Design Patterns: Augmenting User Intention in ParametricDesign Systems


My initial research questions are “how do people’s intentions play out action?” “Is it possible for a designer to discern intention from the record of his or her designing actions?” and “can we create and improve tools to support or activate their intentions?” Refining and generalizing these questions lead me to my PhD research question and its related actions. Holding a Bachelor degree in Architecture and a Master degree in Applied Science in Interactive Arts, I am specifically interested in the cognitive and HCI issues of application design in terms of supporting designers’ intention, creativity and expertise. My current research aims to observe, understand, support and augment user intention and action in parametric design systems. The users I focus on comprise active designers – more specifically, architects and civil engineers. My objective in the study is to understand the mid-level patterns of work that recur across designers and tasks. Making such patterns explicit may result in improved expert work practices, in better learning material and suggestions for improvements in parametric design interfaces. In order to achieve these goals, I would conduct a series of qualitative studies on designers’ user experience in one parametric modeling application - Bentley’s Generative Components.


As a family of computing systems intended to facilitate design activities, conventional architectural CAD (Computer Aided Design) software comes in two flavors: drawing tools with symbol libraries and “intelligent” tools that offer component-level design, for instance, walls, doors and windows. In the former the designer has all the flexibility and freedom of the drawing representation, and access to pre-prepared drawing libraries. The latter creates a building model out of components that “know” how to interact with one another and drawing are derived from this model. The main downside of these modelling tools is that the components are fabricated by the CAD software vendors, limiting adaptability – a designer cannot add a structural element if there are no structural elements in the library to add.

The majority of professional Architects find that these two types of product will cater to most design tasks, although with drawing tools, changes and edits could impact hundreds of drawings, which have to be done manually. At the manual editing stage, the task is pure tedium – error detection and repair. Yet designers must pay full attention during this important contractual and legal process. From a cognitive perspective, attention is the means by which we actively process a limited amount of information from the enormous amount of information available to us through our senses, stored memories and other cognitive processes (Benyon, Turner et al. 2005). Many small mistakes on the drawings are hard to be detected. Actually, human error is a critical contributor to lapses in system design (Wickens, Lee et al. 2004). The “intelligent” solutions aim to overcome this by using object-oriented design but these concentrate on producing documentation and usually fail to model buildings with innovative form.

Parametric modeling software has introduced into practice computational mechanisms and interfaces for representing variation in design. These work best when variation is continuous and distinctly different alternatives are not part of the model. Using parametric modeling, it is possible to develop models that support discrete variation, but it is very difficult to understand the range of possibilities entailed by such models. Parametric modeling interfaces thus provide partial support for expressing variation and, because they are increasingly used in practice, a means by which new variational techniques can be trialled in actual use.

Bentley’s Generative Components (Aish 2003), developed by Robert Aish, Chief Scientist at Bentley Systems, is not about walls, doors windows. This tool provides an environment in which geometry (lines, arcs, circles, solids, and surfaces) can be related, transformed, generated and manipulated within a user-defined framework. The end results are complex and sculpted geometry that can be quickly generated and manipulated in real-time, allowing design exploration and variation. Currently, this system has been used to promote and educate parametric modeling in practice, reaching firms such as Foster and Partners (whose recent works include the British and Smithsonian Museum courtyard roofs and the SwissRE headquarters in London), Arup Sports (Beijing Olympic Stadium) and Kohn Peterson Fox (World Bank headquarters).

Through hosting SmartGeometry workshops around the world, the community of Generative Components users is evaluating and improving the structure and interface to make it more communicative and supportive for architects, civil engineers and constructors. These events provide great opportunity for me to observe, understand and suggest new interface ideas.


My research hypothesis has two parts: the first part is concerned with making work effective: Can a designer’s learning and working process in parametric modeling applications be well modeled by patterns? The second part concerns system design: Can design work be wellsupported by tools that can express intention, such as, “what do I mean here?” or “what is the likely step for me to take next?”

To understand more about these questions, I started to explore the literature from the perspectives of design patterns, expertise, intentional stance and user modeling.

Design Patterns

The term “design pattern” originated with Christopher Alexander to describe an established architectural configuration, its context of use and consequences. Patterns express design work at a tactical level, above simple editing and below overall conception. In the process of parametric design, patterns can be observed as a general repeatable solution to a recurring modeling problem. Architects may use the same pattern in different circumstances and may also derive new patterns as they work. This pattern concept originated in urban architecture but has been adapted quite successfully to software engineering (Gamma, Helm et al. 1995) and extended to other disciplines such as HCI (Tidwell 1998), cognitive theory, and business domains. Centering on the idea of pattern language, architect Christopher Alexander proposed a new theory of architecture, building and planning in the 70’s (Alexander, Ishikawa et al. 1977). He emphasized his belief that people have an innate ability for design that paralleled their ability to speak, and every community has its own shared language particular to its place and needs. He deployed two traditional metaphors (the biological and the linguistic) to answer to questions “what defines a true pattern” and “how do patterns generate form”. Alexander also pointed out that patterns are not isolated: they refer to other, smaller-scale patterns for the solution they describe, and they can only be used in a certain type of context, which is the result of applying larger-scale patterns.

David Week adapted Alexander’s patterns through a highly pragmatic lens. Within a comparatively small domain (for example, hotels in the tropics, local housing in Pacific Island and office workplace design), he introduced “useful patterns” instead of “a pattern language for workplace design”. He underlines that: “A pattern is a hypothesis or rule-of thumb, which can be discussed and evaluated relatively independently of other patterns. Patterns are like the individual codons in the DNA sequence for an organization’s workplace” (Week 2002). Week introduces two ways to express patterns: an informal structure and a formal structure. Although the informal structure only comprises a problem with its context, proposed solution and illustration of the solution, Week’s work in culture driven workplace demonstrates that patterns in such simple form are useful for designers.

Expertise in Design

Understanding designers also includes recognizing and supporting their expertise in design. Expertise consists of those characteristics, skills and knowledge of a system, distinguishing experts from novices and less experienced people. There are broadly two academic approaches to the understanding and study of expertise (Jackson 1998). The first understands expertise as an emergent property of communities of practice. In this view expertise is socially constructed. Tools for thinking and scripts for action are jointly constructed within social groups enabling that group jointly to define and acquire expertise in some domain. In the second view expertise is a characteristic of individuals and is a consequence of the human capacity for extensive adaptation to physical and social environments. Many accounts of the development of expertise emphasize that it comes about though long periods of deliberate practice (Ericsson 1999).

In recent years, the disciplines of cognitive science and artificial intelligence have devoted a great deal of attention to the nature of expert problem solving and decision making in professional-level tasks. The goal of the cognitive science research has been to gain an understanding of the differences between the behavior of experts and novices, and possibly to learn more about how novices can become experts. A key competency of an expert is the ability mentally to stand back from the specifics of the accumulated examples, and form more abstract conceptualizations pertinent to their domain of expertise (Benyon, Turner et al. 2005). Experts are believed to be able to store and access information in larger cognitive “chunks” than novices can, and to recognize underlying principles, rather than focusing on the surface features of problems.

I am concerned with comprehending expertise in the context of design process. Cross (2004) claimed that although designers change goals and constraints as they design, they appear to hang on to their principal solution concept for as long as possible, even when detailed development of the scheme reveals unexpected difficulties and shortcomings in the solution concept (Cross 2004). Duncker defined functional fixedness as being a “mental block against using an object in a new way that is required to solve a problem” (Dunckerm 1945). On the other hand, Cross reported protocol and interview studies with three outstanding designers and drew conclusions on the common aspects of their design strategies. He found that creative design solutions arise especially when there is a conflict to be resolved between the designer’s own high-level problem goals and the criteria for an acceptable solution established by clients or other requirements (Cross 2004). It is the break-down that helps the expert designer to think outside the box and be more creative. I want to investigate if tools can remind designers the functional fixedness in the process and inspire designers during the break-downs in the design.

Intentional Stance

In the process of formulating my research question, Dennett’s intentional stance has become a crucial concept. First introduced in the book “The intentional stance” and revisited in “Darwin’s dangerous idea” (Dennett 1995), it is the strategy of interpreting the behavior of an entity (person, animal, artifact, or the like) by treating it as if it were a rational agent that governed its “choice” of “action” by a “consideration” of its “beliefs” and “desires”.

Here is how it works: first you decide to treat the object whose behavior is to be predicted as a rational agent; then you figure out what beliefs that agent ought to have, given its place in the world and its purpose. Then you figure out what desires it ought to have, on the same considerations, and finally you predict that this rational agent will act to further its goals in the light of its beliefs. As Dennett states, “a little practical reasoning from the chosen set of beliefs and desires will in most instances yield a decision about what the agent ought to do; that is what you predict the agent will do” (Dennett 1995). Dennett’s argument is that the intentional stance gives us predictive power that we can get through no other methods. It is powerful in reverse engineering and claimed to be the best, cheapest way to solve problems. The purpose of this level of reverse engineering is not just for prying out secrets of history but to predict events in the present. The strategy of intentional stance not only helped construct my research question, but also will act as the main route for me to analyze designer’ intentions.

User Modeling

A fundamental objective of HCI research is to make systems more usable, more useful, and to provide users with experiences fitting their specific background knowledge and objectives. The challenge in an information-rich world is not only to make information available to people at any time, at any place, and in any form, but specifically to say the “right” thing at the “right” time in the “right” way. Designers of systems face the formidable task of writing software for millions of users (at design time) while making it work as if it were designed for each individual user (only known at use time) (Fischer 2001). User modeling research is one of the main approaches attempting to address these issues. A n important point about user modeling might be that use time and design time get blurred. If the system is constantly adapting or is being adapted to users, use time becomes a different kind of design time (Henderson and Kyng 1991). The need to support a broad class of different users leads to high-functionality applications with all their associated possibilities and problems. Fischer introduces that a feasible design strategy to support users in their own domain of knowledge is that system designers make assumptions about classes of users and sets of tasks in which they want to engage a design methodology leading to domain-oriented systems (Fischer 1994). Most of the user modeling software appears to use this strategy. However, in Don Norman’s recent article “Human-centered design considered harmful”, he claimed that activity-centered design could be more successful than human-centered design because of activity-centered nature and the communication of intention between builders and designers (Norman 2005). If we focus on analyzing user activities into patterns, it could be another workable approach to provide “right” information to different users.


I will use Bentley’s GenerativeComponents as the study platform to analyze designer’s activities in a parametric modeling system. I became familiar with this software more than 3 years ago. I attended the SmartGeometry workshops in Cambridge (Ontario Canada) in 2004 and a GC tutorial in Seattle in 2005. At the 2006 Subtle Technologies Symposium in Toronto, I also worked as a tutor in the GC workshop (attended by 40 participants). I will attend a Smart Geometry workshop in New York in January 2007 and be a tutor at a parametric modeling workshop in Vancouver in February 2007.

The qualitative study I will conduct includes four stages: pre-pilot study, two pilot studies and a prototype design and evaluation. In the first stage, my literature review of this study will result in an understanding of curent design principles. I plan to interview the designer of GenerativeComponents Dr. Robert Aish, to obtain his original design ideas of the system, his understanding of user experience and his assumptions in design patterns.

In the second stage, I will run a pilot study in the New York’s January workshop, with two aims. Firstly, it will allow me to gain a preliminary understanding of how experts use and adapt these systems. Second is that will develop the codable moments for a complete thematic analysis (Boyatzis 1998). The approach I use is called participant observation. With it I engage designers as a tutor and simultaneously observe and discuss how they are working. I provide immediate value through tutoring and pose little overhead on any particular designer as the discussion and observation is essentially what already happens in a tutoring session. I use the data I collect to discover recurring patterns of work. I plan to invite workshop participants to take part in the study. With their permission, I will engage them in short conversations and take notes and audio-recordings. I would aim to have two or three short conversations each day with each of my subjects. I also aim to study a diverse set of design tasks. At this stage, I would be able to develop themes and codes. I plan to set up a couple possible patterns with examples.

In the third stage, I will conduct the case study in a February 2007 workshop in Vancouver. Apart from selecting more participants to observe and getting more focused on specific patterns, I would invite some of my subjects to evaluate the patterns I conclude. Their feedback is essential to polish the final outcomes.

In the fourth stage I will propose, prototype and evaluate tools for supporting designer intention in Generative Components.

The results of this case study will provide key material for a book on patterns for parametric design currently being written by Drs. Robert Woodbury, Robert Aish and Axel Kilian. In particular, Dr. Woodbury will sponsor my participation in the SmartGeometry workshop as a part of the book project. The result of my study will be useful patterns for parametric design. A collection of such patterns can be used in many ways. Firstly, it can be used to improve CAD tool design so that to provide better support for designers Secondly, it can be a useful approach in teaching parametric design. For example, Axel Kilian’s roof tutorial can be seen as a collection of three or four simple patterns. Thirdly, some designers may have invented their own “tricks” to in their specific design tasks. These tricks may be able to be distilled into patterns that others can use.


Being interested in the cognitive and HCI issues of application design in terms of supporting designers’ creativity and expertise, I set my research hypotheses as using patterns to model a designer’s learning and working process in parametric modeling applications and supporting designer’s intuitions through applying these patterns in tool design. My current literature review is focused on the domains of parametric design, patterns, expertise, intentional stance and user modeling. In this study, my objective in the study is to understand the midlevel patterns of work that recur across designers and tasks. Making such patterns explicit will result in better learning material and suggestions for improvements in the parametric design interfaces. The study will occur in four stages comprising literature review, two pilots and system implementation and evaluation.


1. Aish, R. (2003). Bentley's GenerativeComponents: a Design
Tool for Exploratory Architecture, Bentley Systems Inc.
2. Alexander, C., S. Ishikawa, et al. (1977). A Pattern
Language: Towns, Buildings, Construction New York,
Oxford University Press.
3. Benyon, D., P. Turner, et al. (2005). Designing interactive
systems: people, activities, contexts, technologies Harlow,
Eng., Pearson Education.
4. Boyatzis, R. E. (1998). Transforming qualitative
information: thematic analysis and code development
Thousand Oaks, CA, Sage Publications.
5. Cross, N. (2004). "Expertise in Design: an Overview."
Design Studies 25(5): 427-441.
6. Dennett, D. C. (1995). Darwin's Dangerous Idea: Evolution
and the Meanings of Life New York, Simon & Schuster.
7. Dunckerm, K. (1945). "On Problem Solving." Psychological
Monographics 58(5).
8. Ericsson, K. A. (1999). Expertise. the MIT Encyclopedia of
the Cognitive Sciences. R. A. Wilson and F. C. Keil.
Cambridge, Mass., MIT press.
9. Fischer, G. (1994). "Domain-oriented design environments."
Automated Software Engineering 1(2): 177-203.
10. Fischer, G. (2001). "User Modeling in Human-Computer
Interaction." User Modeling and User-Adapted Interaction
11(1): 65-86.
11. Gamma, E., R. Helm, et al. (1995). Design Patterns:
Elements of Reusable Object-Oriented Software. Reading,
MA, Addison-Wesley.
12. Henderson, A. and M. Kyng (1991). There's no place like
home: Continuing design in use. Design at Work:
Cooperative Design of Computer Systems. J. Greenbaum
and M. Kyng. Hillsdale, NJ,, Lawrence Erlbaum Associates,
Inc.: 219-240.
13. Jackson, P. (1998). Introduction to Expert Systems. Harlow,
England, Addison-Wesley.
14. Norman, D. (2005). "Human-Centered Design Considered
Harmful." Interactions 12(4): 14-19.
15. Tidwell, J. (1998). Interaction Design Patterns. PLOP'98
Conference on Pattern Languages of Programming, Illinois.
16. Week, D. (2002). The Culture Driven Workplace: Using
Your Company's Knowledge to Design the Office. Australia,
RAIA Publication.
17. Wickens, C. D., J. Lee, et al. (2004). An introduction to
human factors engineering Upper Saddle River, N.J.,
Pearson Prentice Hall.