Design v. Computing: Debating the Future of Human-Computer Interaction

Panelists:

ABSTRACT

This debate questions the presumption that the future of human-computer interaction resides in the computing sciences. We propose the following resolution: It is resolved that the CHI community should disassociate from professional computing societies and realign closely with professional design societies. The four panelists will form two teams with Terry Winograd & Jim Faris arguing for the resolution and Paul Dourish & Wendy Kellogg arguing against it. It is our intention to evoke the widest possible range of viewpoints and discussion in the community on this very important topic for the future of human computer interaction.

Keywords

© 1997 Copyright on this material is held by the authors.

INTRODUCTION

Human Computer Interaction is associated primarily with computing. Other perspectives and disciplines, such as design, appear only tangentially related. While this may actually be appropriate, the robust survival of our field requires exposure to, enrichment from, and perhaps even focused realignment with alternative perspectives that may prove profitable to the consumers of our products.

We must seriously consider what the design disciplines can offer human-computer interaction. Thus, the central question of this debate is whether the discipline of design or the discipline of computer science offers the best opportunities for the future of human computer interaction. For purposes of elucidation, and to provoke the widest range of thought and discussion, this panel proposes to debate the following resolution:

It is resolved that the CHI community should disassociate from professional computing societies and realign closely with professional design societies.

ARGUING FOR THE RESOLUTION

Terry Winograd is Professor of Computer Science at Stanford University. He has done extensive research and writing on the design of human-computer interaction. His most recent book, Bringing Design to Software (Addison-Wesley, 1996) brings together the perspectives of a number of leading proponents of software design.

Jim Faris is a principal of Alben+Faris Inc., Santa Cruz, CA. Alben+Faris specializes in graphic and interaction design for multimedia, software & other emerging technologies. Alben+Faris designed the brand identity for the Mac OS. A guiding vision in their work is to use design to humanize the experience of using technology.

Software design is happening all the time, some of it good, and much of it mediocre or bad. We see notable examples of skilled practitioners who have learned through some combination of formal training, experience and apprenticeship. But on the average, the relevant skills aren't well taught in existing institutions and aren't well applied in the commercial world of software development. There is an important difference between the "constructor's-eye view" that dominates computer science and software engineering, and a "designer's-eye-view" that takes the system, the users, and the larger context all together as a starting point. When a constructor says that a piece of software "works" he or she means that it is robust, reliable, and meets its functional specification. When a designer says that something "works" (e.g., a building or a visual layout) there is a much broader sense -- it works for someone in a context of values and needs, to produce quality results in use.

The academic and professional field of Human-Computer Interaction has its roots in computer science (a constructors-eye engineering field), and human factors (an empirical experimental field). In the past few years, the designer's view has been moving to the fore in a number of ways, including changes in the CHI conference (e.g., design briefings), the publication of a design-oriented human computer interaction magazine (interactions), the emergence of human computer interaction programs in art and design schools (e.g., the Royal College of Art), and the centrality of interdisciplinary design in programs to promote human computer interaction teaching (e.g., workshops sponsored by Apple and Interval Research). This creative ferment has been very much to the benefit of the field, but also threatens to pull us away from our traditional academic contexts. Much of the most interesting current work does not find a comfortable home in departments of computer science (such as the one in which I work in as an human computer interaction program), and there are direct obstacles to hiring faculty and admitting students whose background and concerns are too far on the design side, as seen by mainstream computer scientists.

It is possible to lobby the existing structures to be more open, but in the end this is a weak strategy and one which succeeds only at the expense of losing coherence within the discipline. Design-oriented human computer interaction researchers and designers have very little common ground with colleagues whose work is in computational complexity theory or cache memories for parallel processing. The inevitable direction we need to follow is to create a field of interaction design which bears the same relationship to computer science that architecture does to civil engineering. Interaction design draws on the technological expertise of the computational engineering field, but has its own professional stance and educational process which is more akin to the traditional design fields. In the coming years we will see the emergence of this new professional area, and will struggle with the difficulties of building a separate identity and intellectual base.

ARGUING AGAINST THE RESOLUTION

Paul Dourish is a senior research scientist at Apple Research Labs, where he works on advanced architectures for user interface and collaborative systems. While at Rank Xerox, he worked on collaborative systems, open imple-mentation techniques and media spaces. He holds a Ph.D. in Computer Science from the University of London.

Wendy A. Kellogg is a Research Staff Member at the IBM T.J. Watson Research Center where her most recent work has been designing & developing internet software for use by K-12 students & teachers. Previously, Dr. Kellogg was General Co-Chair of CHI'94. She holds a Ph.D. in cognitive psychology from the University of Oregon.

We have to disagree with the resolution that Human Computer Interaction should reside within the discipline of design. However, we disagree with the resolution while agreeing with the fundamental idea which lies behind it. In fact, in some ways, we disagree because we agree with that fundamental idea.

The idea is that the design of interfaces and interaction has too long been the sole preserve of those who design the technology. It argues that these specialists, trained to understand computation and its applications, are not best equipped to create interfaces which are not only useful and effective, but engaging and delightful. We wholeheartedly support this. However, we feel that the resolution reflects a misplaced response to this idea, and we have three primary objections--one practical, one methodological and one conceptual.

First, we would argue that the role of human computer interaction as a discipline -- and of SIGCHI as a professional organisation -- is not just to foster communication among the initiates, but also to take a leading role in educating the wider communities of which we are part about human computer interaction. The people to whom we need to convey our deepening understandings about the importance of design in interaction are computer scientists, not designers. Teaching those involved in computer science education and policy about the broader requirements of human computer interaction would not be accomplished by severing our ties to that domain.

Second, there is significant computer science yet to be done in human computer interaction. We are approaching the point where the complexities of hardware and software systems have outstripped the capabilities of traditional interfaces. To reach deeper into the system, and to provide effective interfaces for complex, large-scale, collaborative and distributed applications will mean not just designing some new widgets, but reconsidering the way in which interactive systems are computationally constituted; we need serious computer science in human computer interaction now more than ever.

Third, and most importantly, we feel that moving human computer interaction from computer science to design would be a response to the symptom, but not the cause, of our problem. The problem is that current practice construes the concept of interactive software too narrowly. To locate human computer interaction within design is to make exactly this same mistake again. The problem isn't thinking that human computer interaction is computer science, or that it is industrial or product design; the mistake is to presume that it is either. If we are to forge a new, multi-faceted discipline of human computer interaction design, then we must first abandon the notion that system software can be broken down across these kinds of disciplinary dimensions. To locate with design rather than computer science is no solution; taking the fundamental idea seriously means abandoning the kinds of blinkers which lead to these disciplinary skirmishes.