Sie sind auf Seite 1von 14

..

..
7
Interface Metaphors and
Conceptual Models
7.1 Verbal metaphors 142
7.2 Virtual interface metaphors 145
7.3 Classification of interface metaphors for applications 149
7.4 Ubiquitous computing 149
7.5 Conceptual models 151
Key points 153
Further reading 153
Aims and objectives
The aim of this chapter is to describe how the cognitive principle of
metaphorical reasoning - that we use prior knowledge to understand new
situations - has been applied to user interface design. After working
through this chapter you should be able to:
describe the difference between verbal and virtual interface metaphors,
explain why the development of composite interface metaphors has
been necessary,
discuss the problems and advantages of using metaphors at the
interface,
decide whether you think the interfaces that you use or develop should
model some aspect of the world [that is, metaphor-based] or that it be
invisible to the user (that is, ubiquitous computing).
understand the different ways in which the term 'conceptual model' is
used in HCJ.
NOTICE
, THIS MI\TEI(IAL MAY OF PROTECTED BY
141
, COI"YI;ICiiiT lAW (Till F 17 US CODE)

I,
I
'
t!
\
I
'
142 Interface metaphors and conceptual models
Overview
The term 'metaphor' is traditionally associated with language use. When we
want to convey an abstract concept in a more familiar and accessible form
we frequently resort to using metaphoric expressions. For example, we
tend to talk about time, which is a very abstract concept, in terms of money
(L.akoff and Johnson, 1980). We save it, spend it, waste it, give someone
our time, live on borrowed time and so on. In fact, the whole of language is
based on these types of metaphorical abstractions.
When we consider how system interfaces have been designed, we
can also see how metaphors play an important role. The objects on the
screen, the types of user interactions we perform, the way the system
responds, the names given to command names, tend to be based on
familiar terms. Desktops, icons, menus, windows, cutting, pasting, copying
are either system objects or actions that are part of a virtual interface
metaphor. In fact, it is difficult to think of a system that is not based on some
form of metaphor. We need to ask ourselves, therefore, to what extent and
how they help users interact with computer systems.
m 1: ::nm :: 1 ::a:: 1 WI!
!i!i li!lii' I' 1i I i ; ill !@! :
[!D Verbal metaphors
When confronted with a new piece of technology, such as a computer, for the first
time people will often compare it to a machine with which they are familiar in a
metaphorical way. The classic example is of people who use a word processor for
the first time; it occurs to them how similar it is to a typewriter. Having activated the
typewriter schema (see Chapter 6) they are then able to interpret and predict more
readily how the word processor functions. On seeing that the computer has a
keyboard the obvious inference is that it behaves like the qwerty keyboard on a
typewriter. Hence, an obvious assumption is that the character keys should act in the
same way as they do for the typewriter.
These links provide the basic foundation from which users develop their mental
models of computer systems. Knowledge about a familiar domain in terms of
elements and their relation to each other is mapped on to elements and their
relations in the unfamiliar domain. Elements include the keyboard, the spacebar
and the return key. Relations between the elements include 'only one character key
can be hit at any one time' and 'hitting a character key will result in a letter being
displayed on a visible medium'. By drawing on this prior knowledge a learner can
develop an understanding of the new domain more readily.
Verbal metaphors 143
As well as similarities between a new and familiar domain, however, there are
obviously going to be many dissimilarities. Two examples are the spacebar and
backspace key, which are both present on the typewriter and word processor
keyboards but which perform different functions. Whereas the backspace key on the
typewriter physically moves the carriage (this is the mechanism that holds the paper
feed rollers and travels across the top of the machine) one space back and the
spacebar moves the carriage one space forward, the backspace key on the word
processor deletes the character marked by the cursor on the screen while the
spacebar inserts a blank character. Moreover, if users want to move backwards or
forwards, as they would do when using a typewriter, then they have to use another set
of cursor control keys (or move the cursor on the screen via a mouse). The effect of
changing the function played by the spacebar and return key, together with
introducing a different way of doing spacing, often causes problems for learners
because it contravenes their expectations about how the elements and their relations
should behave. However, it should also be noted that once users become aware of
the discrepancies and differences between the old and new systems, they can
develop a new mental model, accordingly.
EXERCISE
Can you think of any other similarities or mismatches between the typewriter and the
word processor domains?
A number of studies have been carried out that have investigated the effects of
providing verbal metaphors in the form of written or spoken instructions. For
example, Foss eta/. (1982) looked at the effects of providing an 'advance organizer'
on learning to use a word processor. The advance organizer described how files were
created, stored and retrieved in terms of a filing cabinet metaphor. The results
showed that subjects who had been presented with the verbal metaphor before using
the word processor showed better performance in terms of making fewer errors and
144 Interface metaphors and conceptual models
Drawing by Koren; 1979 The New Yorker Magazine, Inc.
faster times to complete the tasks. Other studies have also shown similar
improvements (e.g. Borgman, 1986).
In general, the results suggest that verbal metaphors can be useful tools to help
users begin to understand the new system. However, it must be remembered that
computer systems are much more complex and have different ways of doing things.
For example, in the typewriter analogy there are clearly a number of properties about
the word processing domain that do not map on to existing features in the typewriter
domain. The most obvious include the ability to save text and manipulate it in a
number of ways (such as copying sections and formatting). Here there are no
obvious links with the typewriter domain, so users have to develop a new
understanding. In addition, it is important that users understand how the new system
works as a computer system. Accordingly, instructions and descriptions of the system
need to be developed that also describe aspects of the structure and function of the
system (see Halasz and Moran, 1982) .
Question 7. 1
What knowledge types are mapped between the familiar and unfamiliar domains of
metaphors?
Virtual interface metaphors 145
Figure 7.1 The desktop metaphor as it originally appeared on the Star screen (Smith eta/.,
1982).
L 7.2 ) Virtual interface metaphors
One of the first computer companies to realize the enormous potential of designing
interfaces to be more like the physical concrete world that people are familiar with
was Xerox (see Box 1.4). Instead of developing verbal metaphors as ways of helping
users to understand the interface, they went one step further and designed an
interface metaphor that was based on the physical office. The outcome was the
'Star user interface' (Smith et a/., 1982) discussed in Chapter I. The core aspect of
the interface metaphor was to create electronic counterparts to the physical objects
in an office. This involved representing office objects as icons on the screen. These
included paper, folders, filing cabinets and in and out trays, as shown in Figure 7.1.
The overall organizing metaphor that was presented on the screen was of a desktop,
resembling the top of a typical office desk.
Instead of being abstract entities, having arbitrary names, files were transformed
into pictorial representations, which everyone could easily identify and understand
(see Chapter 5 for more on icons). Moreover, having this basic understanding would
provide the basis from which users would know how to interact with the icons. Just as
,
146 Interface metaphors and conceptual models
one opens, closes, copies and trashes paper files in the physical world, the interface
was designed so that equivalent actions could be done on the electronic versions.
The trick was to design an input device that enabled such electronic actions to be
performed intuitively. And so the 'mouse' was developed to enable actions
equivalent to physically handling documents, which were achieved by 'clicking',
'pointing', 'selecting', 'moving' and 'dragging'.
EXERCISE
What are the physical equivalent actions for (i) clicking, (ii) pointing, (iii) selecting
and (iv) dragging?
A difference between virtual interface metaphors and verbal metaphors is that the
former are part of the interface. Whereas verbal metaphors invite the user to 'see' the
similarities and dissimilarities between the system and the familiar domain, interface
metaphors combine the system and familiar domains into one entity. Instead of
imagining parts of the computer system to be like a typewriter, the interface
metaphor conflates the familiar domain and the new system domain into one model.
In other words, the desktop metaphor is like an office desktop but it is also the
system interface. The effect is that users will develop mental models of the system
that are more like the metaphor world rather than how the underlying system works.
When they place an icon of a file into an icon of a folder they will assume that the
system is doing just that (rather than it changing the pointer to the file).
Hence, instead of using the metaphor as a basis from which to develop a new
mental model of the new domain, the metaphor is the model that is learned. This
means that users will t ~ n to develop functional-based mental models of the system
and be largely unaware of the structural aspects of the system (see Chapter 6 on the
difference between structural and functional models).
Question 7.2
What is the difference between a verbal metaphor and a virtual interface metaphor?
A design problem with the 'metaphor as model' approach is working out ways in
which to incorporate additional functionality, which is not part of the interface
..
Virtual interface metaphors 1 147
metaphor, but which enables the computer application to be more powerful than
non-electronic means. For example, how has it been possible to represent in familiar
terms the manipulation of documents, like copying, moving, formatting and so on
that could not have be done with the older technology?
Composite metaphors and multiple mental models
Designers have got round this problem by developing composite metaphors at the
interface (Carroll, eta!., 1988a). The desktop metaphor has been combined with
other metaphors to allow users the flexibility of carrying out a range of computer-
based actions. One example is the scroll bar. Such objects do not exist in real life but
they are metaphorical in the sense that they have been designed to capitalize on the
main feature of the concept of a scroll that is, a rolled-up document that has to be
unrolled to be read. In the same way, files can be 'unrolled' by moving a box up and
down a bar adjacent to the text file.
Other examples of interface metaphors based on objects that have been
combined with the desktop metaphor are menus and windows. Both have their own
metaphorical basis that is distinct from the concept of an office or a desk. Another
type of metaphor that has been used is based on a prototypical activity associated
with a professional practice. For example, the 'cut' and 'paste' actions used for
moving and copying text is based on the process used in page layout in the printing
profession.
From a cognitive perspective, it might be assumed that people would have
difficulties with interpreting composite metaphors. In most instances, though, it
seems that people can readily assimilate differing concepts and develop multiple
mental models. The idea of moving around in a file, by scrolling through a window
and selecting items from a menu attached to the top of the window by a bar, appears
to create few conceptual problems. In fact, most people do not actually think about
what they're doing in terms of the various metaphors (Tognazzini, 1992). Instead,
they just interact with the system, thinking in tenns of windows, menus, icons, scroll
bars and so on as if they were everyday terms.
Needless to say, there are some poorly designed composite metaphors that can
cause conceptual problems to their users. One of the main problems is the mismatch
between the user's expectation of what an interface object should and should not do,
based on their previous knowledge, and what the interface object actually does.
Question 7.3
What is a composite metaphor?
EXERCISE
As part of the desktop metaphor it is common practice to have an icon of a waste
basket that serves other functions besides those for much it is conventionally used
148 Interface metaphors and conceptual models
(as a container for discarded objects). One of these is to represent the place where
disk icons are put in order to eject the corresponding disk from the disk drive. This
implies that one has to 'throw away' a disk in order to retrieve it. Do you think that
these contradictions can cause conceptual problems, and if so can you suggest
another way of representing this operation at the interface?
A counter-argument against conceptual confusion is to treat such problems as actual
learning experiences. For example, Carroll et al. (1988a) describe how they
observed a user trying to tear off a sheet of paper from a stationery pad icon by
dragging the cursor across an icon representing the stationery pad. The interface had
not been designed to allow this action; sheets of paper could only be selected from a
menu option. However, what Carroll et al. noted was that the invalid action enabled
the user to understand better the difference between menu-based and mouse-based
operations. Accordingly, she developed a more elaborate mental model of the
interface.
Paradoxically, it could be the case that the more unexpected and the more
bizarre the interface metaphor is, the more likely the user will develop a better
understanding of the system. By carrying out inappropriate actions the user is
enlightened as to what are the appropriate and perrnissable operations. Provided the
user is able to experience the 'ah-ha' phenomenon relatively easily, and not be
humiliated or frustrated in the process, such a strategy may be very effective.
However, where the interface metaphor contravenes deep-rooted expectations ( cf.
the voice mail system and the central heating control models discussed in Chapter 6)
users may find it difficult to switch their models and adapt to the new way of
understanding how an object works. Furthermore, if the interface metaphor elicits
inappropriate actions that have undesirable consequences, such as unexpectedly
deleting files, the users may be reluctant to experiment further with the interface and
hence never learn the full functionality of the system.
The dividing line between what consititutes a good or a poor interface metaphor
is by no means clear cut. But by being aware of users' expectations about different
objects' behaviour in conjunction with having a thorough understanding of the
system's functionality and what aspects users are unlikely to understand about the
Ubiquitouscomputing 149
system, designers can begin to have a better understanding of what metaphors are
most appropriate (Frikson, 1990).
of interface metaphors for applications
The desktop metaphor and its composites have been the most successful and
pervasive of all interface metaphors. There are other metaphors, however, which
have been developed for applications other than information systems (see Chapter
22). Table 7.1 presents some examples with their associated applications.
An important consideration when searching for interface metaphors is the
appropriate effect. This is the subjective and emotional impact that different
graphical representations can convey (V erplank, 1988). The kinds of metaphors like
spreadsheets and multi-agents may be very appealing to adults working in office
environments, but may be inappropriate for schoolchildren. Imaginary characters
like demons and wizards might be far more attractive. Some sketching techniques for
helping you to explore different metaphors by brainstorming are discussed in
Chapter 22.
Table 7.1 Examples of applications and associated metaphors.
Application area
Operating environment
Spreadsheets
Object-oriented environments
Hypertext
Learning environments
File storage
Multimedia environments
Computer supported
cooperative work
Metaphor
The desktop
Ledger sheet
Physical world
Notecards
Travel
Piles
Rooms (each associated with a
different medium/task)
Multiagents
cr4 .. '1. Ubiquitous computing
Familiar knowledge
Office tasks, file management
Columnar tables
Realworld behaviour
Flexible organization of
structured text
Tours, guides, navigation
Categorizing objects in terms of
urgency, projects and so on
Spatial structure of buildings
Travel agents, butlers and other
serving roles
Ubiquitous computing is a tenn coined by Weiser (1991) for 'invisibly enhancing
the world that already exists' (p. 61). The ultimate aim of ubiquitous computing is to
make the interface metaphor invisible to the user in the same way as computer
150 Interface metaphors and conceptual models
Batteries
Control
button
Infrared
light-emitting
diodes
Micro-
processor
Figure 7.2 An example of ubiquitous computing: the 'tab' with an invisible interface
metaphor (Scientific American, September 1991, p. 69: photograph by courtesy of Matthew
Mulbry).
systems are invisible in home appliances, such as the VCR, the microwave oven and
the washing machine. Just as we use these tools unconsciously and effortlessly to
accomplish our everyday tasks (with the exception of the VCR, whose interface has
generally been all too visible and difficult to use) it is envisaged that ubiquitous
computing systems will be effortless to use.
An example of a ubiquitous system is the 'tab' which is a tiny networked
computer that is intended to be worn by the user, shown in Figure 7.2. The tab or
'active badge' can identify itself throughout a building, making it possible to keep track
of the person it is attached to. This can be very useful for buildings with high security
areas in so far as the person wearing the tab can enter secure areas without having to
remember a password and then key it in, as the tab does it automatically. Among other
opportunities such automaticity affords are: on entering a room people can be greeted
by name, telephone calls can be automatically forwarded and computer terminals can
change to the customization specified by the person sitting at it.
Negreponte (1989) also discusses the idea of invisible computers that
intercommunicate with each other with the purpose of doing all the chores in life.
Conceptual models 262
Instead of you having to decide that you need to buy some more milk, or that the
heating needs turning up or the trash needs emptying, a 'society of objects' in the
form of virtual butlers, secretaries and housekeepers will organize and manage
everything for you! However, a problem with such extensive anthropomorphism is
that users may assume that the system is more intelligent than it is. When the virtual
agents fail to behave as expected users may get frustrated. There is also the danger
that designers could mislead users in undesirable ways.
Question 7.4
What is ubiquitous computing?
Conceptual models
Conceptual models is the generic tenn that describes the various ways in which
systems are understood by different people. Primarily these consist of (i) the way
users conceptualize and understand the system and (ii) the way designers
conceptualize and view the system.
As we said in Chapter 6, whether interacting with devices, machines, computers,
people or the physical world, people use their prior know ledge to develop mental
models to enable them to understand and predict their behaviour. A highly
successful approach in interface design is to capitalize on users' existing knowledge
and the use of metaphors. However, the problem confronting designers who follow
this approach is finding a suitable metaphor. The aim for designers is to help users to
develop accurate mental models of the system. As Donald Norman (1986, p. 46)
puts it, 'The problem is to design the system so that, first, it follows a consistent,
coherent conceptualisation - a design model - and, second, so that the user can
develop a mental model of that system - a user model - consistent with the design
model'.
Users' models, design models and the system image
An important consideration of conceptual models is the relationship between
designers' models - the design model - and users' mental models - the user's model
shown in Figure 7.3. As most designers work in teams, it is more accurate to consider
the design model as the product of a collection of individuals rather than the outcome
of any one individual. Ideally, the user model should map onto the design model. That
way the users will be able to use the system's full capability as intended by the
designer. However, in the real world this does not often happen. More often, users
only develop a partial mental model of the design model. Their understanding and
ability to use the system, therefore, is limited. Another problem is that the design
model may be inappropriate for \\bat the user wants to achieve. In this situation the
152 Interface metaphors and conceptual models
Figure 7.3 The design model, the user's model and the system image (Nonnan and
Draper, 1986, p. 46).
users are forced to develop a mental model that is unfamiliar to them. A mismatch
can also arise if the design model is ambiguous, inconsistent or obscure.
Generally, the way users get to find out about the design model is through the
interface, its behaviour and the documentation. Collectively, these are called the
system image as shown in Figure 7.3. A large part of the accessible system image
comprises the physical interface (that is, the knobs and dials or images on a screen).
It is important to bear in mind, however, that the system image also includes the
system's behaviour, that is, the way it is used (for example by pressing keys, moving
a mouse, and so on). The sequences of operations required in using the system
(such as commands or menu selections) and the resulting events are all part of the
system's image, and users learn not just from looking at the system, but also from
their experience of using it.
If the system image is not able to convey to the users the design model in a clear
and obvious way, then it is likely that the users will develop incorrect mental models.
Consequently, they will experience great difficulties in understanding the system,
using the system and knowing what to do when the system doesn't behave in the way
they assumed it would. Much frustration, time-wasting and error-making can result. In
Section 13.7, we shall extend these concepts further in relation to the design of direct
manipulation systems. In particular, we shall consider what happens when the user's
model does not match the designers' model well, creating 'gulfs' between the two,
and how designers can help to prevent this happening.
Question 7.5
(a) What is the difference between a design model and a user model?
(b) What is the system image?
(c) . Why do mismatches occur between the design model and the user's model?
Key points
Further reading 153
Verbal metaphors are analogies based on familiar knowledge, spontaneously
elicited or used in written or spoken instructions, which help users begin to
understand new systems.
Interface metaphors combine a familiar domain with the system structure to
make a concrete system image.
Composite interface metaphors are a combination of multiple and partial
models of familiar domains.
There are several kinds of intelface metaphors, although the most connnon one
is the desktop.
Users and designers both have conceptual models of the system; Norman has
called these the user and the design model, respectively.
A main goal in interface design is to develop a system image that maps the
design model onto the user model.
Ubiquitous computing systems have invisible interfaces, are interconnected and
intended to be effortless to use.
A conceptual model provides a framework from much to consider how to design
appmpriate interface metaphors.
Further reading
Metaphors and language
LAKoFF G. and JoHNsoN M. (1980). Metaphors We Liue By. Chicago: University of
Chicago Press.
A wealth of books are available in the linguistic, psychological and
philosophical literature, but this is one of the most readable, lively and
thought-provoking books, much has now become a 'classic'.
Interface metaphors
CARROLL J.M., MAcK R.L. and KELLOGG W.A. (1988). Interface metaphors and user
interface design. In Handbook of Human-Computer Interaction (HELANDER M.,
ed.) , pp. 67-85. Amsterdam: North-Holland.
ERicKSoN T.D. (1990). Working with interface metaphors. In The Art of Human-
Computer Inteiface Design (LAUREL ,B., ed.) . Reading, MA: Addison-Wesley.
These two chapters provide several useful suggestions on how to generate and
design interface metaphors.
154 Interface metaphors and conceptual models
Conceptual models
NORMAN D. (1986). Cognitive engineering. In User-Centred system Design (NoRMAN D.
and DRAPER S., eds) , pp. 3 1-61. Hillsdale, NJ: Lawrence Erlbaum Associates.
Although dated and rather general, Norman's introductory chapter on cognitive
ergonomics includes an account of his original framework of conceptual
models.

Das könnte Ihnen auch gefallen