HCI at Stanford University

Terry Winograd Computer Science Department Stanford University Stanford, CA 94305-9035 USA +1 650 723 2780 winograd@cs.stanford.edu
ABSTRACT

Bill Verplank CCRMA, Music Department Stanford University Stanford, CA 94305-9035 USA verplank@ccrma.stanford.edu
HISTORY OF THE PROGRAM

We have been working to teach HCI as a design discipline within a conventional university departmental structure. Our experience over the last 15 years has led us to a number of questions and we are entering into a new collaboration to create a multi-disciplinary program in Design.
Keywords

HCI Education, Interaction Design

INTRODUCTION

We began with a strong focus on design, initiating the HCI program with an NSF curriculum development grant that sponsored a workshop leading to the book Bringing Design to Software.[2] The majority of courses at all levels are built around team projects in user-oriented design, iterative prototyping, and evaluation of interactive systems. There are relatively few courses on specific technologies and tools, and a strong emphasis on the processes of design and innovation. From the beginning, the program was strongly shaped by our fortunate location, close to a wealth of product firms, research labs, and consultants doing HCI work in Silicon Valley. Until this year, there was only one regular faculty member (Winograd) and the teaching was primarily done by lecturers recruited from industry (including Verplank, as a consistent contributor over the history of the program). With the addition of a second CS faculty member this year (Scott Klemmer), we can offer more courses, but still depend heavily on the availability of outside resources. In a way we have been bending the conventional education of our department towards the kind of learning that goes on design-oriented schools, such as art, architecture, and some of the recent schools of interaction design, such as the Interaction Design Institute Ivrea (Verplank has served on their Steering Committee and as a visiting professor), the Royal College of Art in London, the ITP program at NYU, and TU/Eindhoven. Those programs position themselves in the design tradition, centering the education on critiques of design project work, and guiding students in the production of portfolios rather than publications. They are structured to lead students to develop the tacit knowledge that goes into doing their profession, much more than to give them the explicit articulation of knowledge of the kind in textbooks. We believe there is a role in the mainstream university for an HCI program that is inspired by the ideals of design, but is also grounded in the technical competence of computing, operating within the structures and constraints of a computer science department. We attempt to combine the design approach with other perspectives on HCI, including the technical and experimental. For a number of years we have collaborated with David Kelley of the Mechanical Engineering Design Division in a

Stanford is about to create a new multidisciplinary educational program in design, informally known as the d.school [1], which will incorporate human-computer interaction design as one of its core elements. It does not yet have formal university status but has obtained initial funding and we have begun plans to renovate a maincampus building as its home. We are in the process of developing the structures. In its initial stages, the focus will be on professional Masters degree students. The basic premise of the d.school is that students need two complementary kinds of training. The disciplinary training provided by conventional departments provides them with depth in the concepts and experience of a specific field. This gives them intellectual tools, but often misses the larger context of relevance and integration with other kinds of knowledge, which are required to innovate effectively in the real world. The d.school will complement traditional education with intensive interdisciplinary team project experiences to foster the learning of design thinking. The model for a d.school course is that a substantive project serves as a focus for students from multiple departments to work together in a dedicated project space. These project courses provide a context for students to learn the process issues of user-centered iterative integrative design and to apply and refine the skills brought by the various members of a team. A key component of the teaching method is active coaching in which individuals and teams get ongoing detailed reviews of their work from someone with relevant experience. This is like the crit of traditional design education and although it is relatively high labor compared to conventional course methods, we are able to enlist coaches (or mentors) from local industry, to complement the faculty teaching resources.

course Interdisciplinary Interaction Design. Students from CS, ME, Communications, Business and other departments have joined for team projects that combined physical and interaction design for a variety of product types. Some years we have had sponsorship and participation of experienced designers from industry (Motorola and Electronic Arts) to set directions for the project and gave ongoing feedback to provide students the benefit of their experience. This, along with other similar one-to-one collaborations across departments and schools, inspired us to develop a unified program in the d.school.
THE PROGRAM STRUCTURE TODAY

analysis, user observation, idea sketching, concept generation, scenario building, storyboards, user character stereotypes, usability analysis, and market strategies. Uses a series of design projects, some individual and some team. CS 247B. Contextual and Organizational Issues in Human-Computer Interaction(taught by Pamela Hinds, in the Dept. of Management Science and Engineering) Focus is on the contextual issues associated with designing and using computer interfaces and technology, providing insights into, experience with, and ways of understanding issues in work and consumer settings that influence the design of computer interfaces. Student team projects develop skills in: observing individuals and groups of people in context, using models of work and other activity to extend their design capabilities, identifying constraints and tradeoffs on designs within the context of use, and observing and working with people in interdisciplinary design groups. Students do an extended group observation and analysis project.
Advanced Courses

We will primarily describe the program in HCI in the Computer Science Department. There is also a variety of HCI-related research and teaching in a number of other departments, including Communications, Education, Management Science and Engineering, and Mechanical Engineering. The HCI program in Computer Science was initiated with a single course in 1990, and has grown to include programs at all levels. There is no distinct HCI degree, but there are HCI concentrations for the MS and PhD in computer science, and for the undergraduate degree in the interdisciplinary Symbolic Systems program. The curriculum is oriented primarily to the CS Masters program, which is a practice-oriented rather than a research-oriented degree. In addition to course work, a number of students are engaged in research. PhD students fulfill the general breadth requirement for the Systems area, and then do dissertation research in HCI, often working with faculty from other departments along with those in CS. No research is required for the Masters degree, but a small number of masters students choose to obtain a distinction in research by working on a research project along with their courses. We have a small number of PhD students, graduating one a year or so, with approximately 10 MS students a year (more during the Internet boom!) and about 30 undergraduate Symbolic Systems majors with an HCI concentration.
COURSES Core Courses

CS376 Research Topics in HCI --- This course is a broad graduate-level introduction to HCI research. Topics include computer-supported cooperative work; audio, speech, and multimodal interfaces; user interface toolkits; design methods; evaluation methods; ubiquitous and contextaware computing; tangible interfaces; haptic interaction; and mobile interfaces CS378 Phenomenological Foundations of Cognition, Language, and Computation --- Critical analysis of theoretical foundations of the cognitive approach to language, thought, and computation. Contrasts the rationalistic assumptions of current linguistics and artificial intelligence with alternatives from phenomenology, theoretical biology, critical literary theory, and sociallyoriented speech act theory. Emphasis is on the relevance of theoretical orientation to the design, implementation, and impact of computer systems as it affects human-computer interaction. CS 447. Interdisciplinary Interaction Design(taught jointly with David Kelley in Mechanical Engineering.) Small interdisciplinary teams develop technology prototypes combining product and interaction design. Focus is on software and hardware interfaces, interaction, design aesthetics, and underpinnings of successful design including a reflective, interactive design process, group dynamics of interdisciplinary teamwork, and working with users.
Diverse Courses

CS 147. Introduction to Human-Computer Interaction Design Usability and affordances, direct manipulation, systematic design methods, user conceptual models and interface metaphors, design languages and genres, human cognitive and physical ergonomics, information and interactivity structures, design tools and environments. Structured around a set of case examples and a team interaction design project. CS 247A. Human-Computer Interaction: Interaction Design StudioSystematic presentation and experience with methods used in interaction design including needs

Much of the variety in the program comes from a series of courses offered under a general topic listing

CS 377. Topics in Human-Computer Interaction Contents change each quarter. May be repeated for credit. Many of the offerings are by people in other departments (cross-listed) or from outside Stanford (as lecturers). More than 40 different topics have been covered, some for a single year, and some on an ongoing basis [3]. This years topics were: Introduction to Cybernetics and The Design of Systems; Design for Autism; Human-Computer Interface Technology for Music Control; The Evolution of Ideas and Products: Pen computing as a case study; Experimental Research on Advanced User Interfaces; Collaborative design and research of technology-integrated; Mobile Interaction; Interactive Media Design for Kids (grades 412) Learning Health and Biology; and Cognition in Interaction Design. A weekly speaker series, CS 547. Human-Computer Interaction Seminar is available freely on line, and is watched by a large number of people outside of Stanford. Over the past 15 years more than 400 speakers have participated, across a broad spectrum of HCI topics. [4]
RESTRUCTURING THE CURRICULUM

2. How can we appropriately integrate teachers with relevant design knowledge and experience, who are not in a standard academic career path. Our program has made extensive use of industry-based researchers and designers, both as course lecturers and as coaches and reviewers in design courses. They have often done this with no pay and little formal status in the university. This makes it hard for them to give it a high priority in their own lives and careers. To make design teaching by this group more stable and credible, it is important to be able to create full-time or substantial parttime positions that have appropriate status and resources, even though they do not fit the publications-oriented tenure-line professoriate. In the teaching of the arts and professions there are faculty tracks with modifiers such as studio, performance, and clinical. What would a corresponding track be in interaction design and HCI? 3. How do we appropriately integrate knowledge about the larger context that inspires and constrains interaction design? One of our Topics courses this year (taught jointly with Don Norman) studied the history of pen computing, focusing on the issues that surround the basic technology: user acceptance, markets and channels, development models, product categories and positioning, etc. Although these are not part of a conventional HCI curriculum, we believe it is important for HCI students to have a fundamental understanding and awareness of them, to avoid a nave view that classical usability is the dominating issue in design. This experience cant be taught in student projects, since they arent embedded in the real world. How can we use case studies (perhaps in the style of business education) to make this knowledge accessible? 4. How are students to be evaluated in design-oriented HCI activities? A key problem in putting design-oriented classes into the engineering environment is that students find it frustrating to be assessed without clear rubrics of right and wrong answers. How do we create a culture of critiquing in which rigorous evaluation and guidance can be provided in areas where the objective measures are few and far removed from the most important qualities of the work? 5. What kinds of materials can be shared across institutions to enhance the overall level of project-based learning? With conventional teaching, there are obvious materials to exchange: Writings, references, topic outlines, lectures, etc. In project-based teaching much is learned by the teachers through experience, but it is often tacit knowledge that is not shared in an explicit way: What kind of knowledge banks can we develop in this dimension? They might include ways of using infrastructural tools (e.g., one of our classes this year used Greenbergs Phidgets to explore tangible interaction); critiquing and evaluation methods;

With the advent of the d.school and the addition of a second faculty member we are at a stage where we want to redesign the curriculum. Since we have been influenced by the design approach throughout the history of the HCI program, our courses have often tended to blend the teaching of broader design methods and paradigms with the technical depth more specific to HCI. This, combined with the loosely controlled content of courses taught by volunteer outside teachers, has often led to lack of clarity in the educational goals and in repetition of material. By collaborating in the d.school program, we will be able to better structure our HCI-specific courses to develop deeper knowledge of HCI concepts and tools, while taking advantage of the broader design-oriented courses for learning about general issues and methods in user-centered design.
QUESTIONS

1. How can the long history of teaching in design disciplines be adapted to the specifics of HCI in a university setting (in our case, a School of Engineering). Much of the approach of the HCI program (and moreso the d.school) has been developed in analogy with traditional design programs. Some interaction design programs, such as the one at TU/Eindhoven have carefully articulated a theory of design learning [5] and have developed effective learning methods for a school oriented as a whole to a design philosophy. For us to fit this in with the existing departmental requirement and curriculum structures requires solutions to new problems. What has been successful in other places along those lines and how can it be adapted? 3

project briefs and critical analyses of experiences with them; a shared lore about the pitfalls to avoid and paths to seek and designing and guiding projects.
REFERENCES

3. Topics in Human-Computer Interaction, http://cs377.stanford.edu 4. Human-Computer Interaction Seminar, http://hci.stanford.edu/cs547/past/alphabetical.html 5. Vinke, D. and Overbeeke, C. J. (2004). Nine Competencies, Six Units - Industrial Design Education at TU/e. In Proceedings of the 2nd International Engineering and Product Design Education Conference, Delft, September 2004 [in press].

1. Design Institute at Stanford (d.school), http://dschool.stanford.edu 2. T Winograd, J Bennett, L De Young, B Hartfield, Bringing Design to Software , Reading, Mass.: Addison-Wesley, 1996.