The Anatomy of Prototypes: Prototypes as Filters,

Prototypes as Manifestations of Design Ideas

YOUN-KYUNG LIM, ERIK STOLTERMAN
Indiana University, Bloomington
and
JOSH TENENBERG
University of Washington, Tacoma
________________________________________________________________________

The role of prototypes is well established in the field of HCI and Design. A lack of knowledge, however, about
the fundamental nature of prototypes still exists. Researchers have attempted to identify different types of
prototypes, such as low- vs. high-fidelity prototypes, but these attempts have centered on evaluation rather than
support of design exploration. There have also been efforts to provide new ways of thinking about the activity
of using prototypes, such as experience prototyping and paper prototyping, but these efforts do not provide a
discourse for understanding fundamental characteristics of prototypes. In this article, we propose an anatomy of
prototypes as a framework for prototype conceptualization. We view prototypes not only in their role in
evaluation but also in their generative role in enabling designers to reflect on their design activities in exploring
a design space. We base this framework on the findings of two case studies that reveal two key dimensions:
prototypes as filters and prototypes as manifestations. We explain why these two dimensions are important and
how this conceptual framework can benefit our field by establishing more solid and systematic knowledge
about prototypes and prototyping.

Categories and Subject Descriptors: H.5.2 [Information Interfaces and Presentation]: User Interfaces; H.1.2
[Models and Principles]: User/Machine Systems
General Terms: Design, Theory
Additional Key Words and Phrases: Prototype, prototyping, design, design space, human-computer interaction
________________________________________________________________________

1. INTRODUCTION
The fields of human-computer interaction (HCI), software engineering, and design
commonly use the term prototype to signify a specific kind of object used in the design
process. The necessity of prototypes in these areas is obvious and unquestionable. Over
the years, researchers and practitioners in HCI have proposed numerous prototyping
techniques; these efforts primarily view prototypes as tools for evaluation of design
failure or success, as evidenced in a recent panel session at one of the most prestigious
HCI conferences, “‘Get Real!’ What’s Wrong with HCI Prototyping And How Can We
Fix It?” [Jones et al. 2007]. A close examination of actual design practices in which

Authors' addresses: Youn-kyung Lim (contact author – the affiliation changed to KAIST, Department of
Industrial Design) - 335 Gwahangno (373-1 Guseong-dong), Yuseong-gu, Daejeon 305-701, Republic of Korea;
Erik Stolterman’s - 901 E. 10th St. Bloomington, IN 47408, USA; Josh Tenenberg’s - Campus Box 358426, 1900
Commerce St, Tacoma WA 98402-3100; emails: younlim@gmail.com or younlim@kaist.ac.kr;
estolter@indiana.edu; jtenenbg@u.washington.edu
Permission to make digital/hard copy of part of this work for personal or classroom use is granted without fee
provided that the copies are not made or distributed for profit or commercial advantage, the copyright notice,
the title of the publication, and its date of appear, and notice is given that copying is by permission of the
ACM, Inc. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific
permission and/or a fee.
© 2001 ACM 1073-0516/01/0300-0034 $5.00
prototypes are pervasively used, however, shows that prototypes as a means for formal
evaluation (such as usability testing) are a relatively small part of the entire design
process. Prototypes are the means by which designers organically and evolutionarily
learn, discover, generate, and refine designs. They are design-thinking enablers deeply
embedded and immersed in design practice and not just tools for evaluating or proving
successes or failures of design outcomes. Buxton [2007] advocates such a view,
promoting the notion of “sketching” throughout the whole design process.
In this paper, we introduce a new way of thinking about prototypes and prototyping
based on the need for exploring and establishing a fundamental definition of prototypes
that extends current understanding and highlights critical roles. With this attempt, we
conceptualize prototypes as tools for traversing a design space where all possible design
alternatives and their rationales can be explored [Goel and Pirolli 1992; Moran and Caroll
1996]. Designers communicate the rationales of their design decisions through prototypes.
Prototypes stimulate reflections, and designers use them to frame, refine, and discover
possibilities in a design space. This view differs markedly from current approaches in
software engineering contexts where engineers use prototypes to identify and satisfy
requirements [Floyd 1984]. These requirement-oriented approaches have their limitations,
especially since design activities are flexible rather than rigid, reflective rather than
prescriptive, and problem-setting rather than problem-solving [Schön 1982]. A design
idea that satisfies all the identified requirements does not guarantee that it is the best
design since a number of ways can meet each requirement. If the focus of prototyping is
framing and exploring a design space, what matters is not identifying or satisfying
requirements using prototypes but finding the manifestation that in its simplest form,
filters the qualities in which designers are interested, without distorting the
understanding of the whole. We call this the fundamental prototyping principle.1
In order to support this perspective and to provide a stable foundation for the study of
prototypes in HCI, we propose a framework for conceptualizing prototypes; we see such
a framework as an anatomy of prototypes. The framework is an attempt to create an
understanding of the nature of prototypes in general and to provide a language for
articulating the characteristics of a particular prototype. Such a framework will enable
designers to specify more effectively the goals and questions to explore when planning
and making their prototypes. It will also better guide designers in thinking critically about
their approach to prototyping.

1
The discussion of the benefits of applying this principle resonates with the new way of thinking about
prototyping in HCI illustrated in [Wong 1992].
 Two fundamental aspects of prototypes form the basis of our framework:
1) prototypes are for traversing a design space, leading to the creation of
meaningful knowledge about the final design as envisioned in the process of
design, and
2) prototypes are purposefully formed manifestations of design ideas.

When exploring a certain aspect of a design idea, designers can focus on

demonstrating various ideas for interaction techniques without determining other
qualities of the design, such as its appearance or its functionality. When exploring only
the design’s form aspect in evaluating portability-related ergonomics, they may develop
various prototypes with different sizes, weights, and shapes without any interactivity or
functionality in place.
As a part of our framework, we identify an initial set of design aspects that a
prototype might exhibit. We call these aspects filtering dimensions. We use the term,
filter, since by selecting aspects of a design idea, the designer focuses on particular
regions within an imagined or possible design space. The designer screens out
unnecessary aspects of the design that a particular prototype does not need to explore.
Designers may purposefully do this so that they can extract knowledge about specific
aspects of the design more precisely and effectively. The decision of what to filter out is
always based on the purpose of prototyping.
When creating a prototype that manifests a certain aspect of a design idea, designers
need to make careful choices about the prototype’s material, the resolution of its details
(which corresponds to the concept of fidelity), and the scope of what the prototype covers
(which can be understood as a level of inclusiveness—i.e. whether the prototype covers
only one aspect of the design idea or several aspects of the design idea). These three
considerations of manifesting a design idea—namely, the material, resolution, and scope
of a prototype—are also part of the prototype’s anatomy. We call these considerations
manifestation dimensions.
A designer can determine the manifestation dimensions of a prototype by considering
the economic principle of prototyping, which we define as follows: the best prototype is
one that, in the simplest and most efficient way, makes the possibilities and limitations of
a design idea visible and measurable. If we keep the economic principle of prototyping
in mind, determining the values of the manifestation dimensions—i.e. the materials,
resolution, and scope of the prototype—can be approached in a rational and systematic
way. Based on this conception of an anatomy of prototypes, we view prototypes as filters
intended to traverse and sift through a design space and as manifestations of design ideas
that concretize and externalize conceptual ideas. Table I summarizes the core proposal of
our definition of the anatomy of prototypes and its key principles.

Table I. The Principles of prototyping and the anatomy of prototypes

Fundamental prototyping principle:
Prototyping is an activity with the purpose of creating a manifestation that, in its
simplest form, filters the qualities in which designers are interested, without distorting
the understanding of the whole.

Economic principle of prototyping:

The best prototype is one that, in the simplest and the most efficient way, makes the
possibilities and limitations of a design idea visible and measurable.

Anatomy of prototypes:
Prototypes are filters that traverse a design space and are manifestations of design ideas
that concretize and externalize conceptual ideas.

In this paper, we first examine current understandings of prototypes in the field of

HCI. We discuss what we mean by prototypes as filters and manifestations. Then we
introduce the details of our concept, the anatomy of prototypes. We explore two
prototyping case studies that serve as our sources in identifying the nature of prototypes
and demonstrate how the identified dimensions can help in generating, conceptualizing,
and comparing prototypes. We end by discussing the benefits and potentials of our
proposal for research and design practice in the fields of HCI and Design.

2. HOW PROTOTYPING IS UNDERSTOOD IN CURRENT RESEARCH IN HCI

In HCI, many researchers and practitioners have developed their own ways of
prototyping for their various purposes. Discussions on prototyping have primarily
focused on the issue of the prototype’s fidelity, largely because fidelity is a matter of cost.
Some have therefore emphasized the benefits of using low-fidelity prototyping
techniques. These techniques include paper prototyping for all types of interactive
products, including computer-based applications, mobile devices, and websites [Grady
2000; Rettig 1994; Snyder 2003]; Switcharoo for physical interactive products [Avrahami
and Hudson 2002]; Calder for physical interfaces [Greenberg and Boyle 2002; Lee, et al.
2004]; Buck prototyping for mobile devices [Pering 2002]; rapid prototyping for mobile
devices using augmented reality technology [Nam and Lee 2003]; and DART for
augmented reality systems [MacIntyre et al. 2004].
Nonetheless, low-fidelity prototyping has brought another round of discussion,
focused on the validity of prototyping [Cockton and Woolrych 2002; Convertino, et al.
2004]. There have been discussions on the validity of less exhaustive usability methods in
terms of the number of users to test [Spool and Schroeder 2001], the length of
observation, in situ versus lab tests, and user profiles. Particularly in the case of in situ
tests, the fidelity of prototypes deeply matters because researchers cannot, in most cases,
conduct the tests in the actual situation as the prototype is not fully functional or is not
very similar to the final product [Reichl, et al. 2007]. Most low-fidelity prototyping
examples focus primarily on design exploration and communication and less on formal
design evaluation.
Although the notion of a prototype’s fidelity is helpful for orienting designers in the
ways of building prototypes, some research results, including our own research, show
that the simple distinction of low- versus high-fidelity prototypes can sometimes be
problematic [Lim et al. 2006; McCurdy et al. 2006]. For example, McCurdy et al. [2006]
suggest that such a binary distinction should be reexamined. They demonstrate the
effectiveness of more sophisticated prototyping, namely a “mixed-fidelity” approach—
i.e., a prototype that combines low-fidelity and high-fidelity on different dimensions of
design consideration. Lim et al. [2006] show that not only the fidelity but also other
contextual factors involved in prototyping, such as the materials of prototypes and testing
conditions, affect the results of prototyping.
Prototyping for externalizing and representing design ideas is another pervasive
technique in designing interactive artifacts. Designers commonly use sketching as a
means to externalize concepts [Buxton 2007]. Many researchers have explored
developing tools for creating interactive prototypes that utilize a sketching technique.
Examples of sketch-based prototyping have existed since the SILK tool by Landay and
Myers [Landay 1996]. SILK uses a tablet based input device to essentially “sketch” an
interface. The program allows for dynamic interaction corresponding to the rough button
and form field shapes drawn by the designer. DENIM is another example of a sketch-
based prototyping environment, following in the footsteps of SILK [Lin, et al. 2000].
DENIM is used to prototype entire websites and allows for an intuitive sketching and
linking scheme to lay out individual web pages. Sketch-based prototyping remains a
popular topic for research, and a number of recent studies and tools have extended the
sketch motif in prototyping—such as DEMAIS, a multimedia sketch-based editor [Bailey
et al. 2001], and DART, a rapid prototyping environment for augmented reality
environments [MacIntyre, et al. 2004].
Participatory design is another popular approach in HCI, and this approach also
utilizes various prototyping techniques. Many of the participatory approaches are used for
understanding user needs and for exploring design ideas. (Some of the representative
techniques include CARD [Muller 2001], game-based design [Brandt and Messeter 2004],
and a role-playing approach [Svanaes and Seland 2004].) In participatory design, the use
of prototypes focuses on actively engaging users in creating and exploring design ideas.
Because the users are not expert designers, the results from participatory design
approaches usually need to be reinterpreted to understand users’ needs and values rather
than directly adapting their design ideas into the final design.
The examples named here are only a few of the many uses and styles of prototyping
in interaction design. In each technique, the prototype that is created filters different
aspects of the design ideas, though none of these techniques solve every aspect of a
design. We argue that these techniques are in some cases used without a reflective
understanding of how they differ from each other in terms of their roles and
characteristics. Some researchers have tried to compare the pros and cons of different
techniques [Avrahami and Hudson 2002; Gutierrez 1989; Liu and Khooshabeh 2003;
Pering 2002; Rudd, et al. 1996; Sefelin, et al. 2003; Thompson and Wishbow 1992; Virzi,
et al. 1996; Walker, et al. 2002], and this represents a first step in understanding how
each style of prototype functions differently. Most of those comparisons, however, are
based on anecdotal experiences rather than empirical studies.
Of course, some of the examples were more rigorously conducted, including [Liu and
Khooshabeh 2003; Sefelin, et al. 2003; Virzi, et al. 1996; Walker, et al. 2002]. Sefelin et
al. [2003] examine if users’ willingness to criticize or make suggestions about a design
differs when using paper-based or computer-based low-fidelity prototyping. Virzi et al.
[1996] claim that low- and high-fidelity prototypes are equally suitable for finding
usability problems. The systems that they use for the evaluation, however, are standard
GUI-based ones, which differ from mobile or ubiquitous computing systems; they also do
not clarify what types of problems were identified by which type of prototyping
technique. Liu and Khooshabeh, who study prototyping techniques for ubiquitous
computing environments, claim that it is critical to choose carefully the fidelity and
automation level of the evaluated prototypes [Liu and Khooshabeh 2003].
 We are primarily concerned with the lack of a fundamental definition of prototypes in
the different ways of using and defining prototypes that many researchers and
practitioners propose. We appreciate some researchers’ attempts to summarize
taxonomies of prototypes based on their different uses in design or development
processes. Lichter et al. [1993] identify four types of prototypes within the context of the
software development process. The first type is the presentation prototype, which
presents aspects of design ideas in order to facilitate communication between a client and
a software manufacturer. The second type is the prototype proper, which describes
certain aspects of design ideas in order to understand and discover problems within those
ideas. The third type is the breadboard, which quickly evaluates “construction-related
questions” within the development team. The fourth type is the pilot system, which
closely resembles the actual application for final refinements. Gutierrez [1989] also
suggests various forms of prototyping that derive from existing examples relevant to
software development activities; they include game playing, exploratory prototyping,
system simulation, scenario-based design, experimental prototyping, production
prototyping, and pilot systems.
All of these existing attempts to define a taxonomy of prototypes are primarily based
on different ways of using prototypes in a development and design process. Developing
generally applicable prototyping methods does not seem viable in face of the complex
variety of interactive artifacts in HCI design. Current prototyping research can best be
described as an ongoing attempt to come up with what to do with prototypes without
understanding what they actually are. Although these attempts eventually enable us to
understand indirectly what prototypes are, we will not be able to establish a fundamental
definition of prototypes that is sophisticated enough to characterize their complex and
dynamic nature if we continue to research only this direction. Although the different
ways of using prototypes need to continue to be explored and practiced, we see a strong
need for a fundamental knowledge about what prototypes are in order to be able to
further advance knowledge and research about prototyping. We believe that the search
for fundamental knowledge about prototypes will not only help researchers and
practitioners become more creative and effective in determining what we can do with
prototypes in design but will also establish a coherent understanding of the different
techniques and approaches of existing and forthcoming examples of prototyping. In
addition, this knowledge will support and inform designers and researchers in their
development of new prototyping techniques.
 The lack of a fundamental understanding of prototypes is what motivates our attempt
to define an anatomy of prototypes. Instead of focusing on the wide variety of purposes
and processes in which prototypes are used, we want to define prototypes of any type in a
systematic and careful manner. Without conscious awareness of how prototypes
influence the way users may interpret them during testing or how designers use them to
identify problems, refine designs, and generate more ideas, the results of using prototypes
can lead to undesirable effects. We propose the idea of an anatomy of prototypes
accommodating two key aspects of prototypes, namely prototypes as filters and
prototypes as manifestations of design ideas.

3. PROTOTYPES AS FILTERS
How, then, do prototypes help designers traverse design spaces? A primary strength of a
prototype is in its incompleteness. It is the incompleteness that makes it possible to
examine an idea’s qualities without building a copy of the final design. Prototypes are
helpful as much in what they do not include as in what they do. For example, a two-
dimensional prototype of a three-dimensional building can help us to determine the
spatial relationship of the rooms, without placing any constraints on the materials used
for walls and floors. This incompleteness structures the designer’s traversal of a design
space by allowing decisions along certain dimensions (appearances of walls and floors) to
be deferred until decisions along other dimensions (spatial relationship of rooms) have
already been made.
This characteristic of a prototype—being an incomplete portrayal of a design idea—is
the reason behind our metaphorical description of prototypes as filters. We view
prototypes as a means for design, and in this sense, our notion of filters is very different
from the notion of filtering out nuisance variables in scientific experiments. In design and
development processes, prototypes are used not for proving solutions but for discovering
problems or for exploring new solution directions. Even though they can serve other
purposes, prototypes in this context are a means of generative and evaluative discovery.
When incomplete, a prototype reveals certain aspects of a design idea—i.e., it filters
certain qualities. For example, let us assume that a designer needs to evaluate her ideas
about the ergonomics of one-thumb interactions with a mobile device. She may make
various three-dimensional forms of the mobile device to figure out which ideas work
better. When testing her ideas with three-dimensional form prototypes, she not only
evaluates which ideas work better than the others, but also, more importantly, she
discovers what factors of the forms make the ergonomics better, leading her to generate
more or new design ideas. Those three-dimensional prototypes open up a new design
space to explore—a space that may offer possibilities and better choices of the forms of
the mobile device that are more effective ergonomically. The competence involved in
prototyping is therefore the skill of designing a prototype so that it filters the qualities of
interest to the designer. In other words, the most efficient prototype is the most
incomplete one that still filters the qualities the designer wants to examine and explore.
Fig. 1 shows an example of showing different possible prototypes representing different
qualities of interest that can be filtered out through each of them when exploring the
design of a digital camcoder.

Fully working 3D form with a Screen-based 3D form with

product hand strap viewfinder and partially working
interface panel breadboard

Examining the Examining the Examining the

ergonomic input-feedback input layout
quality relationship quality quality

Fig. 1. A series of prototypes that represent different qualities of interest to a designer to filter out
different aspects of a design [Lim, 2003]
Normally, a design space is extremely large and complex; it is not feasible to explore
the whole space at one time. One of the most difficult challenges of design is that we
cannot control all possible effects of the design we produce. Prototypes are a tangible
attempt to view a design’s future impact so that we can predict and evaluate certain
effects before we unleash it on the world. Knowing that prototypes filter certain aspects
of a design, we can become more aware of the complexity and responsibility of a design,
and hence be more thoughtful about our design decision-making.
Prototypes are intricately intertwined with the evolution of design ideas throughout
the design process. We constantly evaluate and reflect on the values of what we design—
if those designs are socially responsible, economically viable, experientially pleasing,
culturally sound, operationally usable, technologically compatible, and functionally error-
free. These are some of the important values that designers try to satisfy. Throughout the
design process, prototypes are what manifest the design thinking process to reach such
design outcomes.

4. PROTOTYPES AS MANIFESTATIONS OF DESIGN IDEAS

It is widely accepted that design is a continuous coupling of internal mental activities and
external realization activities. Recent research in education and cognition indicates that
designs are constituted through iterated interaction with external design manifestations.
Within the domain of engineering, Adams [2002] reports, “iteration is a significant
component of design activity that occurs frequently throughout the design process; and
measures of iterative activity were significant indicators of design success ... and greater
engineering experience.” Recent cognitive research informs this view by advancing the
notion of the extended mind: a view of the mind that extends beyond the confines of the
individual brain to include external artifacts. Andy Clark points out the commonsensical
bias we have toward viewing the mind (and cognition) as a purely internal affair: “we are
in the grip of a simple prejudice: the prejudice that whatever matters about MY mind
must depend solely on what goes on inside my own biological skin-bag, inside the
ancient fortress of skin and skull. But this fortress was meant to be breached” [Clark
2001].
Clark describes an empirical study by Van Leeuwen, Vertijnen, and Hekkert [2001]
on the interaction between artist and artifact in the act of creation. “The sketch pad is not
just a convenience for the artist, not simply a kind of external memory or durable medium
for the storage of particular ideas. Instead, the iterated process of externalizing and re-
perceiving is integral to the process of artistic cognition itself” [Clark 2001, p.19]. What
Clark suggests is that externalization of thought gives rise to new perceptual and
cognitive operations that allow for reflection, critique, and iteration. That is, the act of
bringing thoughts into material form is not incidental to the act of creation but is itself
constitutive of and essential to creation. Mind, then, is not simply the sum total of
representations and processes within the brain but also includes external manifestations
of thought. Donald Schön famously captures this perspective when he states that we have
to externalize our ideas so that the “world can speak back to us.” The realized idea
becomes a discussant, a collaborator, helping us to understand and examine our own
ideas [Schön 1987]. Therefore, when a designer creates and envisions an idea, she
necessarily develops the idea by moving it out into the world. She performs this
transformation and externalization by realizing the idea in some kind of “physical”
manifestation [Lim 2003; Tyszberowicz and Yehudai 1992; Zucconi's et al. 1990].
 These manifestations can take almost any form, shape, and appearance, based on the
choice of material. The simplest form, the rough sketch on a piece of paper, is as
important to the designer as it is to the abstract artist. Even simple configurations of
images and text can serve an important design purpose. Looking at our own or a
colleague’s sketch, we can get a sense of eventual possibilities or limitations inherent in
the idea. As an idea evolves and is refined, the need for more complex prototypes or
manifestations increases.
The characteristic of prototypes as manifestations of design ideas is the same in all
design fields, but it is especially interesting and important within Human-Computer
Interaction (HCI) design. One reason is that the material used in the field—digital
material—is of a different kind, a “material without qualities” [Löwgren and Stolterman
2004]. As they can take almost any shape or form, digital materials have very few
intrinsic “material” limitations. Physical materials—such as wood, concrete, or steel—all
have limitations and distinct properties that limit us in the choice of the desired form and
function of a design. Working with the design of a digital artifact means that the material
qualities determine form and function to a lesser degree, and that the design space
therefore is larger and less restricted. We argue that the choice of filters is almost infinite
in interaction design since the design space is itself infinite and not limited in the same
sense as in other design areas.
Due to the greater possibilities inherent in digital material, the choices in prototyping
are even more open-ended. The designer may use very different materials in prototyping
than those in the final target product, especially when she needs to select the most
efficient and cost-effective choices to manifest design ideas. For example, designers can
use paper prototypes to approximate screen-based web designs. The material chosen for a
prototype has direct implications on users’ perceptions when it is used for evaluating a
design concept, (e.g., as in [Lim et al. 2006]). All these material issues lead to an even
greater problem in deciding what prototypes to build and use and for what purposes.
In the definition of the anatomy of prototypes, we incorporate several issues in the
manifestation of ideas, including the implications of the disparity between prototype
materials and the expected real materials of a final design outcome; the dissimilarities
between the manifested details of design ideas with prototypes and the details of the
actual final design—i.e., issues related to the level of resolution; and the differences
between what a prototype covers and what the final design actually contains—i.e., issues
of the level of scope.
5. ANATOMY OF PROTOTYPES
We argue that the purpose of designing a prototype is to find the manifestation that, in its
simplest form, will filter the qualities in which the designer is interested without
distorting the understanding of the whole. We call this the fundamental prototyping
principle. This principle serves as the foundation of our attempt to develop an anatomy of
prototypes. It embeds two notions about prototypes, namely prototypes as filters and
prototypes as manifestations of design ideas. In this section, we propose a beginning
definition and an outline of an anatomy of prototypes. But, before doing that, we need to
identify the difference between the meaning of prototype and prototyping. Prototypes are
representative and manifested forms of design ideas. Prototyping is the activity of
making and utilizing prototypes in design. Current research has primarily focused on the
different types of prototyping without any rigorous analysis of what prototypes are,
except in the notion of a prototype’s fidelity, as we discuss earlier. For the purpose of this
paper, it is important to understand prototypes and prototyping as two separate objects of
study.
Anatomy is commonly defined as the “the science of bodily structure” [anatomy
2006]. We use this notion both literally and metaphorically to sketch an anatomy of
prototypes, to “dissect” or uncover the fundamental dimensions along which to
understand any particular prototype. We use the notion of anatomy descriptively rather
than prescriptively. An anatomy is a description of possible shapes and structures; it
shows how things can be organized. The anatomy itself does not tell designers how to
design prototypes, but it can inform them about the fundamental nature of prototypes and
the possibilities in thinking about them.
Our proposed anatomy of prototypes includes (1) filtering dimensions and (2)
manifestation dimensions. These two types of dimensions correspond to the two
important characteristics of prototypes—prototypes as filters, and prototypes as
manifestations of design ideas.
In defining the set of filtering dimensions, we include appearance, data, functionality,
interactivity, and spatial structure (Table II). These dimensions correspond to the various
aspects of a design idea that a designer tries to represent in a prototype. They also refer to
the aspects of a design idea that the designer must consider in the exploration and
refinement of the design. We define the three core aspects of the manifested forms of
prototypes as materials, resolution, and scope (Table III).
Although they represent two different ways of looking at prototypes, both the
prototype’s filtering dimensions and the manifestation dimensions are tightly related to
each other. For example, designers who explore possible ideas of using a one-handed
mobile device interface—which is the interactivity dimension in terms of filtering—may
consider how to manifest these ideas using prototypes. Here we can readily imagine
unlimited possibilities to manifest an idea addressing the same filtering dimension. In
terms of the prototypes’ material, designers may use foam core as a material to mock up
a prototype design that is the same size as the target design in order to simulate the
holding postures for the mobile device, or they can use clay or wood to give more
realistic three-dimensional forms for ideas related to thumb positions and gestures for
interacting with the mobile device. The designers may continue to use three-dimensional
forms since their purpose is to explore the effects of one-handed interactivity with the
mobile device. This example shows us that, while affected by the filtering dimension, the
choice of manifestation dimensions involves various issues such as resources, cost, and
user perception in the use of a prototype.
Manifestation dimensions other than material are also related to the filtering
dimension. In this mobile device design example, designers are particularly interested in
the possibilities and effectiveness of one-handed interaction, such as different ways of
operating inputs using one thumb or with one-handed gestures. For this purpose,
designers may not need to implement sophisticated details of the interface’s look-and-feel
as long as the prototype provides key interface indicators that are clear enough for users
to understand where they can place and move their thumbs. In this case, the purpose
guides the designers to determine the right level of resolution of the prototype—another
manifestation dimension. It also applies to the scope of the prototype in terms of what
other parts of the design the designer needs to include in a prototype in order to be able to
examine the filtered aspect(s) of the design. For example, a designer can decide whether
or not to include corresponding outcome screens according to her selected aspect(s). Thus,
a designer has to decide what aspects of a design idea should be filtered when forming a
prototype. One prototype might only filter an appearance aspect, while another filters all
aspects at once. The challenge for a designer is to design the prototype that supports her
design intention most effectively.

5.1 Filtering Dimensions

As a part of the anatomy, we define five filtering dimensions that we believe, in a
reasonable way, cover the core aspects of a design idea in interactive systems design. The
appearance dimension is the physical properties of a design. It may include forms, colors,
textures, sizes, weights, and shapes, as well as proportional relationships among these
elements. It is not restricted to visual appearance, since characteristics such as weight,
texture, size, and shape can be sensed by touch as well as by sight. The data dimension is
the information architecture and the data model of a design. It may include the size of
data, the number of letters to be shown in each label, the amount of visible and invisible
data on screen, the semantic organization of the contents, the ways of labeling and
naming, the levels of privacy of data, and the types of information. The functionality
dimension is the functions that can be performed by the design. Focusing on this
dimension, designers may determine preferred functionalities and scenarios associated
with using different functions. The interactivity dimension is the ways in which people
interact with each part of a system. It may include feedback, input behaviors, operation
behaviors, and output behaviors. The spatial structure dimension is how each component
of a system is combined with others. It may include considerations of laying out interface
or information elements in an interactive space. If the design includes partially tangible
and intangible interfaces, such as mixed-reality systems, this dimension may involve the
relationships and interconnections between tangible and intangible interfaces.
This list of dimensions is not meant to be complete; it is, however, meant to be useful,
in ways we elaborate later. Table II shows relevant variables to be discussed in relation to
each filtering dimension.

Table II. Example variables of each filtering dimension

Filtering dimension Example variables
Appearance size; color; shape; margin; form; weight; texture; proportion;
hardness; transparency; gradation; haptic; sound

Data data size; data type (e.g. number; string; media); data use;
privacy type; hierarchy; organization

Functionality system function; users’ functionality need

Interactivity input behavior; output behavior; feedback behavior; information
behavior

Spatial structure arrangement of interface or information elements; relationship

among interface or information elements—which can be either
two- or three-dimensional, intangible or tangible, or mixed

The dimensions are tightly related to and influenced by each other; it is therefore
impossible to treat them separately. For example, the interactivity aspect of the iPod’s
wheel interface drives its basic appearance. The data dimension is likewise tightly
related to other dimensions. The size of music data that the iPod can contain— a decision
about the data dimension—led the decision making on interface design issues related to
interactivity. Since an iPod can hold more than 200 songs, the iPod’s development team
avoided the use of buttons to browse songs, instead inventing the wheel interface to
browse naturally through a large amount of songs with a thumb [Levy 2006]. This new
way of browsing songs is tightly related to the interactivity dimension. The result of this
new idea of interactivity led to novel decisions concerning its appearance.
In spite of the relationships among the dimensions, a necessity of crafting and
working on each dimension separately also exists; the separation ensures that the selected
dimension is itself carefully designed to fulfill important design values. For example,
designers cannot determine the design details of the iPod’s appearance—such as size of
the interface wheel circle, font size of the wheel’s label, color, shape of the symbols used
as labels, and texture of the surface of the interface wheel—without separately exploring
this appearance dimension of the design space using various prototypes. Prototypes can
enable designers to explore a dimension space in order to reach a decision on the final
appearance of the design outcome. Through the process of making prototypes, designers
constantly evaluate their ideas (whether formally using user tests or informally and
heuristically by using their own expertise), generating better ideas.
Prototypes allow designers to do this by filtering a dimension out from other ones but
also enable them to see the relationships among different dimensions as well. The
anatomy of prototypes we propose can guide designers to be aware of and think about
these multiple dimensions even while working on a specific dimension.
It is obvious that the relationships between these dimensions are intricate and
dynamic; no dimension is separate from any other. We see this recognition of intertwined
relationships among the dimensions as an outcome of the prototype’s anatomy.
Attempting to clarify these dimensions reveals the complexity of prototypes. The
anatomy we propose can serve an educational purpose as it enables the articulation of
structures—e.g., anatomies—of different prototypes as design knowledge that can be
taught.

5.2 Manifestation Dimensions

Though it may provide an initial direction for prototype formation, knowing only what to
filter based on the set of filtering dimensions cannot fully determine how to form a
prototype nor provide strategies for forming it. We use the term “formation” instead of
“construction” since a prototype may not need to be “constructed” out of physical matter
but can be formed by invisible triggers or behaviors. For example, a case of experience
prototyping proposed by Buchenau and Suri [2000] used a beeper to simulate a person
having a heart attack in order to understand what kinds of possible situations surrounded
the heart attack accident. They asked participants to journal the surrounding situation
when they heard the randomly activated beeper ringing or vibrating. In this prototyping
example, a prototype is not “constructed” with raw physical materials. A prototype is
“formed” by a situation and an existing object behaving in a certain way—i.e., the beeper
beeping randomly to simulate a heart attack.
What determines the specifics of how to form prototypes are the issues of what
prototypes should be composed or made out of, i.e., the materials (whether visible or
invisible) by which the prototype is made manifest; what level of fidelity the prototype
should be, i.e., the resolution of a prototype; and how complete the prototype should be,
i.e., the scope of a prototype. We call these three dimensions manifestation dimensions.
The meaning of scope is completeness and differs from the notion of resolution. Scope is
how completely a prototype covers the range of aspects of what we design even if those
aspects are not related to what we want to filter through the prototype. Those additional
aspects may help us understand the prototype more effectively. Table III shows the
definition and corresponding variables of each manifestation dimension.

Table III. The definition and variables of each manifestation dimension

Manifestation dimension Definition Example variables

Material Medium (either Physical media, e.g., paper, wood, and

visible or invisible) plastic; tools for manipulating physical
used to form a matters, e.g., knife, scissors, pen, and
prototype sandpaper; computational prototyping
tools, e.g., Macromedia Flash and
Visual Basic; physical computing tools,
e.g., Phidgets and Basic Stamps;
available existing artifacts, e.g., a
beeper to simulate an heart attack

Resolution Level of detail or Accuracy of performance, e.g.,

sophistication of feedback time responding to an input
what is manifested by a user—giving user feedback in a
(corresponding to paper prototype is slower than in a
fidelity) computer-based one); appearance
details; interactivity details; realistic
versus faked data
Scope Range of what is Level of contextualization, e.g., website
covered to be color scheme testing with only color
manifested scheme charts or color schemes
placed in a website layout structure;
book search navigation usability testing
with only the book search related
interface or the whole navigation
interface

As we discuss earlier, the economic principle of prototyping should guide designers to

determine the values of these dimensions when forming a prototype. Based on her
purpose in prototyping, a designer may use paper as a material for prototyping instead of
working computer screens. This is an example of a prototype material decision. The
designer can also vary the details of what is shown in a prototype. Even if using paper,
she can have a very detailed and sophisticated drawing or a rough sketch. This is an
example of a prototype resolution decision. When figuring out which color scheme is
best for her website design, a designer may use color schemes without the details of text,
icons, and menus on the web page. This is an example of a prototype scope decision.
How to decide these values is based on the economic principle of prototyping.
What must be understood here is that a prototype is fundamentally different from the
final product, whether or not it is identical to the final product. Prototypes are means and
tools for design and are not the ultimate target for design. In this regard, the designer’s
mindset in forming prototypes is different from that in forming the final design. When
treating something as a prototype, the designer can start to put in different materials or
take out certain materials based on the purpose of using that prototype for the design.
We argue that the manifestation dimensions influence how well a prototype performs
as an informing tool in the design process. The manifestation dimensions affect the
performance of the prototype—not in the sense of how much the prototype performs like
a final product but how well the prototype performs as a tool for evaluating design ideas
and generating better design ideas—without altering the filtering dimensions the designer
has chosen to evaluate. The manifestation dimensions may modify or influence how well
and to what degree the prototype filters the desired filtering dimensions.
The reason that we name these manifestation dimensions is that these dimensions
influence people’s perception of and reaction to a particular prototype. For example, if
we compare the two cases of evaluating how people perceive the colors of a room’s
walls—one with a three-dimensional virtual model through a computer and the other with
a three-dimensional life-sized foam-board model in a physical space, we recognize that
the two situations will affect the way in which people react to the colors due to the
different materials used to represent the variable of a specific filtering dimension, i.e.,
appearance, and more specifically colors, of the room’s wall. In this regard, the selection
of material—a virtual model or a physical model—modifies the appearance dimension of
the design of the room.
When compared to other prototyping research approaches, one of the most significant
contributions of our approach is that we strive to establish an understanding of the nature
and anatomy of prototypes that can be utilized and extended for both research and
practice. We claim that the anatomy of prototypes can be used to examine and analyze
existing prototypes as well as inform designers in their design of new prototypes. In the
next section, we describe two cases in which we have used the anatomy of prototypes in
our analysis. After these cases, we return to the question of how to use the proposed
anatomy of prototypes and what it can mean for future research, practice, and education.

6. TWO CASES EXPLAINED WITH THE ANATOMY OF PROTOTYPES

In this section, we describe two case studies based on our previous research. These two
cases led us to identify the key dimensions of our proposed anatomy of prototypes. We
present the two cases in order to describe the anatomy of prototypes by applying it to the
real contexts of prototyping.
In the first case, we investigate how different prototypes filter different aspects of a
design and how the prototypes influence the ways in which the users who interacted with
the prototypes interpreted the design concept. In this regard, the first case is for
understanding the effects of the filtering dimensions of prototypes. In the second case, we
investigate how the choice of materials, resolutions, and scopes of prototypes influence
users’ reactions toward prototypes and affect their interpretations of the design. In this
regard, the second case is for examining the effects of the manifestation dimensions of
prototypes.
When we describe the prototypes used in each case, we use the structure of the
anatomy of prototypes. The overall description of each prototype based on the anatomy
of prototypes can be seen as a prototype profile that specifies what was considered in
forming the prototypes. Since each case enables us to examine different parts of the
anatomy dimensions—i.e., the first case for the filtering dimensions and the second case
for the manifestation dimensions, we present the prototype profiles based on those
relevant dimensions. The two studies were carried out separately from each other and by
different groups of researchers. The two studies have both been described in earlier
writings [Skog and Söderlund 1999; Lim et al. 2006] but are reinterpreted for the purpose
of this study.

6.1 Case 1: Prototyping a House Design

The analysis of the first case study led us to understand that a prototype can filter
different aspects of a design. In this case study, two prototypes were formed, both
representing the same design idea. The target design was a typical family house with a
few bedrooms, a living room, a kitchen, a stairway to the second floor, and a couple of
bathrooms. One prototype was a two-dimensional floor plan of the house, and the other
prototype was a three-dimensional virtual model of the same house design. The original
study of this case [Skog and Söderlund 1999] examines how users would convey their
interpretations of the proposed design differently if the same design were represented in
two prototypes focusing on two very different filtering dimensions.

6.1.1 The Prototype Profiles. The first prototype consisted of a two-dimensional

paper-based blueprint of the house. It was a very simple representation that showed the
floor plan of the house—the spatial layout. The blueprint showed precise sizes of spaces,
proportions among spaces, and the structure of the rooms, along with the spatial
relationships. Created with simple three-dimensional modeling software, the second
prototype enabled people to interact with a three-dimensional virtual model of the house
by virtually “walking” into the home, through rooms, turning around, and experiencing
the home as if they were walking around a real home. The home was sparsely furnished.
Even though the two prototypes represented the same house, the very nature of how
the aspects of the house were manifested was very different between the two. For
instance, in a three-dimensional model, you are “forced” to have colors on the walls in
the rooms, which is not the case on a blueprint. On a blueprint, you get a bird’s-eye view
of the house—a perspective not possible in the physical world or in the three-dimensional
virtual space. In a three-dimensional virtual model, the space is something you can feel,
while, in the blueprint, the space can only be experienced as layout and relationships. A
two-dimensional blueprint, in this regard, may better filter the spatial structure dimension
of the interior of the house, while a three-dimensional model may better filter the
appearance dimension of the interior of the house. Table IV shows prototype profiles for
the two prototypes according to the filtering dimensions of the anatomy of prototypes;
these profiles enable us to see the significance of the filtering characteristics of
prototypes.
 Table IV. Prototype profiles for the prototypes used in the first case
Dimensions 2-dimensional Blueprint 3-dimensional Virtual Model
Filtering Addressed filtering Addressed filtering dimensions:
dimensions dimensions: Appearance—colors and textures
Spatial structure—precise of walls; heights and widths of
manifestation of relationships spaces
and proportions among spaces
Interactivity—the possibility to
Not addressed filtering move around and interact with the
dimensions: 3-dimensional space
appearance, data, functionality,
interactivity Spatial structure—precise
manifestation of relationships and
proportions among spaces

Not addressed filtering

dimensions:
data, functionality

6.1.2 Prototyping and Results. The researchers tested the two prototypes with two
groups of users. The study consisted of eight subjects, four men and four women. They
were divided into two groups of four. Each individual in each group was asked to explore
and interact with either (and only) the blueprint or the three-dimensional model. The
experiment was performed individually.
Each person was told to examine the house as if she was considering buying the
house. They were asked to form a reasoned judgment on how they liked the house and if
they could imagine living there. They were allowed to use the time they thought they
needed to get a fair understanding of the house. They used approximately 10–15 minutes.
After each session, they were asked to describe the house, to express specific qualities
and characteristics that they had noted. They were told to describe their experience of the
house in their own words and try to express their judgments. They were not asked
specific questions about the house. At the end of each interview, they were also asked to
draw a blueprint of the house from memory. This was done to determine how the two
different prototypes communicate the spatial structure of the house.
The results from the study show that the individuals interacting with the different
prototypes established clearly different understandings of the house. The individuals that
dealt with the blueprint of the house used a language that, not surprisingly, consisted of
words that referred to the layout and spatial relationships between rooms. They
commented on the overall use of the space, such as “the kitchen seems small compared to
the living room.” The group that dealt with the three-dimensional virtual model
commented on the appearance of the house interior, using aesthetic concepts rather than
structural concepts. They also commented on how they felt about the house; they
mentioned that the house felt large or small, airy or tight. The language they used also
expressed their experience as if they had “been” in the house.
It is obvious that the overall judgment of the house differs markedly between the two
groups. The individuals in the two groups also had different opinions on the functionality
of the house. The blueprint group had many ideas about how to re-design and re-model
the house, how to take down walls, etc. They also commented on the functionality of the
kitchen. The people who interacted with the three-dimensional virtual model, however,
had only minor comments on functionality. A possible explanation for this difference is
that the three-dimensional virtual model is experienced more as a finished product, while
the blueprint is experienced only as a proposal—provisional and open to change. Overall,
the three-dimensional virtual model group commented that the house was small, which
was something that no one in the other group mentioned. The final difference was that
when the two groups were asked to draw the house’s layout. The blueprint group created
accurate drawings, while the three-dimensional virtual model group created completely
inaccurate layouts. Individuals in the three-dimensional virtual model group were unable
to put the rooms in the right places and grossly misjudged the sizes and shapes of rooms.
From this case study, we can observe how the choice of representational forms—i.e.,
blueprint versus three-dimensional virtual model—is critical to how a prototype filters
the properties of a target design. If we look at the two prototype profiles in Table IV, we
can remark on how prototypes lead to different results when each prototype addresses
different filtering dimensions. In this case, it is clear that the prototype profile strongly
impacts the way in which users experience the final design. The prototype cannot give
relevant information about certain aspects of the final design if those aspects are not
manifested since they cannot be experienced. The blueprint prototype, for instance,
cannot filter appearance since it does not manifest any such qualities. The three-
dimensional virtual model manifests the interaction a person can have with a house, such
as moving around, turning, etc., and it gave users the feeling of having been there, an
experience that strongly influenced their judgment of the house. The blueprint group
experienced nothing similar to this sense of being there. Even though both prototypes are
manifestations of the same design, the two participant groups experienced and valued the
spatial structure in distinctly different ways.
 This case shows that a prototype only filters those dimensions manifested in the
prototype. The blueprint prototype works well if the designer wants to find out more
about (or filter) the spatial structure dimension, but it does not inform the designer about
any other filtering dimension. Adding more filtering dimensions creates a more complex
prototype that is more difficult to interpret. This added complexity means that the
designer has to decide what to filter and carefully craft the prototype in relation to her
chosen filtering dimensions. These findings resonate with our economic principle of
prototyping.

6.2 Case 2: Prototyping a Mobile Phone Application

In the first case, the focus is on filtering dimensions and the importance of choosing what
to filter. Deciding the filtering dimensions, however, does not provide fixed options for
the choices in the manifestation dimensions. For example, the two-dimensional blueprint
prototype used in the first case can be either represented on a sheet of paper or shown on
a computer screen. Both manifestations address the same filtering dimension—i.e.,
spatial structure—but the materials are different from each other. In the second case study,
we found that differences in the manifestation dimensions, even if the filtering
dimensions remain constant, lead to different outcomes. When choosing paper or
computer-screen for the blueprint prototype, designers should carefully consider not only
which way is more effective in terms of the economic value of prototyping but also how
the chosen values of the manifestation dimension—in this case, material—may affect
users’ perceptions of the prototype.
For the second case, we analyzed one of our previous research projects [Lim, et al.
2006]; in this project, we compare three different prototypes of the same design idea (a
mobile phone) to determine how changes in manifestation influence user experience. The
results of this study led us to realize the importance of the consideration of the
manifestation dimensions in forming prototypes.
We formed three different prototypes in this study: a paper-based prototype, a
partially working computer screen-based prototype, and a fully functional mobile phone.
With this case, we describe the effects of the use of different values for the manifestation
dimensions as applied in these different prototypes, as well as how these choices relate to
the economic principle of prototyping. Despite being inexpensive material for visualizing
design ideas, paper may cost more than computer-based prototyping tool when it needs to
communicate a complex and detailed level of interactivity. We also discuss this issue in
terms of selecting the right values for a prototype’s manifestation dimensions.
 6.2.1 The Prototype Profiles. The three prototypes constructed in this case study
target the evaluation of usability of a text-messaging feature of a mobile phone—in this
case, the Samsung VI660. The approaches behind the three prototypes are all commonly
used in HCI. First, our first prototype was a paper prototype (Fig. 2). Promoted as an
example of an effective low-fidelity prototyping technique, paper prototyping is claimed
to be beneficial for early concept evaluation and user involvement for idea generation
[Rudd, et al. 1996; Snyder 2003]. In this case study, we focused only on evaluating the
usability of the design, considered an appropriate use of paper prototyping [Snyder 2003].
Second, the computer screen-based prototype (Fig. 3) was used to represent both the
keypad and the screen of the mobile phone. This is another popular approach for testing
mobile phone usability as it is cheaper than making the hardware for these parts and
connecting them together, for example, using augmented reality technology [Nam and
Lee 2003; Pering 2002]. Third, a fully functional prototype, i.e., an actual Samsung
VI660, was used (Fig. 4). Our use of the fully functional artifact was similar to how
clinical trials use a control group in comparison with one or more treatment groups. We
wanted to determine how users would experience different manifestations of the same
design aspects in comparison with the fully functional artifact.

Fig. 2. The paper prototyping setup and its use situation [Lim et al. 2006].
 White paper

Fig. 3. The computer-based prototype and its test setup [Lim et al. 2006].

Fig. 4. The fully functional prototype (Samsung VI660) [Lim et al. 2006].

Unlike the first case study, these three prototypes all focus on evaluating the same
thing—the usability of the text-messaging feature of a mobile phone. In this case, the
target filtering dimension is the same—i.e., interactivity. The values of the manifestation
dimensions, however, differ across the three prototypes. We describe the details and
differences in those values for the manifestation dimensions as prototype profiles for the
three prototypes in Table V.

Table V. Prototype profiles for the prototypes used in the second case
Dimensions Paper prototype Computer screen- Final product
based prototype
Manifestation Materials—paper; Materials—mobile Materials—same as
dimensions foam core board; phone simulation the final product
knife; pen; wooden toolkit; laptop
sticks; glue; yellow computer; mouse
cellophane paper;
 two-dimensional
phone appearance
color-printout

Resolution—rough Resolution— Resolution—the same

and simplified simplified screens as the final product
sketches of screens; using given interface
formats from the
simulation toolkit;

(picture from [Lim et

(picture from [Lim et al. 2006])
al. 2006])
large time lags by (picture from [Lim et
human’s simulating al. 2006])
the product partially working in a
behaviors; buttons on simulated way;
the keypad are not keying with a mouse
push-enabled (not a touch screen)

Scope— Limited to Scope— Limited to Scope—exactly same

the text-messaging the text-messaging as the final product
feature and making feature and making
other parts as “not other parts as “not
available” screens available” screens

When forming each prototype (except the fully working one which is the same as the
actual phone) in the original study, the key constraint was following how each type of
prototype is conventionally defined. For example, a paper prototype should be cheap to
make and easily to sketch; a computer screen-based prototype created by using a toolkit
can easily demonstrate and evaluate real-time interactions without constructing the actual
physical parts of the product. With these examples and their use in test sessions, we
discuss how the different prototypes, which used different values of the manifestation
dimensions, affected the users’ perceptions of the same filtering dimension and the same
design idea.

6.2.2 Prototyping and Results. In this study, the test sessions followed the
conventional formal usability testing method, including a testing session with a think-
aloud protocol while users carried out specified tasks, followed by a debriefing session
where users answered questions in terms of their evaluation on the key aspects of
design’s usability. We recruited a total of fifteen participants—five per prototype. Each
participant was given only one prototype of the three. The testing setup across the three
prototypes was identical in terms of the list of the given tasks, the debriefing questions,
and the script used by the facilitator to lead the testing. This study setup worked well in
allowing us to focus solely the effects of the manifestation dimension variables on the
results of the evaluation sessions.
One of the striking results from this study is that only twenty percent of the total
usability findings are common to all three prototypes although the parts of the design
tested were all same. A detailed analysis of these findings tells us that the manifestation
dimensions—the materials used, the level of resolution, and the covered scope—
significantly matter. Those findings identified by only one or two of the prototypes but
not all three included both false findings—i.e., things that were not problems in the
design itself but were caused by the characteristics of the prototype itself—and missing
findings—i.e., things that were not found by a particular prototype due to its limited
characteristics compared to the fully functional prototype.
With the paper prototype, the material used could not enable users to push the buttons
on the keypad area. This made it difficult for the computer-person—whose task was to
display corresponding feedback and output in response to the user’s input—to know
whether a user had pushed button or even which button was pushed, thus delaying
responses. Furthermore, the abstractness and roughness of the screen images sometimes
made users confused about an image’s precise meaning, a confusion which we did not
observe in the other prototypes. This instance tells us that the resolution dimension also
significantly matters since the level of detail and sophistication of the images affected
users’ interpretations of the interface elements. With the computer screen-based
prototype, the conventional graphical user interfaces (GUI) influenced the users’
interpretations of the labels on the screen images. Since all the parts of the mobile phone
design were shown on the computer screen, many users first tried to click directly on the
screen images instead of using the buttons on the keypad image; some users also tried to
use the keyboard attached to the laptop computer to type the text message even though
we covered the keyboard with white paper (Fig. 3). This instance tells us that the type of
materials significantly affects users’ ways of responding to prototypes. Without careful
consideration of these effects, there is a high probability of obtaining unintended user
interpretations of the design. However, as this prototype has similar feedback behavior to
the fully working product in terms of the response times to users’ inputs, many findings
overlapped those of the fully functional one, despite the material difference. This finding
informs us that a careful plan for forming a prototype—one that considers the dynamics
among the material, resolution and scope dimensions—enables precise projections of
how the design may affect users. Using this result, we can see that it is possible to
explore certain aspects of a design without making it fully working, as long as we
carefully form the prototype while being aware of the effects of the manifestation
dimensions. For the full details of findings in this case study, see [Lim, et al. 2006].

7. DISCUSSION: USING THE ANATOMY FRAMEWORK FOR PROTOTYPING

IN INTERACTION DESIGN
The two case studies support and illustrate our initial idea about two fundamental
characteristics of a prototype—one as a medium for exploring a design space by filtering
certain aspects of design ideas, and the other as a medium that purposefully manifests
those filtered aspects of the design ideas through different means of externalization. The
results we gathered from the two case studies led us to see how significant those two
characteristics of prototypes are in terms of knowledge that can be gained from
prototyping. Based on our notion of an anatomy of prototypes and our case studies, we
see three possible contributions to interaction design.
First, the anatomy framework provides a language that can be used to articulate any
prototype. This capability can contribute to cumulative knowledge production in the
study of prototypes within the field of interaction design and research. As it also creates a
language, this framework can provide support for critique, examination, and analysis of
prototypes used for manifesting design ideas. It can lead to building inventories of
prototype ideas for different filtering dimensions. In addition, the accumulation of such
inventories will reveal patterns of important aspects of designs for different types of
interactive artifacts. Those patterns can also be categorized according to different design
values or criteria, such as usability, ergonomics, aesthetics, performance, sustainability,
and ethics. The idea of capturing emerging design patterns is analogous to what has been
done with the use of the pattern language [Alexander et al. 1977] for design in HCI
[Tidwell 2005; van Duyne et al. 2002].
Second, the anatomy framework of prototypes provides a critical thinking guide when
designing and constructing prototypes. Since designers and researchers using it can better
understand what characteristics of prototypes matter, this framework will help them to
make careful and intentional choices of materials, resolutions, and scopes of prototypes—
i.e., the manifestation dimensions—in relation to the aspects of a design idea—i.e., the
filter dimensions,—that they plan to explore in their prototypes. This will be supportive
not only for design and research practice in HCI but also for HCI and design education in
relation to prototyping activities. The process of designing and constructing prototypes is
a time- and resource-consuming process, making it difficult for students to gain
adequate experience with the pros and cons of prototypes. If presented with carefully
chosen prototypes that they can analyze with the help of the anatomy framework,
students might be able to build an enhanced sensitivity to prototype quality and how they
can serve design.
Third, this framework can be used for constructing prototype profiles in real design
practice that can help designers in producing quick-and-dirty prototyping plans before
they construct prototypes. These plans allow them to discuss and share their prototyping
ideas with others in design teams in advance; such sharing will provide a communication
point. It will also help in comparing and integrating different prototypes that partially
represent a final design outcome. The deep understanding of fundamental characteristics
of prototypes that this framework enables will also allow them to make salient and
economic decisions about their prototyping.

8. CONCLUSION
The results from these studies have convinced us that it is possible to clearly identify and
plan for prototype characteristics and that we must base those considerations on why and
how we intend a particular prototype to support the design process. The studies also
convinced us that it is possible to understand qualities of prototypes in a more
conceptually structured and pragmatically useful way. This understanding means that, by
being aware of an anatomy of prototypes, designers can approach the tasks of forming
and using prototypes in a more deliberate, intentional, and reflective way, and, we hope,
with a higher degree of precision.
We base our definition of the anatomy of prototypes on the fact that prototypes are
not the same as the final design. To create a prototype is to find the manifestation that, in
its most economic form, will filter the qualities in which the designer is interested,
without distorting the understanding of the whole. A designer must be aware of the fact
that the manifested forms of prototypes are different from the final form of the design and
that prototypes can significantly affect the ways of perceiving the manifested ideas in
various situations of using the prototypes.
We do not propose this framework as a prescriptive approach for the design of
prototypes in interaction design. But, designers can learn from the framework and can let
the framework and earlier experiences inform their decisions in a specific design situation.
The anatomy of prototypes represents a way of thinking about prototypes, rather than a
method that may lead to “good” prototypes. The framework can be seen as both an
analytic and reflective tool. It can provide designers with conceptual and reflective
guidance not only on how to design prototypes but also on how to interpret prototyping
results.
We believe that the notion of a “good” prototype can only be understood in relation
to the specific purpose of the design process and to the specific issue that a designer is
trying to explore, evaluate, or understand. The purposes for which prototypes are used
can be broadly categorized into the following areas: (1) evaluation and testing; (2) the
understanding of user experience, needs, and values; (3) idea generation; and (4)
communication among designers. These categories are not meant to be mutually
exclusive, and any one prototype can be used for multiple purposes. The notion of
prototype profiles that we have introduced can be used for planning and specifying
prototypes in design practice according to these different purposes.
This fundamental conception of prototypes is critical in our field as it provides a
systematic way of understanding, describing, and forming the knowledge of prototypes,
which is not established in prior research. It is, however, true that the framework we
propose here is not an absolute one. We expect that our framework for prototyping will
lead to more research comparing different roles and effects of prototypes in design,
perhaps by adapting new and unconventional ways of constructing prototypes not yet
commonly used.

ACKNOWLEDGMENTS
We like to thank our Ph.D. candidate, Justin Donaldson, for his valuable input on this
research.

REFERENCES
ADAMS, R. 2002. Understanding Design Iteration: Representations from an Empirical Study. In Proceedings
of Common Ground International Conference, September 5 – 8, 2002. London, Design Research Society.
ANATOMY. Oxford English Dictionary. Oxford University Press. http://www.oed.com/ (accessed: August 16,
2006).
ALEXANDER. C., ISHIKAWA, S., AND SILVERSTEIN, M., WITH JACOBSON, M., FIKIDAHL-KING, I,
AND ANGEL, S. 1977. A Pattern Language, Oxford University Press, New York, NY.
AVRAHAMI, D. AND HUDSON, S. E. 2002. Forming interactivity: a tool for rapid prototyping of physical
interactive products. In Proceedings of the Conference on Designing interactive Systems: Processes, Practices,
Methods, and Techniques, London, England, June 25 - 28, 2002. ACM Press, New York, NY, 141-146.
BAILEY, B. P., KONSTAN, J. A., AND CARLIS, J. V. 2001. DEMAIS: designing multimedia applications
with interactive storyboards. In Proceedings of the Ninth ACM international Conference on Multimedia, Ottawa,
Canada, September 30 - October 05, 2001. ACM Press, New York, NY, 241-250.
BRANDT, E. AND MESSETER, J. 2004. Facilitating collaboration through design games. In Proceedings of
the Eighth Conference on Participatory Design: Artful integration: interweaving Media, Materials and
Practices - Volume 1, Toronto, Ontario, Canada, July 27 – 31, 2004. ACM Press, New York, NY, 121-131.
BUCHENAU, M. AND SURI, J. F. 2000. Experience prototyping. In Proceedings of the Conference on
Designing interactive Systems: Processes, Practices, Methods, and Techniques, New York, NY, United States,
August 17 - 19, 2000. ACM Press, New York, NY, 424-433.
BUXTON, B. 2007. Sketching User Experiences: Getting the Design Right and the Right Design, Morgan
Kaufmann, San Francisco, CA.
CLARK, A. 2001. Natural-Born Cyborgs? In Proceedings of Cognitive Technology: Instruments of Mind: 4th
International Conference, Warwick, UK, August 6-9, 2001, Springer-Verlag, Berlin/Heidelberg, Germany, 17-
24.
COCKTON, G. AND WOOLRYCH, A. 2002. Sale must end: should discount methods be cleared off hci’s
shelves? Interactions 9, 5, 13-18.
CONVERTINO, G., NEALE, D. C., HOBBY, L., CARROLL, J. M., AND ROSSON, M. B. 2004. A laboratory
method for studying activity awareness. In Proceedings of the Third Nordic Conference on Human-Computer
interaction, Tampere, Finland, October 23 – 27, 2004. ACM Press, New York, NY, 313-322.
FLOYD, C. 1984. A Systematic look at prototyping. In Approaches to Prototyping, BUDDE, R.,
KUHLENKAMP, K., MATHIASSEN, L., AND ZULLIGHOVEN, H. Eds. Springer-Verlag, Berlin, Germany,
1-18.
GOEL, V. AND PIROLLI, P. 1992. The Structure of design problem spaces. Cognitive Science 16, 395-429.
GRADY, H. M. 2000. Web site design: a case study in usability testing using paper prototypes. In Proceedings
of IEEE Professional Communication Society international Professional Communication Conference and
Proceedings of the 18th Annual ACM international Conference on Computer Documentation: Technology &
Teamwork, Cambridge, Massachusetts, September 24 - 27, 2000. ACM Special Interest Group for Design of
Communication. IEEE Educational Activities Department, Piscataway, NJ, 39-45.
GREENBERG, S. AND BOYLE, M. 2002. Customizable physical interfaces for interacting with conventional
applications. In Proceedings of the 15th Annual ACM Symposium on User interface Software and Technology,
Paris, France, October 27 - 30, 2002. ACM Press, New York, NY, 31-40.
GUTIERREZ, O. 1989. Prototyping techniques for different problem contexts. In Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems: Wings For the Mind. ACM Press, New York, NY, 259-
264.
JONES, W., SPOOL, J., GRUDIN, J., BELLOTTI, V., AND CZERWINSKI, M. 2007. "Get real!": what's
wrong with hci prototyping and how can we fix it?. In CHI '07 Extended Abstracts on Human Factors in
Computing Systems, San Jose, CA, USA, April 28 - May 03, 2007, ACM Press, New York, NY, 1913-1916.
LANDAY, J. A. 1996. SILK: sketching interfaces like krazy. In Conference Companion on Human Factors in
Computing Systems: Common Ground, Vancouver, British Columbia, Canada, April 13 – 18, 1996. ACM Press,
New York, NY, 398-399.
LEE, J. C., AVRAHAMI, D., HUDSON, S. E., FORLIZZI, J., DIETZ, P. H., AND LEIGH, D. 2004. The
calder toolkit: wired and wireless components for rapidly prototyping interactive devices. In Proceedings of the
2004 Conference on Designing interactive Systems: Processes, Practices, Methods, and Techniques, Cambridge,
MA, USA, August 01 - 04, 2004. ACM Press, New York, NY, 167-175.
LEVY, S. 2006. The Perfect Thing, Wired 14, 11 (November 2006)
<http://www.wired.com/wired/archive/14.11/ ipod.html>
LICHTER, H., SCHNEIDER-HUFSCHMIDT, M., AND ZÜLLIGHOVEN, H. 1993. Prototyping in industrial
software projects—bridging the gap between theory and practice. In Proceedings of the 15th international
Conference on Software Engineering, Baltimore, Maryland, United States, May 17 - 21, 1993. IEEE Computer
Society Press, Los Alamitos, CA, 221-229.
LIM, Y. 2003. Design Information Framework for Integrating Multiple Aspects of Design, Ph.D. Dissertation,
Illinois Institute of Technology, Chicago.
LIM, Y., PANGAM, A., PERIYASAMI, S., AND ANEJA, S. 2006. Comparative analysis of high- and low-
fidelity prototypes for more valid usability evaluations of mobile devices. In Proceedings of the 4th Nordic
Conference on Human-Computer interaction: Changing Roles, Oslo, Norway, October 14 – 18, 2006. ACM
Press, New York, NY, 291-300.
LIN, J., NEWMAN, M. W., HONG, J. I., AND LANDAY, J. A. 2000. DENIM: finding a tighter fit between
tools and practice for Web site design. In Proceedings of the SIGCHI Conference on Human Factors in
Computing Systems, The Hague, The Netherlands, April 01 – 06, 2000. ACM Press, New York, NY, 510-517.
LIU, L. AND KHOOSHABEH, P. 2003. Paper or interactive?: a study of prototyping techniques for ubiquitous
computing environments. In CHI '03 Extended Abstracts on Human Factors in Computing Systems, Ft.
Lauderdale, Florida, USA, April 05 - 10, 2003. ACM Press, New York, NY, 1030-1031.
LÖWGREN, J., STOLTERMAN, E. 2004. Thoughtful Interaction Design: A Design Perspective on
Information Technology. MIT Press, Cambridge, MA.
MACINTYRE, B., GANDY, M., DOW, S., AND BOLTER, J. D. 2004. DART: a toolkit for rapid design
exploration of augmented reality experiences. In Proceedings of the 17th Annual ACM Symposium on User
interface Software and Technology, Santa Fe, NM, USA, October 24 – 27, 2004. ACM Press, New York, NY,
197-206.
MCCURDY, M., CONNORS, C., PYRZAK, G., KANEFSKY, B., AND VERA, A. 2006. Breaking the fidelity
barrier: an examination of our current characterization of prototypes and an example of a mixed-fidelity success.
In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Montréal, Québec,
Canada, April 22 – 27, 2006. ACM Press, New York, NY, 1233-1242.
MORAN, T. P. AND CARROLL, J. M. Eds. 1996. Design Rationale: Concepts, Techniques, and Use.
Lawrence Erlbaum Associates, Mahwah, NJ.
MULLER, M. J. 2001. Layered participatory analysis: new developments in the CARD technique. In
Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Seattle, Washington, United
States, 2001. ACM Press, New York, NY, 90-97.
NAM, T. AND LEE, W. 2003. Integrating hardware and software: augmented reality based prototyping method
for digital products. In CHI '03 Extended Abstracts on Human Factors in Computing Systems, Ft. Lauderdale,
Florida, USA, April 05 – 10, 2003. ACM Press, New York, NY, 956-957.
PERING, C. 2002. Interaction design prototyping of communicator devices: Towards meeting the hardware-
software challenge. Interactions 9, 6, 36-46.
REICHL, P., FROEHLICH, P., BAILLIE, L., SCHATZ, R., AND DANTCHEVA, A. 2007. The LiLiPUT
prototype: a wearable lab environment for user tests of mobile telecommunication applications. In CHI '07
Extended Abstracts on Human Factors in Computing Systems, San Jose, CA, USA, April 28 - May 03, 2007.
ACM Press, New York, NY, 1833-1838.
RETTIG, M. 1994. Prototyping for tiny fingers. Communications of the ACM 37, 4, 21-27.
RUDD, J., STERN, K., AND ISENSEE, S. 1996. Low vs. high-fidelity prototyping debate. Interactions, 3, 1,
76-85.
SCHNEIDER, K. 1996. Prototypes as assets, not toys: why and how to extract knowledge from prototypes. In
Proceedings of the 18th International Conference on Software Engineering, Berlin, Germany, March 25 - 29,
1996, IEEE Computer Society, Washington, DC, 522-531.
SCHÖN D. A. 1982. The Reflective Practitioner: How Professionals Think in Action. Harper Collins, New
York, NY.
SCHÖN D. A. 1987. Educating the Reflective Practitioner: Toward a New Design for Teaching and Learning in
the Professions, Jossey-Bass, San Francisco, CA.
SEFELIN, R., TSCHELIGI, M., AND GILLER, V. 2003. Paper prototyping - what is it good for?: a
comparison of paper- and computer-based low-fidelity prototyping. In CHI '03 Extended Abstracts on Human
Factors in Computing Systems, Ft. Lauderdale, Florida, USA, April 05 - 10, 2003. ACM Press, New York, NY,
778-779.
SKOG, D. & SÖDERLUND, M. 1999. Virtual information representation. In Proceedings of the 22nd
Information Systems Research Seminar in Scandinavia, Keuruu, Finland, August 7-10, 1999.
SNYDER, C. 2003. Paper Prototyping: The Fast and Easy Way to Define and Refine User Interfaces. Morgan
Kaufmann Publishers, San Francisco, CA,.
SPOOL, J. & SCHROEDER, W. 2001. Testing web sites: five users is nowhere near enough. In CHI '01
Extended Abstracts on Human Factors in Computing Systems, Seattle, Washington, USA, March 31 - April 05,
2001. CHI '01. ACM, New York, NY, 285-286.
SVANAES, D. AND SELAND, G. 2004. Putting the users center stage: role playing and low-fi prototyping
enable end users to design mobile systems. In Proceedings of the SIGCHI Conference on Human Factors in
Computing Systems, Vienna, Austria, April 24 - 29, 2004. ACM Press, New York, NY, 479-486.
THOMPSON, M. AND WISHBOW, N. 1992. Prototyping: tools and techniques: improving software and
documentation quality through rapid prototyping. In Proceedings of the 10th Annual international Conference
on Systems Documentation, Ottawa, Ontario, Canada, October 13 - 16, 1992. ACM Press, New York, NY, 191-
199.
TIDWELL, J. 2005. Designing Interfaces. O'Reilly Media, Sebastopol, CA.
TYSZBEROWICZ, S. AND YEHUDAI, A. 1992. OBSERV—a prototyping language and environment. ACM
Trans. Softw. Eng. Methodol. 1, 3, 269-309.
VAN DUYNE, D. K., LANDAY, J. A., AND HONG, J. I. 2002. The Design of Sites: Patterns, Principles, and
Processes for Crafting a Customer-Centered Web Experience, Prentice Hall, Upper Saddle River, NJ.
VAN LEEUWEN, C., VERSTIJNEN, I. AND HEKKERT, P. 1999. Common unconscious
dynamics underlie common conscious effects: a case study in the interactive nature of perception
and creation. In Modeling Consciousness Across the Disciplines, S. JORDAN, Ed. University Press
of America, Lanhan, MD.
VIRZI, R. A., SOKOLOV, J. L., AND KARIS, D. 1996. Usability problem identification using both low- and
high-fidelity prototypes. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems:
Common Ground, Vancouver, British Columbia, Canada, April 13 - 18, 1996. ACM Press, New York, NY,
236-243.
WALKERS, M., TAKAYAMA, L., & LANDAY, J. 2002. High-fidelity or low-fidelity paper or computer
medium? In Proceedings of the Human Factors and Ergonomics Society 46th Annual Meeting: HFES2002.
September 30-October 4, 2002, 661-665.
WONG, Y. Y. 1992. Rough and ready prototypes: lessons from graphic design. In Posters and Short Talks of
the 1992 SIGCHI Conference on Human Factors in Computing Systems, Monterey, California, May 03 - 07,
1992. ACM, New York, NY, 83-84.
ZUCCONI, L., MACK, G., AND WILLIAMS, L. G. 1990. Using object-oriented development for support
prototyping. In Proceedings of the 12th international Conference on Software Engineering, Nice, France,
March 26 - 30, 1990. IEEE Computer Society Press, Los Alamitos, CA, 129-132.

Received xxx; revised xxx; accepted xxx.