Elizabeth Dykstra-Erickson, Wendy MacKay, Jonathan Arnowitz
This year the SIGCHI Executive Committee appointed two new adjunct members to the Executive Committee: Elizabeth Dykstra-Erickson, Adjunct Chair for Design Policy, and Jonathan Arnowitz, Adjunct Chair for Design Production. Both members are professional designers who not only work in high-tech startups, but also teach human factors and design. They helped the CHI2001 co-chairs bring a "design" focus to the CHI2001conference through the inclusion of two venues: Interactive Video Posters and the Design Expo.
The arduous task of defining these venues, encouraging participation, and managing the production for both analog and digital media is a fascinating story in itself. However, what the reader shall find more interesting is the fundamental topic: what is "design," and how do we define "designers?" Since it is our goal to create a space in SIGCHI for design, how we answer this question is of fundamental importance. This article should be looked upon as a starting point for our discussion, not the conclusion of it.
"...for many American designers, there was no conflict between market-oriented and sales-dominated consumerism and design that has been achieved rationally and which performs properly. Nevertheless, a generation of products have emerged ...that look nice but are difficult to use...Such ergonomic failures indicate that good performance remains more elusive than good looks."
This is not a quotation of current software/HCI practices, though it may seem cogent if not insightful on the current state of much HCI design. The quote above is from a book by Peter Dormer on the state of design just after the second world war. The stage in HCI design in which we find ourselves is by no means anything new to design. Dormer's criticism of industrial design is as relevant to design in 1945 as it is today. Today more than ever, the design of interactive systems and consumerism work hand-in-hand to bring stylish and visually appealing designs to the market. But style alone is not design. Design is much more.
"...for many American designers, there was no conflict between market-oriented and sales-dominated consumerism and design that has been achieved rationally and which performs properly. Nevertheless, a generation of products have emerged...that look nice but are difficult to use... Such ergonomic failures indicate that good performance remains more elusive than good looks." (Dormer, 1993)
The field of HCI has boundless opportunities to illustrate to the public and the software producing community that: design is a complex and thoughtful problem-solving activity; that ease of use is an important measure of marketability; and most important that good design involves appeal in multiple (and sometimes invisible) dimensions.
ACM/SIGCHI was created within the academic community by researchers, professors and students from diverse scientific and engineering disciplines. Today, SIGCHI embraces this broad academic community and is beginning to get its arms around an equally broad practitioner community. Human-computer interaction, as a field, is interdisciplinary. One of the factors that makes our SIGCHI community complex is that these different academic traditions are based on their own distinct goals and value systems, with contrary (if not contradictory) methods and criteria for judging success. This multifaceted heterogeneity makes defining "design" a complex problem. In our recent efforts for increased awareness of "design" within the CHI community, we have realized that the concept of "design" is so broad that usage of the term varies widely. Multiple perspectives can be beneficial by enriching our language and our communication, but only if we learn to distinguish the subtle (and not so subtle) distinctions between these perspectives. We may use different terms to mean the same thing. We can also use the same term to mean different things. We conclude that the word "design" has multiple meanings within CHI.
- A psychologist or ergonomist developing criteria for usability
- An engineer defining an architecture for interactive systems
- A graphic designer laying out a computer screen
- An illustrator creating a graphical language
- An interaction designer developing an interaction model
All of the above people may refer to themselves as "designers" and to their activities as "design." But they do not (and should not) mean the same thing.
The questions we want to pose to the HCI community are:
- How do we raise awareness of the fundamental differences in perspectives among people who call themselves HCI designers?
- How can HCI, as a multidisciplinary field, benefit from this diversity?
One way of communicating across these different design perspectives is to avoid the use of "design" by itself. Instead refer to the "design of" something.
The CHI community's foundation is the academic community. CHI was principally researchers, professors, and students in various disciplines which, taken together, provide the scientific basis for what we know, understand and learn about humans and their interactions with computers. Computer science, cognitive psychology, anthropology, sociology, cybernetics, ergonomics, human factors, graphic communication, industrial design, linguistics, and library science are all academic areas likely to offer anything ranging from a lecture or two to a full degree program in HCI.
Graduates of HCI programs often continue life-long academic pursuits in the domain. Some turn to industrial research firms and commercial product development companies going on to conduct research or create the technology, interaction techniques, and user interfaces that the rest of the world use and take for granted. Those schooled in the community of interest that is HCI often remain a part of the CHI community and devote their talents and energies to diverse activities and events sponsored or co-sponsored by CHI. Often, however, practitioners (even very good ones) of design in software and development are self-taught, or taught in entirely other disciplines (such as the applied arts). These practitioners are unaware or even dismissive of the CHI community from which they could learn much, and to which they could significantly contribute.
Academics are trained in the scientific method; they design and conduct experiments and carefully present their results, often in SIGCHI's publications, especially the conference Proceedings. Non-academic practitioners such as interaction designers, visual designers, and web architects sometimes come from these same academic roots, but may also frequently have a background in applied arts, where process and approach are not essentially "scientific." While there exist multiple important distinctions among the academic components of the SIGCHI communityscience is not engineering is not fine artsthere are also various perspectives and traditions in the practice of HCI in the industrial and consumer development environments. Scientists are trained to study pre-existing natural phenomena as objectively as possible, traversing back and forth between theory and empirical observation. They focus on the "why." Engineers are trained to produce solutions to technical problems. Engineers, then, focus on the "how." Practitioners, on the other hand, have widely diverse educational backgrounds, and their focus is on "what"the production or crafting of HCI artifacts.
Human Interface labs and groups within software (and hardware) development companies vary widely in their practices and their organizational composition (see the rest of this issue and last year's special issue on design for details of more than two dozen companies' and universities' diverse approaches). Whether a designer has a graduate degree in HCI or a related field, or a completely different educational background, they are likely to find themselves struggling with the same issue: good design makes a positive and significant difference in every day life. And "good design," when it happens, is nearly a miracle. Design talent does not dictate the design. All too often politics, schedules, budgets, and personalities determine the realizable design goals.
Recently the CHI community leadership recognized that, lo and behold, there aren't very many of these designers participating in CHI activities. Furthermore, other societies are creating design-focused events that cross over into the HCI domain, which is not to say that CHI has been inactive in the design field. Every year, to be sure, there are paper submissions to the CHI conference that reflect on a particular design; there are tutorials on how to be a consultant; there are workshops and Birds-of-a-Feather activities for designers. At the CHI99 conference, the Visual Design SIG drew more than 50 participants to discuss the topic of design within the CHI community, forming the genesis of the first annual interactions special issue on design. The CHI conference is primarily the domain of researchers. With the conference proceedings as the important mechanism for academics to publish and receive professional recognition (and tenure) for their work. But what of practitioners? What does SIGCHI offer them? The CHI conference and its machinations were not designed to meet the needs of practicing designers; but can it? should it? and more importantly, how?
Don Norman, in his essay "Design as Practiced" (Winograd 1996) said "design as practiced is considerably different from design as idealized in academic discussions of 'good design.'" Donspoke from his experience, from his career at the University of California San Diego, to his HCI leadership in industry. (In 1996 when his essay was written, Don was the Vice President of Apple Computer's Advanced Technology Group. Apple at that time had one of the most philosophically-saturated design communities in industrial software development.) In his essay Don draws out the complexities of "doing design" in a production environment. Not the least of these complexities is culture. He mentioned that human interface designers are only a part of the design story. Engineers, industrial designers, and marketers who develop product plans and requirements are also partners in what eventually emerges as "design." However, Don's view was as a vice president (and a psychologist), not as a designer. Psychologists, like other scientists, have extensive training in how to critique and analyze, but not in how to produce new products. The designers who put pixels to screen, words to documentation, and plastics and mechanics to industrial design view their work as the creative problem solving that brings technology and people together in a first salvo. Implementation then brings its necessary compromises that must be resolved before a product design is realized.
In her essay "Design for People at Work" (Winograd 1996), Sarah Kuhn, a professor and designer from the University of Massachusetts at Lowell, focused her attention on conflicting values, goals, and assumptions that influence the design decisions that produce software artifacts. A strong proponent of sociotechnical design, she contends that designers not only need design skills, but they also need to deeply understand the human component of HCI. Furthermore, designers should ascribe to a code of ethics that governs their behavior and encourages their reflection on the impact of their work. [The IEEE did, in fact, issue just such a code of ethics for engineers last year. But while its intentions are good and just, putting it into practice is nearly impossible, as it requires a fundamental change in values, goals, and assumptionsnot the sort of things that can be legislated nor easily changed.]
A third perspective on design is offered in the same volume by David Kelley of IDEO, a prominent design firm that produces both hard and soft design. In his interview, "The Designer's Stance," David said "basically, design has to do with intuition"; it's messy; it requires the confidence to explore; and, most important, the word "design" is problematic in part because it has such broad usage. As a product designer, David relates that acquaintances who learn he is a designer tend to ask him for his opinion of the color of their drapes. He likes the broad usage of the term "design" but offers his own: design defines what "it" ought to be, whereas engineering does it (implementation). He argues that design comprises three activities: understanding (whereby one tries to understand the "mess"), observing (how a product will be used), and visualizing (deciding what "it" is). While this is an excellent, broad definition of the discipline of design, we unfortunately tend to take the word "visualize" too literally. Design, in a cognitive sense, and often in the case of interaction design, is not necessarily visual at all (e.g., audio design and voice-based systems don't generally have a visual component; gestural input and haptic feedback systems are just as "designed" as the visual components they complement). And herein lies a dilemma for the CHI community: "design" really isn't something that can be narrowly defined.
We propose, instead, that design which encompasses the many roles and skills required to produce human-computer interaction needs to be further qualified. Instead of simply design as an intransitive verb, it would be better to think of a design "of"a transitive verb, requiring an object: design of the user experience; design of the human interface; design of the visual layout, design of... you get the concept. There is room at the CHI table, and indeed room for many more tables as well, to seat a good many practitioners, researchers, and engineers who all "do" design of, in one way or another. And it is time to encourage the conversation to develop between all designers of to share their tools, techniques, goals, values, and assumptions. Hence the continuing development of "design"-oriented venues at CHI, and our request to you to contribute your ideas, submit to these design venues, and volunteer to be a part of the conversation.
Klaus Krippendorff in his essay "On the Essential Contexts of Artifacts" or "On the Proposition that 'Design is Making Sense (of Things)'" (Margolin and Buchanan 1995) gives us the etymology of design: the Latin de + signare, meaning making something, distinguishing it by a sign, giving it significance, designating its relationship to other things, owners, users, or gods. And, based on this original meaning, Krippendorff says design is making sense (of things). Note: not making things; but making sense, of things.
Augusto Morello, in his essay "Discovering Design Means (Re-)Discovering Users and Projects "(Buchanan and Margolin 1995) defines design as "a complex of projectual acts intended to conceive products and services as a whole." He draws a distinction between two types of design. This distinction is conceivably at the heart of confusion over what a designer is and does. Analytic design (deprecated perhaps unfairly by Morello as not really design) is a process to research compromises in complicated structures; synthetic design is a process to manage complexity. In simpler terms, analytic design is solving problems; synthetic design is creating solutions in an ill-defined problem space. This echoes the distinction between the engineer's design as a finely crafted technical problem solution, and the user experience designer's design as a synthesis of human needs and solutions. It is this latter role with which we concern ourselves.
Engineers are trained to solve technical problems; however, they are rarely asked to question whether they are, in fact, solving the correct problem. Part of a designer's training is to include analysis of what the problem is as well as an analysis of what the solution might be.
The design of interactive systems isn't a well-established process that follows an accepted set of rules. It is, instead, a multi-dimensional activity with many context-dependent sets of heuristics, guidelines or techniques.
In the pursuit of "good design" a human factors professional uses rules as a tool for critiquing existing products. A psychologist uses rules to reflect our understanding of how people process and use information. The designer is obliged to articulate the guidelines and explain why they are or aren't followed. Engineers who ask for design rules usually do so only in areas in which they are not trained, for example the guidelines for laying out text on the screen, assuming that they are absolute. Contrarily, an engineer would never expect to write code by mechanically following a set of absolute rules; programming is a creative task that includes many trade-offs and decisions based on strategic and other concerns (an activity otherwise known as design). Similarly, there is no fool-proof recipe for developing a good interactive system design. The process involves complicated tradeoffs requiring the collaboration of many different kinds of professionals.
Susan Fowler, in her interview in this issue, asserts that 90 percent of interface design is done by engineers. How do non-designers end up being the largest population of designers? Mainly due to two factors. Firstly, most development teams do not recognize the need for a designer; or, secondly, they recognize the need but do not have access to designers.
Engineers who do not recognize the need for designers are not necessarily insensitive to users' needs, they are just far more engaged in their strong suitprogrammingthan in the finer points of user experience. Furthermore, some design teams are small and with limited resources. It is for this community who care but really just need to get it done that Susan Fowler and Victor Stanwyck wrote their valuable books. The "who" of design, ultimately, is not really relevant; it's the "what" that counts for an engineer.
For those who do not have access to designers, this problem has two causes: a lack of budget (consequently a lack of management awareness) or a simple lack of qualified designers. This is not only because universities aren't graduating enough technically competent designers; but also because of the nature of the new medium. The web has offered a new dynamic media for a company's presence. This medium also gives them dynamic feedback, confronting them with the importance of information design, branding, usability and visual designthe critical success factors of a good user experience. Unfortunately, this confrontation takes the form of looking to fix only one or two of these critical success factors, at the expense of the other factors. Companies have not learned to understand that all four of these factors are important for an effective design. Many companies concentrate on one of the factors at the expense of the others. Though the need is recognized, the supply of user experience professionals is short. To effectively coordinate and dedicate resources to this type of work requires managers who are equally competent in the objectives and demands of user experience excellence.
Nigel Cross, in his essay "Discovering Design Ability" (Buchanan and Margolin, 1995) believes that there are common themes across different "designing of's": Creativity and intuition are important, problems and solutions are closely interwoven, and design progress is stimulated by sketching, drawing, and modeling. Cross summarizes the major aspects of what designers do as:
- produce novel, unexpected solutions
- tolerate uncertainty, working with incomplete information
- apply imagination and constructive forethought to practical problems
- use drawings and other modeling media as means of problem solving
Furthermore, design ability comprises the abilities to:
- resolve ill-defined problems
- adopt solution-focusing strategies
- employ abductive/productive/appositional thinking
- use nonverbal, graphic/spatial modeling media
Cross suggests that design ability can be fostered. By Cross' estimation, design ability is a specialized form of intelligence, and design is not interdisciplinary, it is a discipline in its own right.
Now back to the CHI community. Is design its own discipline? Is it a sub-discipline within human-computer interaction? In the balance of meaning and aesthetics, the patterns developed through meaning have a more discernable impact than those that are more focused on art. Much like Krippendorff, Mihaly Csikszentmihaly informs us that a design is a pattern that provides an ecology of meaning. In his essay "Design and Order in Everyday Life" (Margolin and Buchanan 1995) Csikzentmihaly argues eloquently that data from his studies "suggest that (at least in our culture and in the present historical period) objects do not create order in the viewer's mind by embodying principles of visual order; they do so by helping the viewer struggle for the ordering of his or her own experience." Visual values are not determined by the viewer, they are determined by the culture, and by the meaning the viewer associates with the observed object. Design is not just art; design is not graphic design or graphic layout; design is the production of meaning involving symbols (whether graphical or not), and it is a collectively determined meaning and familiarity that produces a sense of order. Design is the construction of the user experience, and it comes from many quarters in combination: the visual aspect, the interaction aspect, and positive transference from other experiences balanced against the cognitive load for learning anew.
Is design its own discipline? Yes, and with many sub-disciplines itself. And who are the CHI designers? They are those who exercise their design ability to construct the experience of users with computers. Recognizing that "design" is an overly broad term, "design of" helps us discern those sub-disciplines that we should bring, pointedly, into the CHI discussion.
Under-represented in the CHI fold are interaction designers, graphic communication designers, industrial designers, and (not at all least) engineers who write user interface code and build their own windows, dialogs, menu structures, page layouts, alerts and notifications, icons, and other user interface elements. This is an entire community of skills and foci, and a strong designer has skills in more than one traditional academic category. Figure 1 (next page) illustrates this community of skills by breaking down the development process into stages with different tools that can be employed in each. Where an academic is likely to focus on one cell of the matrix, a production designer must be conversant with many in order to deliver a design that will produce a successful product.
There are many design-oriented conferences some that are in the art domain and not primarily HCI are of the "synthetic" (using Morello's term) variety and focus on interesting and sometimes improbable graphical ideas. The American Institute for Design this past November presented the Living Surfaces conference that addressed design-qua-HCI issues. The Designing Information Systems (DIS) conferences every two years focus on design issues. Seminars and workshops on design for usability abound. So why bother introducing design themes to CHI?
We don't necessarily need to grow our membership. We don't necessarily need to become a different kind of community. But we have within our grasp the ability to bridge the gap between research and development, between well-funded and long-term engagement with research agendas and daily design activities that could be of terrific every day value to anyone who uses a computer-driven device.
There are some voices within the CHI leadership who feel that designers are sketch-pad-totin' "creatives" with a rather low threshold for understanding (much less leveraging) research, and to attract them we need to create entertaining diversions.
Entertainmentgood! But certainly not at the expense of more thoughtful and thought-provoking fare, and not because we designers are too hands-on to sit still for a serious discussion. There are other voices within the CHI leadership who feel that designers are highly visual, have a vastly different vocabulary than academics, and can't be held to high standards for exposing their work. And further, because "designers" are all graphic artists, it's not appropriate for them to be required to use words to explain their designs and rationales. But aesthetics are only part of user experience, and designers are not mute. We do need to create venues that are tailored for interactive experiences and visual demonstrations, but can we afford to dispense with design rationale because we feel we will reduce participation by requiring it? Professional designers' designs are seldom completely unchallenged. But the standards with regards to design work for reviewing and accepting conference submissions must differ in important ways from academic papers that are properly formulated, conducted, and reported experiments. Design submissions should be expected to explain their design rationale, to include the human in the discussion, and to hold up under inspection and questioning by peers. Designers are accustomed to "design crits"peer reviews critical of the approach, the execution, and the meaning of the design.
One problem in any multi-disciplinary field is to help people trained in one discipline to appreciate the complexities of another discipline. One of our challenges in teaching people to work as members of interdisciplinary teams is to appreciate and respect the training and skill brought by people from other backgrounds. We hold that SIGCHI has some work ahead of it, to foster more bridging activities and help develop HCI design as a broad and deep interactive discipline. In the coming months we hope to develop more SIGCHI awareness of how valuable the academic world is to practitioners, and how practitioners can be closely tied to research and development from academia. And most important we invite you, the reader, to tell us how you see our field.
Elizabeth Dykstra-Erickson directs User Experience at Generic Media, Inc., a Silicon Valley startup specializing in digital media access. She spent several years at Apple designing user interfaces, and teaches interface design and human factors at University of San Francisco and University of California-Santa Cruz.
Wendy Mackay is Executive Vice-Chair of the SIGCHI Executive Committee. She has held many prominent positions in the CHI organization. She is currently a researcher at INRIA in France and teaches video prototyping techniques, among other subjects. She has broad experience in research and in designing commercial products.
Jonathan Arnowitz is a Senior Cognitive Architect and HCI Designer for Cambridge Technology Partners-Benelux (Belgium, the Netherlands, and Luxembourg). Jonathan has specialized on the implementation of user-centered design methods in both software companies and internet startups and has worked on a diverse range of production projects.
Generic Media, Inc.
Cambridge Technology Partners
Figure. The following techniques obtain from the coomponent disciplines of Human-Computer Interaction and each is influencedby its parent discipline. All have advantages and disadvantages and no technique is always best in all situations. While it is certainly not necessary to use all techniques in any one project, projects that do not use at least one from each column are likely to suffer.
- Psychologists/Human Factors Professionals:
- analyze and critique what "is"
- provide technical solutions to a defined problem
- figure out what the problem definition should be as part of the design problem
©2000 ACM 1072-5220/01/0300 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2001 ACM, Inc.