How Do Designers Think?
Inspiration is for amateurs. The rest of us just show up and get to work. If you wait around for the clouds to part and a bolt of lightning to strike you in the brain, you are not going to make an awful lot of work. All the best ideas come out of the process; they come out of the work itself.
Chuck Close (Close 2002)
Observing what designers do has helped produce many powerful software tools, but the question of how designers think is important because the traditional pre-digital tools we can observe being used in traditional practice might not, themselves, offer the best support for design. The question of design cognition has been addressed by numerous authors and disciplines over the years, including computer scientists Newell et al. (1958), designers Cross (2011), Lawson (1997), and Schön (1984), design computing researchers Akin et al. (1987) and Gero (1990), and psychologists Kounios and Beeman (2009). Touching, as it does, on human creativity, insight, and inspiration, the answer has proven elusive but worth exploring.
The epigraph touches on two common but seemingly opposed views of how design occurs: an “Aha!” bolt-of-lightning moment of insight (Kounios and Beeman 2009), or a reward for persistent effort (Close 2002). A careful reader might note that Close contrasts persistent work with “wait[ing] around” and finds that ideas “come out of the work,” while Kounios and Beeman acknowledge “studies show that insight is the culmination of a series of brain states and processes operating at different time scales” (Kounios and Beeman 2009). Rather than being opposing views, it seems possible that the two views describe aspects of a single process in which large- and small-scale insights grow out of close attention and engagement with a complex problem. The challenge is to capture and organize the insights while fostering attention and informing engagement.
One fundamental cognitive problem is that products such as buildings are complex, and yet the working or short-term memory of a human holds relatively little information. Long-term memory holds much more but takes longer to access and cannot be used for a dynamic process such as design. External representations, such as drawings, aid in the process, as do models that divide problems into hierarchies of smaller conceptual fragments or chunks that can be manipulated within the limits of short-term memory. Design concepts usually engage the problem by establishing relationships between a relatively small number of elements, creating a theme that can be used to address challenges at different scales and in different aspects of the design. Good concepts are not arbitrary; they fit the design problem. They “come out of the work itself” (Close 2002) and they “are often the result of the reorganization or restructuring of the elements of a situation or problem” which produces “a new interpretation of a situation … that can point to the solution to a problem” (Kounios and Beeman 2009). Before you reorganize a situation, you have to study and understand it; you must build a (conceptual) model of it. This means identifying, labeling, and organizing parts (building a representation), and defining changes or “moves” that can be carried out on the parts. If a subsequent insight or design concept re-organizes the elements of the problem, you may need a new model.
Design, whether insightful or sweat-stained, does not require a computer. Still, as information technologies have been successfully applied to observable productive design behaviors (typing, drafting, modeling), we are moving towards supporting or automating the less observable cognitive aspects of the activity, where architects’ professional identities and egos are at risk. That risk arises not only from the core difficulties of defining any design profession, but also from the notion of making design tasks computable. Nonetheless, beginning in the mid-1990s, conceptual models and a series of conferences have wrestled with aspects of creativity, design, and cognition (Gero 2011). In asking how designers think, we not only open the door to significant productivity gains for designers, we open the door on the computation and possible automation of design—replacing designers.
Whether we aim to automate design, enhance designer productivity, or just support one aspect of design, such as collaborative interaction, we need to develop a model of the process. Out of this may emerge points of leverage where computing might be well applied, or hesitancy, where (current) computing technology may be more hindrance than aid. We must look at the design process, at divergent thinking and the role of memory, media, and social process, and at convergent thinking leading to incremental improvement. In addition, as digital media increasingly take center stage in offices, we need to consider how their use interacts with our thinking and the possibility that their use has unanticipated side-effects.
Designing as Process
Invention is “one percent inspiration, ninety-nine percent perspiration.”
Attributed to Thomas Edison
One view of design and designers tends to focus on the idea itself, the designer’s proposition about what should be done, not where the idea came from. Consequently, there is a certain amount of mysticism regarding the process. People are said to “be creative,” not “act creatively” and we marvel at examples of “thinking outside the box” without necessarily being aware of or appreciating the amount of work that preceded the “Aha!” moment. Production of the idea is then sometimes seen as an act of will and design as a domain of willfulness. However, while professional designers respect the possibility of sudden inspiration, it is more often seen as a fickle source of design solutions.
The general consensus is that design is a behavior that follows a fairly common pattern that can be taught and cultivated, and that certain techniques (e.g., precedent studies, brainstorming, change of perspective) enhance creativity. In The Design of Everyday Things, Don Norman presents a very general model for the “seven stages of human action” that aligns with the stages of design shown in Figure 7.1 remarkably well (Norman 1988). His model begins with a goal (the design problem, or “preferred state”), formulates an intention of a change (“analysis”), and offers a specific plan of action meant to achieve it (“ideation”). The plan is executed (“representation”), after which one must “perceive the state of the world” (“simulation”), interpret the state of the world (“evaluation”), and evaluate the outcome (“assessment”). If not completely successful, a new goal is set and the cycle repeats. This model reminds us that the effect of a change is not always knowable in advance, and that feedback loops containing perception, interpretation, and evaluation form important steps in any action, including design.
Even if design were entirely dependent on a flash of inspiration and there is no way to promote it, there is a role for digital tools, because it will still be necessary to record, evaluate, refine, and document the idea in an external medium such as that provided by a 3D modeling environment or sketching program. While creative production remains a crucial element in design, there are a number of related cognitive activities that are also candidates for support, including problem exploration, solution representation, solution evaluation, decision support, and recording, as we will see in later chapters.
Creativity represents only one aspect of design cognition or what goes on in someone’s head during design problem solving. Other aspects include problem formulation, mental representation, and internal manipulation. Depending on what we think the core of design consists of, software augmentation or replacement will take different forms. In the following sections we’ll look at design as process or method, intuition or inspiration, memory, selection, drawing or dialog, social process, and as the end product of incremental improvement. None captures all of design; each has been the starting point for software projects in the past.
As mentioned in Chapters 1 and 6, and shown in Figure 7.1, design is usually described as an iterative or cyclic process that gradually, though not directly, proceeds from problem statement to solution. Beyond that foundation, differences appear. Some distinguish between strong design procedures or methods and strong design knowledge (Kalay et al. 1990). Others stress the role of social interaction among the participants, or the gradual construction of both problem and solution, described by more than one author as puzzle making or coevolution (Rittel and Webber 1973; Archea 1987).
Most models of design cognition identify two kinds of thinking, described in the “brainstorming” technique as divergent and convergent thinking (Osborn 1953). During divergent thinking (also called associative thinking), the designer is actively collecting as many disparate solutions as possible, without regard for their practicality or details. This is usually seen as the more intuitive and creative aspect of design. During convergent thinking the designer is refining the design idea and considering the ramifications. Divergent thinking is often creative “shower thinking” (Bryce 2014), while convergent thinking is sharp-pencil, “working it out at the drawing board” thinking. The premise in the literature on creativity, as well as remarks by Chuck Close, Thomas Edison, and others, is that process and practice are the primary ingredients of creative productivity.
While being creative may be something one can practice, mastering creative production, as with other expertise, requires time and experience. Experts view problems more abstractly than beginners, so one aspect of developing expertise probably consists of learning templates, frames, or categories of action that can be applied to multiple problems. Young architects are often advised and inclined to travel, in part because it provides grist for the mill of design thinking by associating their own human experience with as many examples of the built environment as possible. They are also advised to study different buildings and cultures, and to record their thoughts in sketchbooks and drawings. Done regularly, the habit builds both direct personal memory and documentary resources.
Design Moves
When considering design computing strategies to provide cognitive support to designers, we need to consider each stage of design, not just one, such as drawing. The creative practice of design has deep roots in the designer’s lived experience, and their mastery of the design process in a way that combines divergent and convergent thinking, but it also requires the designer to evaluate alternative designs and devise transformations—what Donald Schön calls “moves”—that can be applied to make the design better (Schön 1984).
Design moves may be subtle or bold. They very often nest, one inside another, as when an entire floor plan is flipped or rotated. Much of learning to design, as Schön discusses, is concerned with developing a repertoire of moves and judgment regarding their appropriate use. Constraints on available moves, through technological limits or designer inhibition, restrict design exploration. Undoing a move, backtracking, is common. Making increasingly detailed moves following a hierarchy of importance is generally thought to be efficient.
Unfortunately, the list of possible moves seems to arise from the design problem itself, making them difficult to identify or catalog in advance and giving rise to what Schön (1984) characterizes as “reflection in action.” Expert designers also apply abstractions that simplify, standardize, or routinize aspects of design, as when they size timber elements in a residence or select door hardware for an office.
The Role of Drawing in the Design Process
It is through drawing that we not only explore the possibilities of new design but also acquire the fundamental language (the “meta-language”) of architecture.
Simon Unwin (2014)
[R]elatively unstructured drawing or sketching is important in the early stages of the design process and … more structured types of drawing occur later in the process.
Purcell and Gero (1998)
Architects have a deep respect for drawings, and for the act of drawing, to the point where the physical act of drawing and the cognitive processes of design are often linked (Scheer 2014). Describing their own design processes, architects frequently speak of “talking” to a drawing, giving the act of drawing a conversational character and the drawing a role as dialogic partner. Ideas are explored by “drawing them up” using appropriate wall thicknesses, furnishing dimensions, etc. The drawing serves as an analytic platform, a means of overcoming human memory limitations, and as a means of experiential simulation. During schematic design, most drawings go through many revisions, though it is said that Frank Lloyd Wright drew the plans for his iconic Falling Water in the few hours it took the client to drive from Chicago to Taliesin, and that he talked to himself as he drew, describing how the clients would experience the house (Tafel 1979).
Drawings also have meanings beyond the explicit delineation of geometry—latent content. They capture, express, and test ideas, but within the practice they also demonstrate prowess, persuade, explain, and organize work (Robbins 1994). Drawing is part of the identity of architects, and has undergone radical transformation in the last two decades. Among the changes: Most drawings are done using computers; drawing sets contain more 3D details; senior management is less able to jump in and help with production during times of heavy work; and the infrastructure costs of supporting additional designers have grown due to required hardware, software, and training.
Drawings also have implicit meaning for contractors. In the past, they might look at a hand-drawn detail and estimate the architectural knowledge and experience of the designer who drew it based on the drawing’s lettering and line quality. The assumption was that drawing skill develops along with professional experience, so a well-drawn detail is also a well-informed detail. While CAD drawings are more uniform than hand drawings, they do manifest their author’s drawing skills and knowledge; it just takes time to learn to read them.
Finally, because clients often interpret hardline drawings as complete and unchangeable, designers may use software to add squiggle to the drawings, imitating back-of-the-envelope sketching and reminding clients that decisions are mutable.
Drawing Conventions Organize Cognitive Work
One of the significant, and unstudied, changes which digital drawing technology has brought about relates to the relationship between the designer and their visual field. Traditional drawings, on large pieces of paper, adhered to a preset scale, which establishes a visual field within which the drawing is built. Detail finer than a pencil point cannot be represented, and spatial organization larger than what can be viewed standing at the drawing board is difficult to think about.
A change in scale is both cognitive and graphic. You cannot put much detail on a 1:100 scale drawing, and a pencil-width line is probably all you need to represent wall thickness. A 5× or 10× scale change permits a change in representation detail, from single-line walls to double-line walls, to details of framing and trim. This change accompanies, or is accompanied by, a change in focus, from arrangements of space to wall locations to wall composition to assembly details.
The Role of Memory
One of the most powerful aspects of human thought is our ability to sort objects and experiences into categories and then apply our knowledge about the category to deal with the case at hand.
Earl Hunt (Hunt 2002)
Individual designers use memory to draw on personal experience in many ways. Current problems may well bring to mind solutions developed in either the recent or distant past and personally experienced or studied by the designer. Alternatively, models of use, society, and construction developed by others and encountered through study may be recalled at an opportune time, suggesting a solution to a problem. Finally, during projects designers acquire a great deal of client, site, and project information that may not be explicitly recorded, but which influences the design development.
Design firms, and their many individual designers, collaborate on projects through formal and informal memory systems. Preliminary sketches are shared. They become formal documents recording decisions, but throughout a project alternative solutions jostle to become (or be reinstated as) part of the current solution. Desktop “litter” speaks to passers-by. Past presentations and projects decorate the walls. Projects develop a narrative; memory of design rationale is important in preserving decision-making coherence over the course of a project, and, of course, memory is the foundation for learning.
Design by Analogy
We often rely on solution by analogy, finding an example in history or experience that is like the current problem and borrowing or adapting the previous solution. We might call this a problem-space search, in contrast to a solution-space search. Its prevalence in practice might account for the observation that architects, unlike scientists, often reach their full power later in life, once they’ve had time to acquire enough experience.
When you’ve solved a problem once, the “not obvious” solution of Chapter 6 becomes available as an object of memory, which means future presentations of the problem have obvious or easier solutions. In writing about the problem of problems, Kees Dorst describes five levels of expertise, corresponding to “five ways of perceiving, interpreting, structuring and solving problems.” These are novice, beginner, competent, proficient, and expert, the last of which “responds to specific situation[s] intuitively, and performs the appropriate action, straightaway” (Dorst 2003). Similarly, in his book Notes on the Synthesis of Form, Christopher Alexander (1964) distinguishes between unself-conscious design, as practiced in cultures where tradition dictates how to solve problems, and self-conscious design, in which changing conditions are so challenging that reference to history is inadequate and the designer must actually posit a new solution.
Design as Selection
While inventive creativity is important, in many ways design is about selection—selection of strategies, materials, configurations, manufactured systems and objects such as door hardware, lighting systems, appliances, and carpet patterns. Building types such as “split-level ranch-house,” “semi-detached house” or “three-story walk-up” convey a lot of information about the design, information drawn from a foundation of life experience and a pool of culturally validated examples. Without copying any particular design line for line, the designer may use such memory to organize a new response and create a new design. Memory and experience help guide selection, as does knowledge embedded in books such as Building Construction Illustrated (Ching 2014), Architectural Graphic Standards (AIA 2016), and the venerable Sweets® Catalogs (Sweets 2016), all of which provide condensed and simplified selection opportunities. In addition, most firms maintain detail libraries of tested solutions to particular configurations, and it is common for designers to review, or mine, past projects for ideas.
It is worth noting, in passing, that the increasing availability of digital fabrication technology has released (or challenged) designers to design where they might have selected in the past. This, in turn, has led to proliferation of unique buildings and objects and has both advanced and benefited from the use of sophisticated 3D modeling technology, illustrating the degree to which these topics are intertwined.
Digital support or fuel for a memory-based design process suggests one of several directions: experiencing or presenting architecture through digital media, recording personal travel in text, photos and sketches, as well as cataloging and recalling travel or design history.
Problem Solving Through Subdivision
When memory and the related strategies discussed above fail, as they often do in at least some aspect of design, designers fall back on various strategies that aim to recast recalcitrant problems in ways that are more tractable. Plato’s suggestion to “separat[e] the Idea into parts, by dividing it at the joints, as nature directs”—what we might call divide and conquer—is one of the most common.
Modern attempts to formalize design processes are rooted in time-and-motion studies on mass-production assembly lines and the field of operations research, beginning in the early twentieth century, as well as more recent investigations of artificial intelligence. The time-and-motion approach tends to view the whole as simply the sum of the parts. Once isolated, each partial process could be streamlined on its own. Industrial production in World War II demonstrated the power of the divide-and-conquer approach. By dividing problems into smaller, more malleable, parts, solving them individually, and aggregating the results, it was possible to resolve many challenges, while taking advantage of collaborative multi-member teams of average people. Individual action and ingenuity took a back seat to methodological rigidity and the study of process.
What isn’t always appreciated is that the very act of dividing the larger problem into smaller ones involves selecting one solution paradigm over other possible approaches. As Thomas Kuhn explored in his Structure of Scientific Revolutions (1962), it is possible to force-fit old solutions onto new problems right up to the point where the evidence of mis-fit becomes overwhelming and a paradigm shift occurs as a new model replaces the old.
Applied to routine problems, in the absence of overwhelming evidence demanding a paradigm shift, this approach often works adequately. It also describes the system of disciplines, consultants, and contracts that generally governs the process of design, and which is supported by most current software systems. These, in turn, correspond fairly well to the cognitive frameworks found in the IBIS and CBR approaches, as we will see shortly. There are, however, research directions that challenge this model, where the whole is more than the sum of its parts.
Developing Memories
Memories do not simply happen; they are built. The common admonition that young designers should “see the world” and sketch or take notes in their sketchbooks clearly seeks to augment their store of available memories as well as their sensitivity to cultural nuances through reflective study. Traditional paper sketchbooks permit travelers to construct a powerful blend of text, diagrams, and pictorial drawings at an almost infinite range of scales. Digital alternatives in the form of tablet or laptop computer, which preserve both sketch and text, might facilitate subsequent search or cataloging, but have thus-far proven elusive, though photo-sharing, and internet technology in general, certainly represent a giant step forward.
Personal travel, while valuable, is an expensive and time-consuming process. Historically, efforts have been made via words, painting, still photography, and film to convey experience of places and times not otherwise accessible. More recently, various attempts have been made to use internet-based technologies and sensory immersion to enable cyber tourism (Prideaux 2015). Such digital travel is not limited to the present—there are digital reconstructions of the ancient city of Pompeii as well as the Taj Mahal.
Many of these digital media recordings or reconstructions are unsatisfying. In addition to being limited to sight and maybe sound, they are often difficult to navigate without drifting through walls or floors and becoming disoriented because they lack kinesthetic feedback; in some cases they are nauseating. Head-mounted displays are notorious causes of motion sickness, but even flat-screen displays are susceptible when viewing very wide-angle images. There is a need for developers who understand how people visualize and move through the built environment to improve both the media and technology for these experiences. Numerous initiatives employing video-game engines such as Quake or Unity have made only slight progress, but rapid evolution of digital camera and display technology continues.
Nor is all this travel divided cleanly between real and virtual. The availability of location-aware interactive mobile devices such as smartphones and tablets has enabled the deployment of augmented reality systems that can “watch over your shoulder” as you travel and offer information about buildings, menus, shops, sights, history (e.g., http://etips.com). Some also allow you to post location-specific comments of your own, for subsequent travelers.
The recent explosion of personal photography has made image cataloging, tagging, and organization software widely available, and a few researchers have sought to leverage those collections to support design creativity (Nakakoji et al. 1999).
Collective Memories
Design organizations have collective memories built on the memories of individual designers present in the organization, but augmented by stories, project archives, and manuals of standard work, and supported by libraries of reference material drawn from the larger professional culture. Being outside the individual designer’s head, many of these projects utilize computers to store and retrieve information, augmenting individual memory in complex settings.
Case-Based Reasoning
Individual personal memory is fallible, and when discussing design we often care as much about human motivations and decisions related to an environment as we care about the physical product. Capturing intent, in words, can be important. According to its proponents, case-based reasoning (CBR) “means using old experiences to understand and solve new problems” as a systematic activity (Kolodner 1992). It is a problem-solving technique that uses similarities between current problems and past problems and their solutions to propose, critique, and build solutions to current problems. Past experience (cases) are recalled, compared to current circumstances, adapted as needed, evaluated through feedback and/or simulation, and repaired (learned from) for next time. Computational challenges exist at each of these steps, but substantive CBR systems have been built.
Issue-Based Information Systems
A related approach focuses on the problems or issues that emerge during design, explicitly capturing their argumentation, decomposition, and resolution. The resulting issue-based information systems (IBIS) have been used to address very dense design environments with many interacting and possibly conflicting design issues, arguments, and resolutions (McCrickard 2012). By structuring knowledge of known problems and their extant solutions, the approach attempts to organize and apply experience to new problems in a systematic way. PHIDIAS, one such system, offered an interactive graphical environment for capture, review, and resolution of issues and design rationale connected to NASA spacecraft design (McCall et al. 1994).
Pattern Language
Memory, acting through personal association, or structured through CBR or IBIS, provides individuals and groups with tools for addressing complex design situations. On a more analytic or reflective track, designers and theoreticians have sought to organize or analyze human-built environments in order to identify the “atomic” design elements of configuration. Christopher Alexander and colleagues developed the theory of what they called a pattern language by segmenting, characterizing, and categorizing the oft-used parts, or patterns of design. A single pattern, #242, might be a “front door bench” (Alexander et al. 1977). A complete design consists of an assembly of elemental patterns, organized and orchestrated to work together. In the context of design computing it is interesting to note that computer science has embraced the concept of pattern languages in interface design as well as algorithm design and programming instruction.
Space Syntax
Attempting to understand larger organizations of human-created space, researchers at University College London and Georgia Institute of Technology, have developed space syntax theories and processes (Hillier and Hanson 1984). From the SpaceSyntax.net website:
Space syntax is a science-based, human-focused approach that investigates relationships between spatial layout and a range of social, economic and environmental phenomena [including] patterns of movement, awareness and interaction; density, land use and land value; urban growth and societal differentiation; safety and crime distribution.
(Space Syntax 2015)
Understanding that buildings are both meaningful from the outside as civic or cultural products, and meaningful from the inside to the inhabitants that work in and move through the spaces enclosed by the building, the concept of space syntax seeks to understand how the texture and organization of space in human habitation interacts with social and cultural factors in predictable and analytic ways (Scoppa and Peponis 2015).
Coupled with the development and application of computer-based geographic information systems (GIS) to major urban areas world-wide, historical research, and a theoretical orientation towards spatial metrics such as street width, sight lines, walking distance, and pedestrian behavior, space syntax researchers use large data sets to support their analyses, as well as observational data from video cameras and other sensors, to advise civic and private developers on many real-world projects (e.g., http://spacesyntax.com).
Shape Grammars
Individual designers and historical eras often develop stylistic consistencies in the visual or topological character of their buildings. Thinking of these buildings as cultural utterances in a language or design, it makes sense to take a linguistic approach to understanding architecture. George Stiny and James Gips first developed and publicized their shape grammar theory in 1971 (Stiny and Gips 1971). The approach seeks to identify, isolate, and characterize sets of generative rules for producing new members of a particular oeuvre (Stiny and Gips 1971; Stiny 1980). Each rule specifies an initial condition and a replacement condition. The process of applying rules begins by searching a design representation for conditions that match the initial condition, or left-hand side (LHS) of a rule. Where multiple rule matches are found, a choice must be made as to which rule to apply. Often this is left to a human operator. Once a choice is made, the system replaces the portion of the representation matching the LHS with the replacement condition, or right-hand side (RHS), after which the rule-matching starts again. Examples of architectural grammars that have been developed include Palladio’s villas (Mitchell 1990), Queen Anne houses (Flemming 1987), and Frank Lloyd Wright’s Prairie-School Houses (Koning and Eizenberg 1981). While the process may sound automatic or even arbitrary, the generative rules of the grammar can be seen as embodying significant design knowledge, and their production can be likened to design itself (Königseder and Shea 2014, citing Knight 1998). While early grammar rules were derived by experienced researchers and expressed by hand, recent research aims to use a space syntax approach to uncover a grammar in existing work (Lee et al. 2015).
The Roles of Certainty, Ambiguity, Emergence, and Flow
In Google’s world, the world we enter when we go online, there’s little place for the fuzziness of contemplation. Ambiguity is not an opening for insight but a bug to be fixed.
Nicholas Carr (2008)
Some designers work with broad gestures and very soft leads—even charcoal or watercolor—in the early stages of design, resulting in representations that are open to substantial amounts of interpretation. While antithetical to the precision of construction documents, the resulting ambiguity can also be seen as a resource, inviting collaborative interpretation and participation in the design process, engaging clients and coworkers alike.
In contrast, the unambiguous quality of crisp laser-printed drawings has often been cited as a problem when presenting preliminary design ideas to clients—who perceive the design as more immutable than it really is. As with any interpersonal exchange, designer and client communications need to be fluid and comfortable to tease out the full depth and extent of thought. Projecting too much certainty too early can cut off discussion, leaving reservations and concerns unuttered, possibly causing friction or expensive changes later.
The problem exists at a deeper level as well. As we will see in Chapter 8, external representations are also a designer’s means of communicating with their “future self.” Most designers approve of ambiguity early in design, but recognize that it must ultimately be resolved. This might tempt us to view the design process as simply the systematic elimination of ambiguity. Digital systems might seem to fit the task admirably since they thrive on specificity, on precise taxonomies that distinguish rather than blur. However, ambiguity seems to be an important component of emergence as well—“the process of converting an implicit property into an explicit one” (Talbott 2003, quoting Gero and Yan 1994). Emergence happens when a representation invites and is subjected to reinterpretation. In Figure 7.2 the left side represents “two overlapping rectangles.” On the right, after only a slight reinterpretation, are the “two L-shaped polygons” implicit in the first shape, but moved slightly apart. While visually similar, their data representations are quite different, as we’ll see in the next chapter. Further, this isn’t the only possible emergent form, creating two problems. Can the software discover emergent alternatives to the current representation if asked, and might it possibly prompt reinterpretation by suggesting them? This turns out to be one of the sub-problems of shape grammars (interpretive flexibility is necessary in order to match grammar rules broadly). However, given the complexities of representation and the many possible reinterpretations, supporting reinterpretation and emergence is certainly going to be a challenging problem.
Figure 7.2 Emergence: A new interpretation of the representation requires new data.
Parameters and constraints can interact in complex ways. William Mitchell (2009) observed: “building up parametric object families … becomes a fundamental design activity, requiring deep architectural understanding, technical skill and considerable investment.” Indeed, an architecture firm with a contract for a series of new library buildings tried to simplify their work by creating an extensive set of parametric BIM elements during design of the first, only to find that the parametric constraints actually impeded their work on subsequent libraries and had to be removed.
The challenge of creating a family of related parametric forms is one of cognitive capacity; being able to think through the implications of multiple dynamic interactions. This is exactly the challenge of design. Complex user interfaces, no matter how conceptually elegant, can require so much cognitive overhead that there is little room left to pay attention to the design problem. Further, the attentive state of flow (Csikszentmihalyi 1991), in which a designer is able to fully engage a problem and often make significant progress, may well be as unattainable in a cognitively demanding environment as it is in a distracting one, including a complex interface or one that inhibits emergent reinterpretation. The joy of drawing, for many designers, resides in the power available through a very simple interface that requires reflection, but almost no other cognitive load. Reproducing this in a tool that also supports detailed building design at a semantic level remains one of the grand challenges.
Design as Social Action
Buildings and related environmental interventions are rarely an individual creation, but are, instead, a form of collective and collaborative human expression. They exist as cultural productions, or statements, in an evolving cultural and social context. A certain amount of ambiguity facilitates emergence of collaborative interaction, and an undisturbed interaction with the problem may allow useful reinterpretations to emerge. As a product of a shifting cultural foundation, the appropriate metrics for assessing the result may well be unknown in the beginning, and it is possible that their meaning in these contexts arises as much from the process through which they take form as from the particular dimensional and material choices of their production. Modern sociological theory sees negotiated meaning as a product of interaction.
Wicked Problems, Situated Action, and Puzzle Making
The particular solution and the problem definition fit together and emerge together from the process. This view of design as puzzle making (Archea 1987) or co-evolution (Dorst 2003) is very different from the more mechanistic divide, conquer, and aggregate view of design methods presented by Newell and Simon. On the flip-side, the wicked problems identified by Rittel and Webber can be seen to emerge from the mis-fit of problem and solution, arising from conflicts in the negotiation of meaning. Given the expense, permanence, and possibilities of most building projects, it is then unsurprising that at some level most design problems seem to be at least a little wicked.
Research into “situated action” by anthropologist Lucy Suchman (1987) and others (e.g., Heath and Luff 2000) has focused increasingly on the emergent character of action, arising less from individual cognition than from social interaction and technological affordance. While focused on the social and emergent aspects of action, this research does not entirely discount planning, which has a role in constructing the outline of the action. The ultimate relationship is likened to that of a jazz musician to the musical score.
Such research raises questions about the ability or long-term suitability of systems of pre-defined production rules to serve the emergent character of design, unless there is a mechanism for improvisation, though perhaps if the number of rules is large enough the designer can improvise within the system. The challenge for design computing is the amount of freedom needed. Pencil and paper, CAD, BIM, and a copy-machine exist on a continuum from complete freedom to complete pre-definition. At the deepest level we can ask whether systems can encode and capture “best practices” without limiting the designer to “standardized riffs.” At a shallower level, we can develop strategies to best use existing constraint models and families of parts to provide productivity without predestination.
Digital support for puzzle making necessitates creating a system that does not require a problem definition to precede the pursuit of a solution, and which is able to move easily between editing one and editing the other, or perhaps makes no distinction between the two, but considers them two aspects of one production. Wicked problems are more difficult, because they challenge the notion that it is possible to generate or identify solutions through systematic action; they fall back on action mediated on communication.
Tools Interact with Cognition
Tool Neutrality and Cognitive Side Effects
[A]rchitects’ tools aren’t neutral. They have particular affordances and biases, and they implicitly embody particular assumptions and values.
William J. Mitchell (Mitchell 2009)
There is a persistent tendency to treat technologies as interchangeable, distinct from the human cognitive processes—something we might call a theory of tool neutrality. It is implicit in the argument often made in the 1990s that CAD was “just another pencil.” Yes, 2D CAD was brought into most offices as a one-to-one replacement for hand-drafting and was thought to require largely the same knowledge and skills and produce the same results. However, studies of technology projects in other environments have demonstrated the complex interaction of digital technology with human cognition and social interaction patterns in many work settings (Lakoff and Johnson 1980).
The magnitude of the change wrought by the introduction of IT into architecture offices came home to me a decade ago when visiting a medium-sized local firm. Their monthly IT bill had just passed their monthly rent—in an industry where drafters used to bring their own tools and needed little more than a door-on-sawhorses drawing board. In an industry where status is frequently linked to drawing prowess and where senior office staff could “jump on the boards” and draft during a deadline push, CAD tools have redistributed power and introduced new paradigms to production, such as “pair work” (Hancock 2013). Whether these changes are simply transitional or persistent remains an open question.
Infinite Zoom Variation
Traditional drawings were produced at only a handful of different scale factors. While final plots may still honor these conventions, the need to fit small and large parts of projects onto a relatively small computer screen has made precise scale factors a thing of the past during editing. As a side-effect, it isn’t uncommon for students to produce dramatically out-of-scale elements during early design as they learn to adjust their visual estimates to the drawing.
The ability to zoom in, combined with the persistence of data, offers the designer a seductive escape from dealing with big-picture problems first. When stymied by a problem at one scale, they can feel productive by zooming in to a smaller part of the project and working on it in the meantime. Where conventional scale-changes might enforce attention to large-scale issues early in the design, the CAD designer needs to develop an internal discipline about problem decomposition and priority.
The Decline of Hand Drawing: Side-Effects
Designers generally work from the outside in—e.g., establishing organizing principles on large-scale drawings before allocating uses and configuring individual spaces or picking door hardware. In the past this sequence was encouraged because a change-of-scale required manually creating a new drawing. The cost of creating that somewhat redundant representation encouraged the designer to make the changes big enough to matter. Since scale changes also enable new information to be seen or shown, creating the new drawing also provided an opportunity for the designer to recapitulate, and check, the design logic of the smaller-scale drawing. Such recapitulation is lost in the CAD world.
More fundamentally than we think, the act of drawing, not just the production of drawings, may be important to design thinking. James Cutler, a prominent Pacific Northwest architect, came to us a few years ago looking for help with a research idea. He has run a successful award-winning office for many years, but in recent years had observed that his interns are less able to execute requests requiring that they independently develop the initial sketch of a design idea. He had discussed his observation with senior management at a number of offices across the country and found anecdotal consensus on the observation. His own conclusion is that modern students are not sketching enough by hand.
In a culture with a deep belief in the efficacy of digital technology, this concern nonetheless finds support in several places. First there is fMRI evidence that it matters whether a child learns to hand-write text rather than type it. Something about the neural activation of forming the letters by hand facilitates development of writing skill (Richards et al. 2009). Second, recent studies at the university level, based on the educational psychology theory of desirable difficulty (Bjork 1994), have found that students taking longhand notes learn better than those taking notes on a laptop (Mueller and Oppenheimer 2014). Finally, within the realm of architecture and design, research has shown that “conventional modeling methods … satisfy the requirements for intermittent divergence, while parametric modeling methods … undermine them” (Talbott 2003).
These observations suggest that rather than being neutral one-for-one replacements for traditional tools, digital tools are different in important ways that affect the cognition of their users, and that the most appropriate tool for an expert may well be different from that which best suits a neophyte.
Summary
Unsurprisingly, human cognition is complex. Design cognition is especially complicated by the existence of several quixotic phenomena, including emergence, creativity, and flow. Design process and design knowledge both contribute to solving problems, capturing and supporting process, and building personal and institutional knowledge. Constructing software systems that learn from users, accrue knowledge, facilitate browsing of design rationale and alternative design options, is a substantial challenge. Creating systems that do this without interrupting flow with complex interfaces and cognitively demanding interactions will be even harder. Doing it in a way that supports life-long learning and development will require skills from design, psychology, sociology, and human–computer interaction.
Suggested Reading
Archea, John. 1987. Puzzle-making: What architects do when no one is looking, in Computability of Design. Edited by Y. Kalay, 37–52. Hoboken, NJ: John Wiley.
Cross, Nigel. 2011. Design thinking: Understanding how designers think and work. Oxford: Berg.
Lakoff, George, and Mark Johnson. 1980. Metaphors we live by. Chicago, IL: University of Chicago Press.
Lawson, Bryan. 1997. How designers think: The design process demystified (3rd edition). Oxford: Architectural Press.
Norman, Don. 1988. The design of everyday things. Cambridge, MA: MIT Press.
Suchman, Lucy. 1987. Plans and situated actions: The problem of human–machine communication. Cambridge: Cambridge University Press.
References
Akin, Ömer, Bharat Dave, and Shakunthala Pithavadian. 1987. Problem structuring in architectural design. Research Publications, Department of Architecture, Carnegie Mellon University.
Alexander, Christopher. 1964. Notes on the synthesis of form. Cambridge, MA: Harvard University Press.
Alexander, Christopher, S. Ishikawa, M. Silverstein, M. Jacobson, I. Fiksdahl-King, and S. Angel. 1977. A pattern language. New York: Oxford University Press.
American Institute of Architects. 2016. Architectural graphic standards (12th edition). Hoboken, NJ: John Wiley.
Archea, John. 1987. Puzzle-making: What architects do when no one is looking, in Computability of Design. Edited by Y. Kalay, 37–52. Hoboken, NJ: John Wiley.
Bjork, R. A. 1994. Memory and metamemory considerations in the training of human beings, in Metacognition: Knowing about knowing. Edited by J. Metcalfe and A. Shimatnura, 185–205. Cambridge, MA: MIT Press.
Bryce, Nessa Victoria. 2014. The aha! moment. Scientific American Mind 25: 36–41.
Carr, Nicholas. 2008. Is Google making us stupid? The Atlantic (July/August).
Ching, Francis D. K. 2014. Building construction illustrated (5th edition). Hoboken, NJ: John Wiley.
Close, Chuck, Kirk Varnedoe, and Deborah Wye. 2002. Chuck Close. New York, NY: The Museum of Modern Art.
Cross, Nigel. 2011. Design thinking: Understanding how designers think and work. Oxford: Berg.
Csikszentmihalyi, M. 1991. Flow: the psychology of optimal experience. New York: Harper Collins.
Dorst, Kees. 2003. The problem of design problems. Expertise in Design: Proceedings of design thinking research symposium 6. www.creativityandcognition.com/cc_conferences/cc03Design/papers/23DorstDTRS6.pdf.
Dorst, Kees. 2011. The core of “design thinking” and its application. Design Studies 32: 521–532.
Flemming, U. 1987. More than the sum of parts: The grammar of Queen Anne houses. Environment and Planning B: Planning and Design 14: 323–350.
Gero, J. S. 1990. Design prototypes: A knowledge representation schema for design. AI Magazine 11: 26–36.
Gero, John (ed.). 2011. Design computing and cognition ’10. New York, NY: Springer Science and Business Media.
Gero, J. S. and M. Yan. 1994. Shape emergence by symbolic reasoning. Environment and Planning B: Planning and Design 21: 191–212.
Hancock, Lillian. 2013. Visualizing identity: Perspectives on the influences of digital representation in architectural practice and education. Unpublished Master’s thesis, University of Washington.
Heath, C. and P. Luff. 2000. Technology in action. New York, NY: Cambridge University Press.
Hillier, Bill and J. Hanson. 1984. The social logic of space. New York, NY: Cambridge University Press.
Hunt, Earl. 2002. Thoughts on thought. Mahwah, NJ: Lawrence Erlbaum Press.
Kalay, Yehuda, L. Swerdloff, and B. Majkowski. 1990. Process and knowledge in design computation. Journal of Architectural Education 43: 47–53.
Knight, T.W. 1998. Designing a shape grammar: Problems of predictability, in Artificial Intelligence in Design ’98. Dordrecht: Kluwer
Kolodner, Janet. 1992. An introduction to case-based reasoning. Artificial Intelligence Review 6: 3–34.
Königseder, Corinna, and Kristina Shea. 2014. The making of generative design grammars, in Computational making workshop: Design computing and cognition 2014 (DCC’14).
Koning, H. and J. Eizenberg. 1981. The language of the prairie: Frank Lloyd Wright’s prairie houses. Environment and Planning B: Planning and Design 8: 295–323.
Kounios, J. and Mark Beeman. 2009. The aha! moment: The cognitive neuroscience of insight. Current Directions in Psychological Science 18: 210–216.
Kuhn, Thomas. 1962. The structure of scientific revolutions. Chicago, IL: University of Chicago Press.
Lakoff, George, and Mark Johnson. 1980. Metaphors we live by. Chicago, IL: University of Chicago Press.
Lawson, Bryan. 1997. How designers think: The design process demystified (3rd edition). Oxford: Architectural Press.
Lee, Ju Hyun, Michael J. Ostwald, and Ning Gu. 2015. A syntactical and grammatical approach to architectural configuration, analysis and generation. Architectural Science Review 58: 189–204.
McCall, Ray, P. Bennett, and E. Johnson. 1994. An overview of the PHIDIAS II HyperCAD system, in Reconnecting: ACADIA ’94 (Proceedings of the 1994 conference of the association for computer aided design in architecture). Edited by A. Harfman, 63–74.
McCrickard, D. Scott. 2012. Making claims: Knowledge design, capture, and sharing in HCI, in Synthesis lectures on human-centered informatics. San Rafael, CA: Morgan & Claypool.
Mitchell, W. J. 1990. The logic of architecture. London: MIT Press.
Mitchell, W. J. 2009. Thinking in BIM, in Architectural transformations via BIM, 10–13. Tokyo: A+U Publishing.
Mueller, Pam A. and Daniel M. Oppenheimer. 2014. The pen is mightier than the keyboard: Advantages of longhand over laptop note taking. Psychological Science 25: 1159–1168.
Nakakoji, Kumiyo, Yasuhiro Yamamoto, and Masao Ohira. 1999. A framework that supports collective creativity in design using visual images, in Creativity and cognition 1999, 166–173. New York, NY: ACM Press.
Newell, Allen, J.C. Shaw, and H.A. Simon. 1958. Elements of a theory of human problem solving. Psychological Review 65: 153–166.
Norman, Don. 1988. The design of everyday things. Cambridge, MA: MIT Press.
Osborn, A.F. 1953. Applied imagination: Principles and procedures of creative thinking. New York, NY: Charles Scribner’s Sons.
Prideaux, Bruce. 2015. Cyber-tourism: A new form of tourism experience. Tourism Recreation Research 30: 5–6.
Purcell, A. T. and J. S. Gero. 1998. Drawings and the design process. Design Studies 19: 389–430.
Richards, Todd, V. Berninger, P. Stock, L. Altemeier, P. Trivedi, and K. Maravilla. 2009. fMRI sequential-finger movement activation differentiating good and poor writers. Journal of Clinical and Experimental Neuropsychology 29: 1–17.
Rittel, Horst and Melvin M. Webber. 1973. Dilemmas in a general theory of planning. Policy Sciences 4: 155–169.
Robbins, Edward. 1994. Why architects draw. Cambridge, MA: MIT Press.
Scheer, David Ross. 2014. The death of drawing: Architecture in the age of simulation. New York, NY: Routledge.
Schön, Donald. 1984. The reflective practitioner: How professionals think in action. New York, NY: Basic Books.
Scoppa, M. D. and J. Peponis. 2015. Distributed attraction: The effects of street network connectivity upon the distribution of retail frontage in the City of Buenos Aires. Environment and Planning B: Planning and Design 42: 354–378.
Space Syntax. 2015. Space Syntax Network. www.spacesyntax.net.
Stiny, G. 1980. Introduction to shape and shape grammars. Environment and Planning B: Planning and Design 7: 343–351.
Stiny, G. and J. Gips. 1972. Shape grammars and the generative specification of painting and sculpture. Information Processing ’71. Edited by C.V. Freiman, 1460–1465. Amsterdam: North-Holland Publishing Company.
Suchman, Lucy. 1987. Plans and situated actions: The problem of human–machine communication. Cambridge: Cambridge University Press, Construction building materials directory. http://sweets.construction.com
Tafel, Edgar. 1979. Years with Frank Lloyd Wright: Apprentice to genius. New York: Dover.
Talbott, Kyle. 2003. Divergent thinking in the construction of architectural models. IJAC 2: 263–286.
Unwin, Simon. 2014. Analysing architecture (4th edition). London: Routledge.
