Vol.29 No.4, October 1997
Education always plays an important role in the annual CHI conference. The tutorial program provides a valuable opportunity for both HCI practitioners and researchers to explore new topics. Other venues, including workshops, panels, special interest group sessions, and papers are also used to explore educational issues. This year HCI Education was represented by a panel, a Special Interest Group, and several short papers discussing issues important to HCI education. This article provides a summary of these events and is intended to encourage even more activity at CHI 98.
There were several short papers that summarized the strategies used to teach HCI within specific organizations. One particularly interesting article described the experience of developing a multimedia tutorial to support continuing education in HCI (Carey, Peerenboom, and Lytwyn, 1997). In the article, the authors explain that LAUD (Learning About User-centered Design) is geared toward persuading team members that UCD can be rewarding, cost-effective, and explains the concepts so team members are aware of their roles and commitments.
The Special Interest Group provided an opportunity for people interested in HCI education to get together, discuss important issues, and share their experiences and resources (Dringus & Cohen, 1997). Approximately 25 people attended this session and numerous resources were exchanged. A more comprehensive summary of this event appears elsewhere in this issue of the Bulletin.
The goal of the panel was to present several different perspectives on what HCI education is, how it should be provided, and who should be providing it, while encouraging the audience to participate (Sears & Williams, 1997). The panelists were Tom Hewett (Drexel University), John Karat (IBM T.J. Watson Research Center), and Gail McLaughlin (Electronic Data Systems). Jean Gasen (Virginia Commonwealth University) was scheduled to participate, but was unable to attend.
The panel's activities actually began on the first day of the conference with a message board. The goal was to solicit questions from CHI attendees that would be asked during the panel. The message board resulted in more questions than could be asked during the panel, so it was considered a success. Then the real activity began. During the panel, individuals that submitted questions were invited to join the panelists to share the motivation for their question. Each panelist was then allowed a few minutes to address the question. Fortunately, the panelists agreed to answer each of these questions once again, this time in writing. The next section includes the questions asked during the panel and the panelists' written responses. The names of the individuals that asked the questions are included when available.
Tom: These remarks are made in the context of three important points. First, there is a difference between education and training. Second, it is the societal role for universities in general that they provide education rather than training. Third, faculty are best thought of as being small business entrepreneurs who are asked to play many roles, including that of educating students.
Gail: I entered the workforce years ago as a programmer in a time when a degree in computer science didn't exist. I learned to program by reading IBM programmed instruction manuals, by looking at other people's code and by testing, testing, and more testing. I learned to "Know Thy User" because it was the only way you could get program specifications. Here are a few things about the corporate world as I know it:
I know this because my current job is to teach HCI to technical leaders in EDS. I teach them what HCI is, why it is important, and how to do things like conceptual design and usability testing. Most importantly, I teach them that they need to "Know Thy User" and to involve the user throughout the process.
These technical leaders are then expected to go back to their individual organizations and act as change agents. That means that they must influence their peers and their managers to change the way they think about systems development and to begin to incorporate HCI activities into their systems development methodology.
At the other end of the spectrum, when EDS hires new people, they are taught different skills, such as: how to work effectively as a team, how to manage projects, how to follow a system development methodology, and the essentials of the EDS corporate culture.
To recap: Corporations don't recognize the value of the knowledge and skills of the HCI graduate. Universities don't teach students basic skills necessary to be effective in the corporate world.
The bottom line is that academic and corporate communities need to work together for mutual benefit. Corporations benefit because they see universities as suppliers of a product -- manpower -- and they can influence the development of that product. Universities benefit because they can use the corporation as an extension of the classroom, where students can apply what they have learned in a real world situation. Students, faculty and employees benefit through the transfer of knowledge and the relationships that form as they work together. These alliances can lead to corporate funding for academic programs as well as job opportunities for new graduates.
Corporations and universities can work together to provide what I think is really essential in HCI education. HCI graduates must:
What is the value of including academic classes which teach students about specific interface guidelines rather than solely usability principles? -- Elaine Gilman
John: At their best, interface design guidelines are intended to capture "best practice" or "accepted standards" in interface design. As such, any professional designer needs to be aware of what they cover and what they do not cover. Knowing when to apply a guideline or standard and when you are designing outside of the scope of the standard is a part of good design. Making students aware of both general principles and specific guidelines is a part of well rounded HCI education.
Tom: The value of guidelines is that they instantiate principles. Just as students may need two programming languages before they understand the core aspects of data structures, there is probably value in having students exposed to a variety of different sets of guidelines.
Gail: Academic classes that focus on specific interface guidelines are one way to reinforce usability principles. A better way, I think, is to provide a multi-platform environment where students must perform a certain percentage of their work. Individual platform standards will "flow in through their fingertips" (as Tog says) as they work. In addition to learning usability principles, students will have the opportunity to use, evaluate, and contrast different interfaces in a working environment, where they are the end user.
John talked about design and development. What about testing? -- Mike Sandin
John: I prefer to talk about "evaluation" rather than "testing". Evaluating ideas is an integrated and natural part of design -- it goes on all the time as we consider different approaches to what we might do. As such, evaluation can be more or less formal -- knowing when to subject competing ideas to more formal evaluation rather than informal discussion is more important than giving students extensive training in a specific family of evaluation methods. In academics, "testing" still usually refers to formal methods that are of lesser value in design and development.
Tom: The value of testing can not be over estimated. However, testing should not be thought of as being different than design and development. Testing should be an influential source of information which helps support design choices. Arguing about design decisions is less efficient than testing the alternatives.
Gail: Testing is inherent in the process of design and development. Design and development cannot be considered complete without testing. The product must be designed to meet the needs of the user; the only way to know that it does, is to test it with the user. The product must perform in certain ways consistently; the only way to ensure that it does, is to test it over time in different situations.
Why not train students in how to argue for HCI (& PD, etc.) in their organizations? -- David Novick
John: I think we should. Perhaps not as a specific course, but as a part of an increased emphasis on "design exercises" as a part of HCI training. Working as a part of a team gives individuals many chances to practice skills in arguing for or against different ways of allocating resources to a project -- and this is what "arguing for HCI in an organization" really boils down to.
Tom: Students should be educated to be working professionals. That should include an understanding that HCI may have to be justified to others and the knowledge required to explain the importance of HCI considerations in development.
Gail: Why not, indeed! Most endeavors in the corporate world must show value of some sort. HCI is value-based, being a discipline focused on ensuring that a product is useful and usable for the humans who use it. Anyone trained in the discipline, and related disciplines (computer science, most specifically) should be able to espouse the benefits of HCI in the context of their working environment. Students should expect that people in the corporate world may have never heard of HCI and will require exposure to many HCI lectures, and, even then, will require a significant breakthrough to actually begin applying HCI principles. So, it is not enough to be trained to argue for HCI; a heavy dose of patience and persistence is also required.
What are the essential HCI topics to cover in 1) an introductory course for industry 2) an introductory course for undergraduates? How and why do they differ? -- Joe Konstan
Tom: Most of the essential topics are covered in the ACM SIGCHI Curriculum Report and in the NSF New Directions in HCI Research, Practice, and Education report. For industry education it may be necessary to do an analysis of the existing skill base but the more likely educational need is a change in basic attitudes about the importance of and need for iterative evaluation and design.
Gail: An introductory course in HCI for industry would differ from one for undergraduates. The audience is different, with different needs and goals. The courses would contain many of the same elements, but differ in focus and depth.
An introductory course in HCI for industry would typically be a short course, offering CEU credit, probably in the evening, mostly lecture, with a few quick hands-on exercises, about 8 hours total. It would focus on the following:
The "students from industry" could be encouraged to take the "undergraduate course" if they have an interest.
The undergraduates would want a longer course, daily classes conducted over a semester, for college credit, focused on all aspects of HCI, with more depth and lots of hands-on practices. It would focus on the following:
What forces of change are at work in HCI education?
John: We are seeing many of them now. Senior people in the HCI field did not graduate from HCI programs -- they didn't exist 15 years ago. Academic institutions cannot anticipate the need for new fields such as HCI, but they can gradually respond to what seem to be long lasting needs. HCI is such a field -- new programs are emerging with curricula that are much different than those we might have assembled 10 years ago. Practice in design is much more important than it used to be, and techniques other than formal testing are now considered as a standard part of the field.
Tom: HCI education is becoming more regularized and institutionalized with a growing number of courses, programs and degrees being offered.
Gail: The 1990s may be remembered as the "decade of change" and many of the changes have a profound effect on HCI education. Each change brings with it new ways of thinking about how people interact with computers.
The biggest change is the speed with which technology is changing. Hardware is becoming less of a limitation to the types of interfaces possible. Prices keep dropping as speed increases: faster processors, more memory, huge storage capabilities, high-speed modems. Rich media like photographic images, sound, video, and animation make it easier to create more natural, engaging interfaces. New software packages make it easy to incorporate rich media and interactivity to interfaces. The Internet has taken the corporate world by surprise, and makes possible world-wide commerce on a scale and with an ease never before imagined.
The digital convergence bringing together computers, telephony, television, movies, music, and more demand totally new interface devices with hybrid capabilities. The Internet with infinite hyperlink capabilities has changed the way we think about and present information. The Internet is also shrinking our world, facilitating casual contact with people we could have never interacted with before and making us aware of the need to design interfaces for a global audience.
HCI education struggles to keep up with all these changes. We are challenged to be more creative and innovative in the way we approach interface design.
My experience with teaching HCI is that doing iterative design with usability testing creates a "paradigm shift" which cannot be learned from books or lectures. Question: Theory vs. "studio work?" -- Dag Svanes
John: I agree to a large extent. Effective HCI design is learned by doing it. I do not mean to suggest that "books have no place in HCI education". Writing enables us to pass on knowledge that might have been obtained from practice. But unless you have tried doing design and evaluation, you cannot fully appreciate what it entails.
Tom: Both theory and studio work are needed in design education. Furthermore it is important to recognize that this should be a cyclic process with several iterations on theory and studio work. An either or opposition is dangerous here since both theory and practice feed and reinforce each other.
Gail: My experience is the same as Dag's. I would like to see more courses that emphasize both theory and applied HCI, especially in the fundamental courses. Most of my students are experienced computer systems engineers who prefer to work in a semi-vacuum, designing and developing systems with little contact with the end user. Most of them have never developed a system iteratively, much less involving users in the process. The idea of a paper prototype is an anathema to them. Some refuse to do the exercise using paper prototypes. In the end, they all do it, and most become converts when they see the benefits of rapid iterations possible with paper.
Can academics teach a professional HCI program if they have never done HCI design in industry?
John: I think that they can. However, they need to be at least familiar with what the issues of the applied activity are, and not treat HCI as a purely scientific discipline. Effective teaching is a skill in itself, and I think that there are different skills involved in teaching effectively than in practicing effectively. Certainly there is considerable overlap, but I think sufficient practice in design in a non-industry setting can give an academic enough insight into the industry setting to enable them to be an effective teacher.
Tom: Yes, but it helps if they have been involved with at least one industrial strength project and met a deadline. However, if we required this as a prerequisite to teaching, several good teachers would have to stop and then we would have an even greater shortage of HCI professionals than we do at present. Good students will find ways to fill the gaps in their education if they are convinced that they need to continue learning after they leave the academy.
Gail: Yes, they can. The real question is, how effective will they be? As a student, I would prefer that the professor have some industry experience and insights from that experience. The instructor may have strong ties to industry in some way (coordinating industry internships, for example) and have developed insights based on multiple associations which could be more illuminating than experiences in a single company. Another alternative is for the instructor to invite guest speakers from industry to share their experience and insights with the class.
If HCI education has a goal of preparing students to work in a corporate environment, how do we teach effective use of multi-disciplinary design/development teams? -- Kari Swanson
John: In my experience, engineering schools and programs do this much better than "science" programs do. I would place more emphasis on practice -- on individual and team exercises focused on the creation and evaluation of interfaces and systems.
Tom: HCI education should have the primary goal of preparing educated citizens who understand the importance of designing and developing human centered technology. In addition, some of those citizens will want to work in an environment where they can make a difference. However, because of the differences in time scale, it is difficult to provide the requisite education about working in multidisciplinary design/development teams. A typical University term runs between 10 and 15 weeks, and often the project time is only a few of those weeks. It is not unusual to have industrial development projects which run for several years. Furthermore, in a typical class the students are self selected and multidisciplinary groups may be hard to find.
Gail: I would hope that one of the goals of HCI education is to prepare students to work in a corporate environment. In the corporate environment, working effectively in multi-disciplinary design/development teams is very much desired. One way to teach students how to work effectively in such an environment is to have them do just that. This is an area where universities and corporations can jointly provide the optimum learning environment for the student. Ideally, a multi-disciplinary team consisting of students and employees of the corporation could work together on a real project. The corporation benefits from the influx of new ideas and different perspective of the students; the students benefit from the experience of working in teams on a real project with real customers; the instructor benefits from the experience and insights of each such project.
Suppose you could hand students a package of HCI materials as they leave academia for the "real world". What would be in the package?
John: Proceedings of the professional society meetings in the field contain lots of pointers that I think students can find useful.
Tom: Rather than give them a package of materials I'd want them to recognize that "The users are always right", and, "The users typically do not know or understand their own needs", are typically both true simultaneously and that the world of that intersection is where they should plan on living.
Gail: The two most important "materials" I would include in the package are mentors: a mentor from academia and a mentor from the real world. It is scary to enter a new, unfamiliar world alone. The mentor from academia would be an old, trusted friend who understands the student's strengths and weaknesses and can provide good advice and encouragement. The mentor from the real world would be a new friend, a guide who knows the ways of the real world, who can help the student adapt to the corporate culture and create a place for themselves in the corporate world.
Should every HCI designer know how to program a GUI?
John: I think the short answer is that every HCI designer should know about how the artifacts they are designing are constructed. This applies to other fields such as building architecture and design, as well. Whether they elect to "roll up their sleeves" and actually write code or not should be considered a matter of personal preference for allocation of their efforts. I do not think that being able to write "production quality code" is a necessary part of HCI design. I think that "having some knowledge of programming" is different than "being able to be considered a programmer".
Tom: No more so than every programmer should know assembly language. No more so than every programmer should know how to run an appropriately designed evaluation study. No more so than every programmer should know how to do professional quality graphic design.
Gail: If you plan to design user interfaces on a computing device, then you should have some idea how to program a GUI. The experience will help you understand the capabilities and limitations of programming, techie terminology, and the mindset of the computer systems engineer. At the least, knowledge of programming will help you get a job where eventually you can practice your HCI skills.
CHI 97 provided several opportunities for attendees to participate in activities focused on HCI education. The participatory style of the panel provided an opportunity for everyone to contribute questions and for people that posed questions to briefly join the panel. The SIG provided an informal opportunity for educators, or those interested in education, to discuss problems they face and share resources they find useful. While many of the CHI 98 submission deadlines will have passed before you read this, at least one opportunity still exists. The deadline for Special Interest Group proposals is not until January.
Carey, T., Peerenboom, D., & Lytwyn, M. (1997). Learning About User-Centered Design: A Multimedia Case Study Tutorial. CHI 97 Extended Abstracts, New York: ACM, pp. 267-268.
Dringus, L. and Cohen, M. (1997). The HCI Educator's Open House: Exchanging Resources, Delivery Formats, Learning Strategies, and Future Concerns. CHI 97 Extended Abstracts, New York: ACM, pp. 123.
Sears, A. and Williams, M. (1997). None of the above: What is really essential in HCI education? CHI 97 Extended Abstracts, New York: ACM, pp. 109-110.
Vol.29 No.4, October 1997