Back to Tom Erickson's home page

Last Updated: 12/20/96 by now at

Putting It All Together:
Pattern Languages for Interaction Design

An extended description of a workshop at CHI '97


Thomas Erickson
Apple Research Laboratories
(now at)

John Thomas
NYNEX Science and Technology, Inc.


The Challenge: Complexity and Diversity

Interaction design is becoming an increasingly complex and diverse activity. It is becoming more complex in that communications and computational technology is being integrated into more devices and environments as it becomes cheaper and smaller. This, in combination with advances in sensing and effector technologies, provides a new arena in which interaction designers can solve (and create!) new problems. Even as the space of design possibilities increases, workplace studies are making us increasingly aware of the complexity of the socio-technical systems within which we are working to integrate our new technologies. How are we to manage this complexity?

Interaction design is becoming more diverse in that a wider range of people are becoming involved in it. Within CHI, it is well accepted that anthropologists, psychologists and visual designers, as well as computer scientists, have roles to play in systems design. As computing systems shrink in size, industrial and product designers need to work hand in hand with systems designers. The advent of virtual spaces create roles for architects and interior designers. The commercialization of video and multimedia technologies create roles for musicians, film producers, et al. While the multidisciplinary nature of interaction design brings much richness, it is also challenging because no common perspective, set of practices, or theoretical orientation can be assumed.

Another factor driving the diversification of interaction design is customization. As systems become increasingly customizable due to technologies like component software, more and more design -- in the sense of front end creation, application programming, and software configuration -- is being done in-house. Sometimes this means that traditional MIS departments are playing a role; sometimes it means that external consultants are involved; sometimes it means that end users themselves participate. In many cases, these participants lack the time, resources, or inclination to engage in research on the needs and practices of their users. And, in many cases, these participants lack formal training in design, and hence any common perspective.

A Possible Solution: Pattern Languages

So, we have a rapidly expanding game: more players and more technology projected onto workplaces which we are learning more and more about. This increasing complexity and diversity can be source of richness, or of chaos. If CHI is, indeed, looking to the future, we need to explore ways of dealing with the increasing complexity and diversity of the interaction design field. In this workshop we would like to explore one possible way of putting it all together. Our model is the work of Christopher Alexander and his colleagues who over the last few decades have looked at what works and what doesn't work in architecture and urban design. The basic approach is to closely examine particular cases, attempt to identify recurring patterns and integrate them into a language of relatively concrete patterns. Their work is codified in the book A Pattern Language [1]: each pattern is described; examples are give; empirical data supporting the pattern are referenced; and the relationship to other patterns are defined. The way of using the patterns, that is, the process of design, is described in a companion volume, The Timeless Way of Building [2].

Let's take a brief look at Alexander's Pattern Language. The language consists of a network of over 250 patterns. The patterns cover a wide range of scale, ranging from a pattern for the distribution of towns and cities to patterns for walls and room sizes. The patterns are loosely connected across levels: any given pattern typically points to smaller scale patterns which can support it, and larger scale level patterns in which it may participate. For example, a pattern called Identifiable Neighborhood (aimed, obviously enough, at creating neighborhoods with their own particular, sense of place) will involve a number of lower level patterns, possibilities including Street Cafe, Individually Owned Shops, Corner Grocery, Beer Hall, and so on. At the same time, Identifiable Neighborhood participates in larger scale patterns that specify characteristics of communities.

Now let's look at a few examples of individual patterns, though these short summaries do not give the full flavor of the careful analysis and data that are contained in the full patterns. "Eccentric Neighborhood Centers" points out that neighborhood centers should be off-center; that is, closer to downtown because people will tend to go toward the city center rather than away from it. "Beer Gardens" points out that community pubs should have activity around the edges and large tables in the middle. This encourages people to cross through the center and sit at the tables. By contrast, many bars have very small tables, and it becomes uncomfortable for a stranger to "casually" approach another because the space is too intimate. Thus, the opportunity for the pub to serve as a cohesive force is diminished. "Gradient of Privacy" says that there should be a gradual gradient from public to private space in a house; e.g., from porches that look on the street life to entry ways to public rooms to family rooms to the bedroom.

These patterns focus on the interactions between the physical form of the built environment, and the way in which its form inhibits or facilitates various sorts of personal and social behavior within it. This emphasis on the overlap between the physical and behavioral worlds brings to mind the Gibsonian concept of affordances, although what is going on here is much subtler: whereas saying that certain properties of the environment afford certain actions simply means that it makes them possible (most tables and chairs afford being stood on by a person), the Alexandrian emphasis is on characteristics of the environment which facilitate or inhibit (the presence of a table and chairs can facilitate people working together or sharing food).

While these patterns describe basic ways that space should be organized in order to have a positive impact on human feelings and behavior, we believe there is a possibility to create something analogous for processes of human interaction. As representations, Pattern Languages have very interesting properties:

We believe that these properties might enable Pattern Languages to serve as a lingua franca for the diverse community of interaction designers. (It should be noted that there is considerable interest in Pattern Languages as a vehicles for object-oriented software design reflected both books (e.g., [3, 4, 5]), and in mailing list activity (see the Patterns home page at -- however, as yet, there has been little attention given to pattern languages for interaction design.)

Goals of the Workshop

We believe that Pattern Languages have two important advantages for interaction design. First, due to their concrete nature, they offer the potential for functioning as a lingua franca among the multiple disciplines involved in design, and between designers and domain experts (i.e. end users). Second, because patterns embrace both physical and social worlds, they offer a tool for representing and reflecting on the sociotechnical systems that pervade the workplace.

This workshop will explore ways of applying Pattern Languages to interaction design problems. The approaches we're interested include (but are not necessarily limited to):
HCI Guidelines:


Submissions, etc.

This workshop will be small and intense, limited to 12 people.
Participants will be selected by the organizers on the basis of a bio
and a short (2-3 pages) position paper. In the position paper, tell
us about your interest in and experience with Pattern Languages, and
expand upon any of the above approaches that are relevant to your own
work. Optional: We would like the workshop to be more than a series of
presentations interspersed with discussions, and would thus be
interested in hearing proposals for pattern-making, -applying, -using
projects that might serve as foci for workshop activity.

When submitting your materials, please include full contact information
(e-mail, snail mail, business phone, fax, and URL's if any). Send your
material to the organizers below by e-mail, snail mail or fax. Email
is the preferred mode of submission, unless you have figures or
other non-ascii representational requirements.

Thomas Erickson
Apple Computer/Remote Office
Email: (preferred)
Vmail: (408) 974-3767
Fax: (612) 823-1576

John C. Thomas
NYNEX Science and Technology
500 Westchester Avenue White
Plains, NY 10604
F: (914) - 644 -2107

SUBMISSION DEADLINE: Friday, February 7, 1997



Participants will be expected to spend time preparing for the workshop, so that we have a common ground for discussion. Participants are expected to be familiar with Christopher Alexander's work: At a minimum, participants should read "A Pattern Language," pp. ix- xliv and Patterns 14, 18, 30, 43, 88, 90, 125, 137, 150, 165, and 243 and examine the paper by Erickson at We will also circulate the position papers of accepted participants.

Workshop Activities

We would like the workshop to be more than a series of presentations interspersed with discussions. We encourage prospective participants to suggest activities suitable for either the workshop, or smaller groups within the workshop. Examples might include exploring how to transpose "Street Cafe" [pattern #88] into cyberspace, or doing some fieldwork (e.g. observing, collecting, and generating patterns that describe Hotel Lobby interactions).

Workshop Output

The output of the workshop will be a summary report to be published in the SIGCHI bulletin and perhaps other venues, and a poster for the CHI 97 poster session. Participants will be asked to assist in preparation of the workshop report, and invited to attend the poster presentation. We will also discuss the desirability and logistics of starting a mailing list, web-based discussion area, etc. (or participating in existing pattern language fora).


1. Alexander, C., Ishikawa, S., Silverstein, M., Jacobson, M., Fiksdahl-King, I.,& Angel, S. A Pattern Language . New York: Oxford University Press, 1977.

2. Alexander, C. A Timeless Way of Building . New York: Oxford University Press, 1979.

3. Coplien, J. O., and Schmidt, D. C. (eds.) Pattern Languages of Program Design . Reading, Mass: Addison-Wesley, 1995.

4. Gabriel, R. P. Patterns of Software: Tales from the Software Community . New York: Oxford University Press, 1996.

5. Gamma, E., Helm, R., Johnson, R., and Vlissides, J. Design Patterns: Elements of Reusable Object-Oriented Software . Reading, Mass: Addison-Wesley, 1995.

Back to Tom Erickson's home page