HCI approaches centering on designing for experience rather than task completion attempt to address a need for computational artefacts that are enjoyable, fun, appealing or meaningful. Tracing historical influences on computer science and user interface design, we highlight some limitations of resulting scientific approaches to design for experience. An example of an information display based on the concepts of ‘informative art’ illustrates a means of overcoming some of the limitations through a more holistic approach.

INTRODUCTION

Fun, enjoyment and other emotional aspects of experience are goals of design not well accounted for within the traditional scientific/cognitive methods which are available to HCI practitioners today. Recent developments in the design of user interfaces towards experience-centred design recognise a need to go in a direction that is different from the production of plain efficient and usable interfaces.
Despite the increased and sometimes subtle integration of computers within everyday life and diverse spaces and places, the majority of devices continue to resemble largely the IBM PC of the 1980s. Computers, as physical objects, are quickly identifiable as serious machines; often unattractive and  consequently hidden away in workrooms and offices. Aspects like fun, enjoyment and pleasure are frequently lesser concerns in the construction of such artefacts and not a part of their design goals.
From a technical viewpoint, the interface defines how a product achieves a  particular task, determining what the user is able to do and signifying how the system reacts to user actions [10]. Such technical assumptions form the basis for many approaches to interface design, drawing on concepts from cognitive science in order to maximise functionality and efficiency of a product. Accordingly, user interface design is most commonly approached as an engineering discipline in which the task of developing an interface is reduced to a problem-solving process which is done by applying a set of well-prescribed steps [13].
Fallman [3] refers to this way of creating interfaces as the conservative account. "Here design is thought of as a scientific or engineering endeavour, borrowing methodology and terminology from the natural sciences, mathematics and systems theory, drawing on a philosophical base in rationalism".
Within this approach to design, user experience is reduced to usability [13]. Design is restricted to the intelligent part of humans, failing to respect the whole person as user. Simultaneously, emotional aspects are often narrowed down to superficial outcomes of fun and excitement, thereby sidestepping deeper and more relevant issues [8].

WHY ARE WE HERE?

Sengers has suggested that our current ideas of computers and user interfaces are rooted in the industrial revolution which took place around the beginning of the 20th century [12]. At that time, concepts like automation and massproduction were used to optimise costs and production processes. People skilled in crafts were collected into factories; their work was standardised and split along a soulless mechanical production line. The invention of the scientific management approach by Frederic Winslow Taylor and the introduction of the assembly line placed greater emphasis on maximising the efficiency of human labour. Workers were observed and analysed to extract optimal techniques for achieving a task and eliminating any wasteful motions [12]. Such efficiency, however, was to the detriment of the overall experience of workers who were left with little individual or collective agency and no opportunity for creative activity.
The influence of these scientific practices on modern approaches to computer science and user interface design is evident in the guise of computational taylorism, whereby processes are broken down into small tasks which are analysed and performed by optimised algorithms. Models are built to represent reality at an abstract level and software is produced based on such models. In this process, everything that is unnecessary is removed from a product [12].
This way of solving computational problems is also applied to user interfaces. Here too, everything that is unnecessary is removed and interface elements are arranged in such a way as to allow the user to work with a product as efficiently as possible.
Widely recognised proponents of usability suggest the benefits of such functionally adept interfaces for reducing stress and frustration in use [10]. However, the wider reductionism of human endeavour represented in such design is dehumanising and limiting in the same way as the automatic loom must have been for the master weaver.  

THE NEED TO DESIGN FOR EXPERIENCE

The importance of a computer functioning and appearing to function efficiently within working environments - where we are paid to complete a task as efficiently as possibility - does not preclude the need for workers to be satisfied in and gain enjoyment from their jobs. Working life is often overvalued and separated out from leisure time in many cultures. Nonetheless, concepts from workplaces make their way into private life [12]. Tayloristic concepts such as time management and to-do lists are employed to make sure that we are equally as efficient at home as we are expected to be at work [12]. This is clearly reflected in the computers and software programs we use in everyday private life, which hardly differ from those machines and programs we use at work.
By bringing computers and software into private life that are more enjoyable to use, it is also likely that working with such products can become a relaxing task. After all, it is increasingly suggested that work tasks are improved if they are enjoyable [1].

A MODEL OF EXPERIENCE

Design for experience recognises the need for a good overall experience, especially for computers and software we use in private lives. But how can we achieve experience? Is it possible to build experience right into a product? As computer scientists, we often create formal models of the reality in order to conceptualise the real world [12]. We than implement our computer programs based on these models, hoping that they will reflect the real world. Hence, it appears to be a good idea to have a model of user experience that can be used to build products which provide a particular set of experiences.
A number of researchers have proposed such models of experience. Hassenzahl [5] for example suggests that a product carries particular features, like content, presentation, functionality and interaction. These features lead to a certain product character that in turn leads to a number of consequences for the user. For example, a user might feel pleasure when using a product because it's character provokes a nice memory. Accordingly, different character attributes are said to lead to different consequences. The assumption behind this is that that designers can shape a product's character by providing it with certain features. However, the character of a product as it appears to the user is not always the character that was intended by the designer. Hence, the consequences for a user are not always the same.
The model suggests that the ‘apparent’ character of a product is constructed by users in in a usage situation. To justify this the user is constructed as working in one of two modes in which they are either primarily interested in completing a particular task as efficiently as possible (‘goal mode’) or merely wants to ‘play around’ and determine goals on the fly (‘action mode’).

PROBLEMS WITH MODELLING EXPERIENCE

Models like Hassenzahl’s, while well argued and structured, are a typical scientific response to the apparent problem of experience. It belongs firmly to the conservative account, in attempting to provide a foundation on which designers can more or less predict the experience a user will have.
Different users of a product will have different experiences, depending on a number of factors. One of these factors, as Hassenzahl mentions, is the situation in which a product is used. The model reduces this aspect basically to two different modes a user may operate in. The fact that a user might work in both modes at the same time or in some other mode is not considered. Further, the use of technical terms like ‘modes’ locates such work in a tradition of seeing the user from a rather technical point of view. Again, the fact that we are humans is left aside and the user is reduced to an entity able to function in one of two definable ways. 
This is not the only problem that such models present. Experience is also subjective in a way that users construct experience by sense making, a reflexive and reflective process. Other authors [4, 9, 14] maintain on the basis of aestheticphilosophical and psychological accounts respectively, that experience is not innate to, ready-made or built into a product. The process of sense making is instead individual, continuous and directly incumbent upon the relationship between self and object. Experience also changes over time as people reflect about a product and themselves. By suggesting that experience can be built into a product by giving it a certain character, conservative models play down this nature of experience. In doing so, tayloristic ideas are clearly visible, "as we try to clean up, formalise, and organise what is an inherently messy and perhaps fundamentally incomprehensible problem" [12].
Despite such criticisms, Hassenzahl’s model of experience is among many other emerging approaches that may be classified as equally ‘scientific’. For example, Norman [7] identifies three levels of experience: the visceral, the behavioural and the reflective. His assumption is that the levels, although they interact, can be treated independently by the designer. As with Hassenzahl's model, Norman's approach shows a reductive tendency. Users are classified within categories of predictable behaviours, disregarding “the wealth of experience brought to the interaction by a person's prior experience and individual way of being, as well as an object's meaning laden history and the uniqueness of a situation" [4].
Others suggest the use of heuristics to provide designers a guidance for experience. Sengers for example, argues that we "must realise that we cannot fully represent experience within the software, and instead try to set up more nuanced relationships between (internal, formal, clean) code and (external, messy complicated) experiences" [12]. While the proposed heuristics are rather vague and do not represent a concrete approach, they nonetheless give us a starting point for aspects to be considered when designing products for experience.

CHANGING THE PERSPECTIVE

Engineering approaches to experience don't work well. Heuristics are a first step but they are not really suitable to get a grasp on experience. As Wright et al point out, we need a more holistic approach that accepts the nature of experience and does not try to formalise it [13].
To this end, Ljungblad et al have attempted to design a weather forecast display to be not only informative, but also enjoyable for viewers [6]. The concept they came up with is based on informative art. Informative art is described as "computer augmented, or amplified, works of art that not only are aesthetical objects but also information displays, in as much as they dynamically reflect information about their environment"[11].
Ljungblad et al based their work on the paintings of Piet Mondrian, an artist from the Netherlands who lived from 1872 to 1944. His paintings appear trivial but are actually complex representations. Mondrian was known to be a contributor to the De Stijl art movement, also known as neoplasticism. Proponents of this style sought to express a utopian ideal of spiritual harmony and order. Paintings in this tradition reveal a reduction to the essentials of form and colour. Compositions are simplified to horizontal and vertical directions and only the primary colours are used.
Figure 1 shows the weather display which was designed by Ljungblad et al. Although this display does not immediately reveal the information, it is quite simple to understand and to read. Each square in the picture represents a day. Starting with today in the top left corner, the display is read western-style, from left to right and top to bottom. The size of each square represents the temperature: the higher the temperature, the larger the square. The three primary colours are used to represent the weather condition. Yellow means sunny, blue means rainy and red represents clouds. Note that today is included in the display in order to give the viewer the possibility of comparison with the following days [6].

Figure 1. A four day weather forecast in the style of Piet Mondrian [6]

The weather forecast display was installed at the University of Göteborg and students were asked about their experiences with the artefact. The interviewees were divided into to two groups. One group received a short introduction about how the display works. This group reported that it was easy to get the information and that it is "an interesting and different way to show an uninteresting weather forecast" [6]. Students also said that they were quite pleased looking at the display. It blended nicely into the background and was well integrated with the environment.
Students who did not receive an introduction, however, reported different impressions. None of these students was able to read the display. They didn't even have an idea of what kind of information was presented. Most students recognised it solely a piece art and some wondered why a computer display was used to display a static painting. They hardly recognised that the display changed at all. Nevertheless, most of the students "who did not understand the information could still appreciate the piece as a decorative item” [6].
The Mondrian-style weather display does not visualise information in the most efficient way. It is perhaps unsuitable for situations where many people need to get precise information quickly (e.g. at airports or railway stations). This is related to the fact that the approach does not work without providing the context to viewers. We are simply not used to reading the weather this way and we don’t expect a painting to provide such information.
Nevertheless, the display is quickly describable and understandable. It is therefore suggested that pieces of informative art are more suited to be installed in private areas, such as living rooms, or in bureaus where they not only display information but also beautify the environment [6].
Such work is interesting to highlight if only for the way in which it illustrates that experience can be achieved in a non-formal way. Instead of falling back to traditional tayloristic/scientific ideas, other disciplines, such as art, can be used to inspire interface designers.
Ljungblad and colleagues did not seek to predict emotions but accepted the subjective and changing nature of experience. They nonetheless wanted to create an artefact that provokes emotions. By taking the romantic account wherein the product is a functional piece of art [3], they based their work on an artistic style, the De Stijl, and on the work of individual artist, Piet Mondrian.
Art is all about emotions. The rational cannot adequately describe what we feel when we experience an aesthetic object. Ljungblad et al changed their perspective away from traditional science and used this nature of art to create an information display that enables engagement in a way that precludes neither experience nor functionality.

CONCLUSION

Computer science is heavily influenced by the ideas of the industrialisation and taylorism. Computers are clearly perceived as serious machines and the software that runs on those machines is optimised for efficiency of use. But as users, we are human beings and have needs which are aside from or go beyond usability. Especially in our private life, we want to have fun and enjoy ourselves when working with a computer, but we also want to reflect and to dip ourselves in the emotional enjoyment of an aesthetic experience [2]. Design for experience aims to face this challenge by trying to create products in such a way that they provide an overall experience and not only good usability. However, traditional computer science has problems in finding a suitable approach. The scientific/tayloristic ideas of reduction and model building don't work well because they don't accept the subjective and changing nature of experience.
Looking at other, less scientific disciplines can provide a valuable inspiration for creating experience. The example from informative art has shown how a weather display can be designed in way that is not only informative but also appeals to our sense of aesthetics. The designers accepted the nature of experience. They created a product that supports experience by provoking and encouraging emotions without trying to predict or control them.

REFERENCES

1. Blythe, M. and Wright, P. From usability to enjoyment - Introduction. In Blythe, M.A., Overbeeke, K., Monk, A.F. & Wright, P.C (Eds.) Funology - From usability to enjoyment. Kluwer Academic Publishers, Dordrecht, Netherlands, 2003, XIII-XIX.

2. Dewey, J. Art as experience. Capricorn, USA, 1934.

3. Fallman, D. Design-oriented human-computer inter-action. In Proc. CHI 2003. ACM Press (2003), 225-232.

4. Fiore, S.G., Wright, P. & Edwards, A. A Pragmatist Aesthetics Approach to the Design of a Technological Artefact. In Proc. Critical Computing 2005. ACM Press (2005), 129-132.

5. Hassenzahl, M. The thing and I: Understanding the relationship between user and product. In Blythe, M.A., Overbeeke, K., Monk, A.F. and Wright, P.C (eds.) Funology - From usability to enjoyment. Kluwer Academic Publishers, Dordrecht, Netherlands, 2003, 31-53.

6. Ljungblad, S., Skog, T. and Holmquist, L.E. From usable to enjoyable information displays. In Blythe, M.A., Overbeeke, K., Monk, A.F. and Wright, P.C (eds.) Funology - From usability to enjoyment. Kluwer Academic Publishers, Dordrecht, Netherlands, 2003, 213-221.

7. Norman, D.A. Emotional Design: Why we love or hate everyday things. Basic Books, New York, USA, 2003.

8. Overbeeke, K., Djajadinigrat, T., Hummels, C., Wessels, S.W. and Frens, J. Let's make things engaging. In Blythe, M.A., Overbeeke, K., Monk, A.F. and Wright, P.C (eds.) Funology - From usability to enjoyment. Kluwer Academic Publishers, Dordrecht, Netherlands, 2003, 7-17.

9. Petersen, M. G., Iversen, O. S., Krogh, P. G. & Ludvigsen, M. Aesthetic Interaction – A Pragmatist’s Aesthetics of Interactive Systems. In Proc. ACM DIS2004, 269-275.

10. Raskin, J. The humane interface: new directions for designing interactive systems. Addison-Wesley, Mass., USA, 2000.

11. Redström, S., Skog, T. and Hallnäs, L. Informative Art: Using amplified artwork as information displays. In Proc. Designing Augmented Reality Environments 2000, ACM Press (2000), 103-114.

12. Sengers, P. The Engineering of Experience. in Blythe, M.A., Overbeeke, K., Monk, A.F. and Wright, P.C (eds.) Funology - From usability to enjoyment. Kluwer Academic Publishers, Dordrecht, Netherlands, 2003, 19-29.

13. Wright, P.C., Blythe, M.A. and McCarthy, J. The idea of design in HCI. http://www-users.cs.york.ac.uk/ ~pcw/papers/DSVIS 20paper18_08.pdf

14. Wright, P.C., McCarthy, J. and Meekison, L. Making sense of experience. In Blythe, M.A., Overbeeke, K., Monk, A.F. and Wright, P.C (eds.) Funology – From usability to enjoyment. Kluwer Academic Publishers, Dordrecht, Netherlands, 2003, 43-53.

Published in:

User Experience - Towards a unified view, Conference Workshop, NORDICHI 2006.