graduate students was Richard Young, who was the most sophisticated student in protocol analysis and invented probably the earliest form of inheritance with Newell in the Merlin project. In computer science, there was Tom Moran, who did protocol analysis of visual tasks. Both were also Newell's students. I caused my office to be moved (more organizational hijinks) to be next to Richard's so that he would have to walk past me to go anywhere and so that I could get to know him. Later I got an office with Tom Moran.
What was the general outlook of your peers on the importance of HCI issues?
Foley: The computer science and electrical engineering community at large was either neutral or, in some cases, hostile (Real people code in octal, after all). Vic Wallace, whom I mentioned above, understood. No one else really cared.
Card: There wasn't an outlook. It didn't exist.
To your knowledge, how many of these peers went on to be come HCI professionals?
Card: Tom Moran came with me to Xerox. He founded one of the first HCI journals so he counts. On some days Richard admits he does HCI. But his real interest is cognitive architecture and HCI is an application area for that.
What are the salient events that have changed the HCI student experience over time? What specifically was their effect?
Foley: The emergence of HCI as a discipline. This was marked by several seminal events. There were several important conferences:
Card: The main change in the student experience is the creation of a community that does this sort of thing. The first step was the creation of the CHI conference. This step led to the shake-out that resulted in the decision that this area would be centered around computer science as opposed to psychology or human factors. The next step was some books about the area. Then there were steps that gave students defined roles in the community -- student volunteers at the conference, Marilyn Mantei's invention of the Doctoral Consortium, etc. Then there were the efforts at defining a curriculum, such as the CHI Curriculum Committee. Then there was the founding of the ACM TOCHI journal. All of these things create faculty, literature, funding, community -- the building blocks for an area.
If given the chance, would you have done anything different in your course of graduate study?
Foley: I would have taken a lot more psychology and a lot less circuit analysis and math. But I would certainly have taken an engineering degree.
Card: Yes, I would have persisted in my original intention of doing more engineering and math.
For the record, what was your thesis title?
Foley: Optimum Design of Computer Driven Display Systems
Card: The Psychology of Human-Computer Interaction
As the field has matured, the involvement of students in official activities has changed. What were the most significant factors in the evolution of student participation in the field?
Foley: The evolution is in two areas:
Card: As I previously mentioned, the increasing involvement of students was completely by design. People planning conference and SIGCHI activities always had an interest in trying to help the students become part of the community, so their role developed as the field developed. Students are the lifeblood of a field.
Both of you have a unique vantage from which to evaluate students & their research, first as students & then as professionals. In your experience, what have been the primary characteristics of successful HCI students?
Foley: An ability to get outside of themselves, to look at a problem as someone else might, to put oneself in another's shoes. This is the heart of understanding UI design. We design not for ourselves, but for our users. Coming at HCI from the computer science side, I find the biggest roadblock to be what I call "the arrogance of the technologist" who understands computers and can't accept that others might not understand them.
Card: Good programming and math skills help a lot, although it's not the only possible way. Good students often develop interests that really get them into some topic or other that they start being an expert on. So a good combination of practical skills and theoretical knowledge.
In choosing students to work with, what do you look for beyond common interests?
Foley: Creativity and determination, and internalized high standards of excellence.
Card: Some technical skill that can make something real happen: programming, statistics, math modeling. Also a kind of problem-solving, practical orientation. The ability to be a reasonable person to interact with counts for something too.
What weaknesses are common among some HCI students you've been in contact with?
Foley: The basic weaknesses are twofold:
Card: An inability to break big mushy problems into a series of small, doable, concrete sub-problems. This, of course, is one of the things students go to graduate school to try to improve on. Everything is possible when talking about big ill-defined problems. But getting some of them to focus, commit, and complete is harder. Lack of practical skills is sometimes a problem.
Based upon your experiences, what have been critical factors that influence students' determination of career paths (academic vs. industry)?
Card: Students have different motivations. Some really want to avoid the pressure of being a professor with its necessity to face all those classes and fight for tenure. Some want to entrepreneur. Then there is the factor of how many jobs are available and for what sorts of people at the time students are making this decision. You can more or less look at the way they walk and talk.
In your opinion, has the growing acceptance/adoption of HCI changed the caliber of incoming students?
Foley: It has broadened the pool of potential students -- but there have always been good students. Early in the development of a discipline, you tend to get proportionately more risk-takers -- people with a broad intellectual curiosity who are attracted by the chance to learn new things in an exciting and different area. Now that HCI is established, perhaps the proportion of risk-takers has decreased -- but fortunately it is certainly not insignificant, either.
Card: I think so. We have been getting incredibly good students at PARC for the summer recently. But we always had pretty good students. I have long believed that the key to developing this field is its academic institutionalization. Human factors is another attempt to do a version of this field, but it didn't get enough main-stream academic institutionalization (there are a few good schools, but not enough). You need university acceptance to get thesis work on foundational issues and to propagate training in new methods. An industrially-based field can't do this. As you get the academic institutionalization, then the quality goes up and then you get better students and faculty in the area and the quality goes up more. HCI is a very difficult sub-field to build because it's hard to get techniques of adequate quality. Right now, this seems to be working for the moment.
How do you measure dissertations? What characteristics determine the impression you walk away with after reading a dissertation?
Foley: I essentially ask two questions:
Card: My impression of a dissertation depends on whether it actually defines and solves a problem, whether someone could use the results and build on them.
Finally, these last few questions provide the soap-boxes for you to preach to the choirs. How might today's students better prepare for careers in academia?
Foley: Be sure to develop your teaching skills -- not because you think you should in order to get a teaching job, but because you truly enjoy teaching and want to excel as a teacher. The tenure pendulum is swinging back from its excessive emphasis on research to a more balanced view of research and teaching. In my own experience, the best teachers are also the best researchers. Both take inquisitiveness, logic and reasoning, and good communication skills.
Card: If a student wants to enter into the university, he or she should figure out what department to aim at. While interdisciplinary training gives a student the best shot at solving novel problems, which often consist of merging ideas from two conventional fields, most academic departments take the attitude for hiring that you should be good at whatever they are first, then you can do this HCI stuff. So computer science departments want to hire people who are computer scientists first and HCI people second. Psychology departments want to hire people who are psychologists first, HCI people second. You've got to know the secret handshake. So my advice is for a student aiming for an academic career to shape his or her education to fit into the shape of a conventional department and specialize in HCI from there, rather than to do a purely interdisciplinary program.
How might today's students better prepare for careers in industry?
Foley: Naturally, take summer jobs in research labs which you might consider for a full-time job. When visitors come to your lab from industry, don't wait for your professor to ask you to help host them -- volunteer, and ask a lot of questions about how your work might fit in, or how you might change it so it does fit in, and what the steps are of taking a research prototype all the way to product.
Card: The question here is whether students want a career in which they can discover/invent things and publish or a career that is development-oriented. For a development career, getting experience working on projects to build things is important. There is a discipline to be able to build things, especially when done in teams. Then there is making sure to learn some of the big ideas that there isn't time to learn later, but which can make you much more productive: graph theory, signal processing, wavelets, things like that.
What do you see as the hottest areas in HCI for the next 5 to 10 years?
Foley:
Card: Well, just look at technology trends and social trends. There's all those MIPS coming, there's all those MBPS coming, memory (at least disk memory) will be cheap. So things relating to high-bandwidth, cheap communications. Things to do with 1000 MIPS computers. These have deep implications, not just for new buildable widgets, but for new social realities: the transformation from the real to the virtual, the loss of locality, virtual immigration, augmented reality, the successor to the paper document, pan-galactic information-access. The notion of the workstation may disappear in favor of the notion of computational media and computer-augmented environments.
If you had to recommend one book for HCI students to read, which one and why?
Card: Either the latest edition of Foley, Van Dam, and Feiner or Newman and Lamming.
Please give one word of advice/wisdom to HCI students:
Foley: HCI is the last frontier in bringing computers to everyone -- at work, at play, at school. There are many opportunities -- and some of you will make the breakthroughs! I have no doubt that 25 years from now, we'll all look back and be amazed at how primitive the user-computer interface was in 1996 -- just as now we look back 25 years and marvel at the inadequacies of the time-sharing systems of the early 1970's in comparison to our contemporary GUIs. Said in the vernacular, "We ain't seen nothing yet."
Card: Vegetables.
Issue |
Article |
Vol.28 No.1, January 1996 |
Article |
Issue |