Keith V. Nesbitt, Charles Sturt University, knesbitt@csu.edu.au
Robert Shen, University of Sydney, xshen@cs.usyd.edu.au

Abstract

In the modern world of electronic data and global communication information displays can change quickly and so require constant attention from users. However, the nature of the modern world is also that users are often required to perform tasks in parallel and there may be competition for the user's attention. Ambient displays are information displays that are designed to function at the periphery of a user's attention. Unlike alarms, ambient displays should not interrupt a user from their current task. Changes in the display may, or may not be relevant to the user as the display is usually designed to show non-critical data. Because ambient displays are not critical to short term tasks, it may be difficult to develop a business case for using an ambient display. The questions that are most difficult to quantify are: "How to design a good ambient display?" and "When is an ambient display good?". This paper addresses these two questions by describing our research with a particular ambient display called; "Butterfly/Dragonfly". This display is designed to show stock market data and provide the user with short-term market trends for a particular stock. The application domain and implementation platform are described. The design of the specific display is described in the context of "good" design criteria. Finally a general framework for evaluating ambient displays is introduced and the preliminary results from an extended field trial are discussed.
 

Introduction

Ambient displays are: "aesthetically-pleasing displays of information which sit on the periphery of a user's attention"(Mankoff et al., 2003). They reside, but do not intrude upon the user's normal location (public, work or home) and are typically designed to display dynamic information that is non-critical to the user's main tasks (Grass et al., Ishii et al., 1998). The related term, ambient media was originally used by Ishii (Ishii et al., 1997) to describe background displays that form part of a seamless interface between people, digital information and objects in the world (people, bits and atoms). There is a notion that ambient media occupies the background, and so contrasts with other information that is present in the foreground of our attention (Skog et al., 2003). Ambient displays therefore, use ambient media to support the ability for people to do a number of tasks in parallel while also being able to shift and focus their attention on a single task (Ishii et al., 1998).

This idea of foreground and background attention, or that users have a central focus within the context of peripheral information has been explored in the domain of Information Visualisation. In this domain displays may be designed to support "focus + context" (Card et al., 1999; Lamping et al., 1996; Rao et al., 1994). Human vision is composed of two parts: foveal and peripheral vision (peripheral version, accessed on May 2006). Foveal vision has a narrow angle (that is, around 10 degrees) but also supports higher visual acuity. In contrast, peripheral vision covers a wider angle but also has lower visual acuity. At a physiological level, Y-cells in the peripheral part of the retina respond directly to movement (Goldstein, 1989). Changes in peripheral scenes are also known to cause rapid shift of foveal attention in order to investigate the change.

Early work by Weiser also focuses on peripheral vision (Weiser, 1991). Weiser believes that “the most profound technologies are those that disappear. They weave themselves into the fabric of everyday life until they are indistinguishable from it” (Weiser, 1991). Weiser proposed the idea of ubiquitous computing which would in part inform the ambient display work of Ishii (Ishii et al., 1997). More recently, Weisser's idea of making the computer more "transparent" was also explored by Streits (Streitz et al., 2005) who described the two main trends of the disappearing computer as mental disappearance and physical disappearance. Mental disappearance occurs when users no longer notice computers because they have been so smoothly integrated into the environment. Physical disappearance refers to the shrinking size of computer components (for example, using PDAs to replace larger desktop computers).

Another field that is related to the idea of ambient displays is augmented reality (Lamping et al., 1996). Augmented reality tries to augment human perception by seamlessly integrating computer generated data in to the real world (Wellner et al., 1993). Because ambient displays describe a type of user interface, research into user interfaces, particularly research involving design and evaluation is also relevant. However, may techniques in traditional interface design focus on specific user goals and tasks. Previous researchers have noted the difficulty of clearly defining the less tangible goals of ambient displays (Mankoff et al., 2003).

The relevance of information visualisation to this field has also been previously described (Skog et al., 2003). Design of information visualisations and ambient displays both require decisions about some mapping between abstract data and a visual representation. However, there is no restriction of ambient displays to the visual sense and examples can be found that rely on visual, auditory and haptic displays.  The digital family portrait (Mynatt, et al., 2001) is a visual display which uses the density of icons in a band to show a person's regular activities. Musicbox (Ullmer, 2006) plays music in response to activity around a remotely located piano. The haptic door knob (MacLean et al., 1999) can display both torque and thermal information. In fact the haptic door knob (haptic & sound), the musicbox (visual & sound) and the lumitouch (Chang et al., 2001) (visual & haptic) utilise a range of senses. Even in his early work, Ishii noted that ambient displays may benefit from considering a range of sensory modalities (Ishii et al., 1998). This implies that the many factors associated with the design and evaluation of multi-sensory displays (Nesbitt, 2003) is also relevant when designing ambient displays.

"How to design a good ambient display?" and "When is an ambient display good?" are two key question our work addresses. These two issues are interconnected as evaluation criteria can also drive design decisions. Indeed design criteria, evaluation techniques and the process that ties them together are often interconnected parts of a design methodology (Nesbitt, 2003).

Having introduced some of the key ideas of ambient display we will now consider some of the issues related to the design of ambient displays. We follow this by discussing the design of the Butterfly/Dragonfly display for stock market data. Next we discuss evaluation of ambient displays and present the results of a three month evaluation of the Butterfly/Dragonfly display.

Designing Ambient Displays

A number of projects have been designed and implemented in the field of ambient displays although most, such as, the digital family portrait (Mynatt et al., 2001), musicbox (Ullmer, 2006), and lumitouch (Chang, 2001), have largely been used within research settings. The ambient weather forecast umbrella (Ambient-Devices, accessed on May, 2006) and the ambient orb are examples of ambient displays that have been commercialized.

Grasso et al. suggest a range of dimensions to consider in the design of ambient displays: collection of information; processing; presentation and; user interaction (Grasso et al., 2003). The collection of information and the processing are relatively simple software engineering tasks for most systems. The presentation of the data and the interaction are the most difficult, as with many interface designs, "fundamental issues of attention, perception, and mental representation come into play" (Wisneski et al., 1998). Ambient media is also likely to have a learning effect, with a person's perception changing over time (Ishii et al., 1997).

The sensory modalities used by the display need to carefully consider the persons foreground task (Wisneski et al., 1998). A good ambient display should be "seamless" and should "allow for a smooth transition between the users' focus of attention between background and foreground" (Ishii et al., 1997).

The context of the ambient display may also be a factor because an ambient display can be designed for public places or for more personal spaces such at home or work (Grasso et al., 2003). In public places the designer may need to consider social impact and factors such as: "value, emotion, privacy and trust" (Theofanos et al., 2005). The degree to which the display is passive or interactive also impacts on the design (Grasso et al., 2003). User interface concepts such as, affordance (Wisneski et al., 1998) and metaphor (Ishii et al., 1997) may also be relevant in some ambient display applications.

We base our design on the four key considerations of: distraction, comprehension, usefulness and interoperability (see Figure 1). Distraction concerns the way the display shifts the user's attention between foreground and background tasks. Distraction is itself decomposed into two parts: display-distraction and self-interruption (Shen et al., 2005). Comprehension concerns the level of information understood by the user. Usefulness refers to the suitability of the information being displayed. Interoperability concerns how well the display, the user and the environment operate as a whole. These four key considerations are used not only for design but also form the basis of our evaluation procedure.

Figure 1.  The key considerations of design are, Usefulness, Interoperability, Comprehension and Distraction. These criteria are also used to evaluate the ambient display.

Butterfly/Dragonfly Display

The goal of our application is to design a piece of informative art (Redstrom et al., 2000) that represents a short term trend in prices for a single share traded on the Australian Stock Exchange. The display is designed for a private office space (see Figure 2), and so has a single user The single subject is an experienced information visualization researcher with background knowledge on stock market trading and ambient displays. The subject is also a painter and one of their paintings, called "Information Metaphor" was used as the basis for the ambient display (see Figure 3). The office environment also includes a number of other paintings in the same style.

Figure 2. Ambient display installed in the users office environment.

Paintings on a wall are not designed for interaction except in a passive way. The have limited affordance except to suggest they should be looked at. In general paintings have an aesthetic quality and they are at home in both public and private spaces. For designers of information they offer a wide range of styles and metaphors on which to map information, although we might expect a picture to remain fairly static in appearance.  The idea of turning paintings into ambient displays has been called informative art (Redstrom et al., 2000). For example, paintings in the style of Mondrian have been used to show email traffic, weather forecasts and bus timetables (Skog et al., 2003). The InfoCanvas application (Miller et al., 2001) allows users to design their own paintings using simple electronic paint tools.
 

Figure 3. The ambient display is based on a painting by the user called "Information Metaphor" (Nesbitt, 2003).

The Butterfly/Dragonfly display uses this same notion of passive, informative art to try and minimize the distraction to the user. Because the picture was painted by the user it expected to meet the user's aesthetic criteria. Paintings are a natural part of many offices and so it is expected the display will have good interoperability with the users work environment. To maximize the comprehension and usefulness of the display, the user was directly involved in specifying both the data and the display metaphor. A face-to-face interview was conducted before actual implementation to confirm the customized content of the display, namely the data source and the visual metaphor.

The stock market data being displayed is the trading price for BHPBilliton (BHP, 2006). BHPBilliton is a blue-chip stock that has a high turnover on the Australian Stock Exchange (ASX, 2006). Real-time BHP stock information is the data source in this system. This data is fetched from the server by a RSS feed every minute.

The display shows four pieces of relative pricing information (Table 1). A butterfly (good) indicates that the price has risen. A dragonfly (bad) indicates that the price has fallen. The display is read from left to right, with four butterfly/dragonfly symbols being shown. The left-most symbol shows the comparison between the current stock price and the previous day's closing price. It is updated every two minutes as trades occur on the market. The next symbol is more static, and shows the comparison between yesterday's closing price and the previous day's closing price. The third symbol compares the closing price from two days ago with the price from three days ago. The final, right-most symbol shows a comparison between the closing price from three days ago with the closing price from four days ago (Table 1, Figure 4). Weekends are ignored and only trading days are considered in the calculation. The display can be in only one of two states, either the market is open or closed. When the market is closed the picture becomes blurred and out-of-focus and no butterflies or dragonflies are visible (Table 2).     

Table 1. The symbols displayed for different market conditions

Table 2. The display state when the market is opened or closed.

The user of this display is interested in monitoring short-term fluctuations of the stock price to find possible patterns. This type of market analysis is sometimes called charting or technical analysis and makes the assumption that price discounts other factors in the determination of trading trends (Nicholson, 1999).

There are two types of complementary analysis, which are used to trade on the stock market. They are known as technical analysis and fundamental analysis (Nicholson, 1999). Fundamental analysis studies the underlying factors that determine the price of a financial instrument. For example, factors such as, a company's profit, market sector, or potential growth can influence the share price. These factors can be considered against more global considerations such as the general economic trend. Fundamental analysis is the more traditional form of analysis used to trade the market. Technical analysis is defined as "the study of behavior of market participants, as reflected in price, volume and open interest for a financial market, in order to identify stages in the development of price trends" (Nicholson, 1999). Unlike fundamental analysis, technical analysis ignores the underlying factors that determine price and makes the assumption that the price of a financial instrument already quantifies all these underlying factors. Technical analysis is based on patterns that can be found directly in the data.

The design of this ambient display is based on technical analysis. Technical Analysts may look for patterns in the market over a range of time periods from minutes to weeks and months. They may carry out a range of tasks which include: identifying viable trading strategies based on patterns in the data; developing trading strategies over different time frames; identifying trends in the movement of prices; determining the balance of supply and demand of a financial instrument and; monitoring changes in the price of a financial instrument.

Figure 4. Combinations of butterflies and dragonflies indicate a ranging market.
 

Figure 5. Four butterflies indicate a downward trend over four days of trading.
 

Figure 6. Four dragonflies indicate a downward trend over four days of trading.
 

Figure 7. This figure shows a typical price chart of stock prices over time. Each bar represents one day of trading. The small stroke on the right of each bar indicates the day's closing price. The trend of butterflies and dragonflies that would appear in the ambient display is shown below the bars.

The intention of this display is not to assist in a defined trading strategy but to allow the user to monitor short term market trends and perhaps identify some useful patterns. Of course when market analysis is the user's primary task, other charting techniques contain more detail. However, these techniques also require greater effort and concentration. A further intention of this display is to make the monitoring and identification a background activity. For example, an up trend in prices over four days is shown by four butterflies (see Figure 5) and a down trend by four dragonflies (see Figure 6). During the evaluation these combinations rarely occurred and more frequently the display showed some combination of butterflies and dragonflies (see Figure 4). This type of fluctuating behavior is common in stock market prices (see Figure 7). If the number of butterflies and dragonflies are equally distributed over time then this would indicate a ranging market. While a predominance of butterflies over time might indicate a long term up trend and a greater number of dragonflies a long term downtrend.

Evaluating The Ambient Display

Compared to the design of ambient displays, little standardization has occurred in the evaluation of ambient displays. For example, there are no standard evaluation metrics and criteria available. Therefore it is difficult for inexperienced evaluators to compare displays and learn from each other.

A few ambient display systems have conducted evaluations during development. BusMobile and Daylight Display (Mankoff et al., 2003) are evaluated by heuristic evaluation, which is one of the most successful evaluation methods in ambient displays. The digital family portrait (Mynatt et al., 2001), is evaluated by combining wizard of oz with interviews and questionnaires to evaluate the form, content, and presentation techniques used for the portrait. InfoCanvas (Miller et al., 2001) conducts an in-depth longitudinal and in situ evaluation using a carefully designing questionnaire and interviews to collect subjective that can be analysed in a quantitative way.

To date, the main evaluation methods have relied on interviews and questionnaires. To better explore the evaluation of ambient displays, we systematically divide the evaluation of ambient displays into the two styles of intrusive evaluation and non-intrusive evaluation.

An intrusive evaluation is where the user is consciously aware of the evaluation process. This kind of evaluation often focuses on usability tests in a laboratory and runs for a short period of time. Most often, such experiments are conducted using mainstream evaluation methods in user interface design, such as; heuristic criteria; questionnaires; interviews and; the think aloud method.

By contrast, in a non-intrusive evaluation, the user is not consciously aware of the evaluation process. This focuses on the actual use of the display in the normal operating environment (in situ). The evaluation typically runs for a longer period of time compared to the short-term nature of intrusive evaluations. Currently, there are few evaluation methods that have been developed in this manner.

In this paper, we present a non-intrusive evaluation, which conducts the evaluation in a real life environment. It is expected to be more ecologically valid than an intrusive display as it conducted in a real life environment, which we assume has less impact on the results than an unfamiliar laboratory setting. It has the disadvantage that the display is designed for, used and evaluated by a single user in their specific work environment. Therefore, we cannot make generalized statements about the suitability of the display for a broader audience or operating conditions.

The evaluation of Butterfly/Dragonfly is conducted in a personal office with a single subject over a 3 month period. The individual subject has participated in the design of Butterfly/Dragonfly display and has defined the metaphor used in the display. 

The hardware used is one desktop computer with a special monitor. A wall-mounted, 19 inch monitor is disguised with its border covered by an old-fashioned picture frame (see Figure 2). There is no other input and output equipment involved in this evaluation.

The subject is involved in two tasks during the evaluation. The primary task is to concentrate on the subject’s own work. The peripheral task is to shift focus to the Butterfly/Dragonfly display and fetch information by a simple glance.

We have two hypotheses for the evaluation. Firstly that the greater the distraction the greater the comprehension. Secondly, we expect the subject's comprehension will improve over time.

There is no video or audio recording equipment installed in single subject office to avoid offending privacy issue. All the results are collected by questionnaire and interview. To avoid the impact on the awareness of Butterfly/Dragonfly display by these questionnaires and interviews, we set the interval between each questionnaire to 1 month. The Butterfly/Dragonfly display is temporally closed during each questionnaire and interview. Four evaluation attributes are measured in each questionnaire:  Distraction, Comprehension, Usefulness and Interoperability. The detailed questions for each specific evaluation attribute are described in Table 3.

The answers to the comprehension questions over the three monthly questionnaires are used to calculate the Mean Comprehension Rate (MCR). A larger MCR indicates better understanding of the ambient display. The Mean Comprehension Rate is seen to increase as time goes (see Table 4). This suggests that the longer the display is used the better subject comprehension becomes.

Table 3 - The questions used in the evaluation (ordered by design criteria)

Table 4 - Mean comprehension rate from three evaluations.

Table 5 - Mean display-distraction from three evaluations.

The answers to the distraction questions are used to calculate the Mean Self-Interruption (MSI) and the Mean Display-Distraction (MDD). A larger MSI depicts a higher curiosity from the subject. A lower MDD is indicative of a calmer display. Results for both display-distraction and self-interruption are shown in Table 5. The general trend of the Mean Display-Distraction is increasing but the Mean Self-Interruption fluctuates for the three evaluations.

For the two questions dealing with usefulness (Does our display reflect the state of information well? and Do you think our display is useful generally?) the subject answered “Yes” to both in all three evaluations. It indicates that the subject is satisfied with the usefulness of “Butterfly/Dragonfly” display. The typical comments for Q1 includes “I can easily see two trading states (open and close)” and “It is easy to detect the BHP stock price pattern from the display and it is useful for me to use it as a visual cue”. A typical comment for Q2 was “I like this idea and I want to have one at home” and “I think it’s enough for me to monitor the trend of BHP stock price”.

The results from the questions designed to check on interoperability (Do you think the painting style is suited to your office? and Do you think of the display as part of your furniture?) indicate the subject is satisfied with the interoperability of “Butterfly/Dragonfly” display. The answers for each question are “Yes” in each of the three evaluations. Typical subject comments on Q1 include; “I really like this kind of implementation. It makes my office more like an art gallery. Furthermore, I can get interesting information from this display……” and “My colleagues also like this idea”. For the second question, the subject comments that; “I treated the Butterfly/Dragonfly picture as a display at the beginning but now I think of it like it's just a part of the furniture of my office. It just looks like a picture on the wall but I can still find information in it.”

Discussion

Our work is generally motivated by two important questions: "How to design good ambient displays?" and "When is an ambient display useful?" The specific aims of this evaluation were to conduct a non-intrusive evaluation that explored the relationship between Distraction and Comprehension and to explore a new design style with a trade off between communicating information while maintaining aesthetics.

Distraction and comprehension are two important variables measured in our evaluation. The results suggest a relationship between comprehension and display-distraction while there is no apparent relationship between comprehension and self-interruption. This result is contrary to our initial thinking that the more times the subject is distracted by the display the better their performance on comprehension would be. One contributing reason to explain this is that display-distraction is caused by the data value changing in the display and this kind of distraction is easily noticed by the subject. On the other hand, self-interruption is affected by the characteristics of the subject. It may be that when the subject looks at the display no change has occurred, the display appears static and thus the subject comprehends no patterns. We make the conclusion that comprehension may be improved in ambient displays by adjusting the level of display-distraction.

The design style of this ambient display attempts to trade off communication with aesthetics. Communicating information is usually the general purpose of information visualization (Friedrich et al., 2002) while aesthetics is largely treated as a secondary goal of the design. However, aesthetics is a key issue in the design of ambient displays as the effective visual appeal of ambient display allows it to perform a dual role as decoration for an architecture space (Ishii et al., 1998). Furthermore, it may also draw attention of users to probe potential subjects in the future evaluations. Subjective feedback from the subject indicates that this display achieves both an aesthetic and information function. For example, the subject comments that; “It’s just like a picture on the wall but can supply me information (general trend) at a simple glance” and “It’s great because I can mainly concentrate on my own task”. This leads to our second conclusion that a trade off between aesthetics and communication of information can improve the usefulness of ambient displays.

Conclusion

In this paper, we have presented an ambient display system that presents real-time stock market data as a painting. We have discussed a number of design issues and also conduct a non-intrusive evaluation to evaluate this system. As a result of the evaluation we suggest two principles to assist with the design of ambient displays:

1) Comprehension can be improved in ambient displays by adjusting the level of display-distraction.

2) A trade off between aesthetics and communicating information can improve the usefulness of ambient displays.

Acknowledgement

X.B. Shen would like to thank Professor Peter Eades for his advice on the evaluation design.

References

Ambient-Devices, Ambient Orb. 2006, url: http://www.ambientdevices.com/cat/orb/orborder.html

Ambient-Devices, Forecasting Umbrella. 2006, url: http://www.ambientdevices.com/cat/index.html.

ASX. Australian Stock Exchange, 2006, http://www.asx.com.au

BHP Billiton, 2006. http://www.bhpbilliton.com

Card, S. K., J. D. Mackinlay, et al., Eds. 1999. Information Visualization. Readings in Information Visualization. San Francisco, California, Morgan Kaufmann Publishers, Inc.

A. Chang, A. and B. Resner. Lumitouch: An emotional communication device in CHI’01 Extended Abstracts on Human Factors in Computing Systems. 2001: ACM Press.

C. Friedrich, C., Animation in Relational Information Visualization, in School of Information Technology. 2002, University of Sydney: Sydney.

A. Grasso, M. Muehlenbrock and D. Snowdon, Visual Ambient Displays in Work Organizations: Looking for Design Dimensions.

A. Grasso, M. Mühlenbrock and D. Snowdon, D. (2003). Supporting Communities of Practice with Large Screen Displays. In O’Hara, K., Perry, M., Churchill, E., and Russell, D., Hrsg.: Public and Situated Displays: Social and Interactional Aspects of Shared Display Technologies. Kluwer International CSCW Series.

Hiroshi Ishii and Brygg Ullmer, Tangible Bits: Towards Seamless Interfaces between People, Bits and Atoms. Proceedings of CHI '97, March, 1997.

Hiroshi Ishii, Craig Wisneski, Scott Brave, Andrew Dahley, Matt Gorbet, Brygg Ullmer, and Paul Yarin. ambientROOM: Integrating Ambient Media with Architectural Space, pp173-174, Proceedings of CHI '98, 1998.

Lamping, J. and R. Rao, 1996. “The Hyperbolic Browser: A Focus + Context Technique for Visualizing Large Hierarchies.” Journal of Visual Languages and Computing 7(1): 33-55

K E. MacLean and J. B. Roderick, Smart Tangible Displays in the Everyday World: A Haptic Door Knob. Proceedings of the IEEE/ASME Intl Conf on Advanced Intelligent Mechatronics (AIM'99), Atlanta, GA, September 1999.

J. Mankoff, K. D. Anind, G. Hseih, J.Kientz, S. Lederer and M. Ames. Heuristic Evaluation of Ambient Displays. In: Cockton, Gilbert, Korhonen, Panu (ed.): Proceedings of Human Factors in Computing Systems Conference, 2003, Ft. Lauderdale, Florida, USA. p.169-176.

T. Miller, J. Stasko. The InfoCanvas: Information Conveyance through Personalized, Expressive Art. Conference on Human Factors in Computing Systems archive. CHI '01 extended abstracts on Human factors in computing systems. pp305-306. 2001

 ADDIN EN.REFLIST

E. D. Mynatt, et al., Digital family portraits: supporting peace of mind for extended family members. Proceedings of the SIGCHI conference on Human factors in computing systems. 2001, Seattle, Washington, United States: ACM Press. p. 333-340.

K. Nesbitt, “Designing Multi-Sensory Displays for Abstract Data”. Ph.D. Thesis, Information Technology, Science. Sydney, University of Sydney. 2003.

K. Nesbitt. Information Metaphor. 2001. url: http://www.csu.edu.au/faculty/sciagr/sit/admin/nesbitt/42_topic8.htm

C. Nicholson, Technical Analysis. Course Notes E114. Sydney, Australia, Securities Institute and Australian Technical Analysts Association. 1999.

Peripheral-Vision, Human Peripheral Vision. 2006: url: http://webexhibits.org/colorart/ag.html.

Rao, R. and S. K. Card, 1994. “The Table Lens: Merging Graphical and Symbolic Representations in an Interactive Focus + Context Visualisation for Tabular Information”. ACM Conference on Human Factors in Computing Systems, New York, USA.

E. B. Goldstein, 1989. Sensation and Perception, Brooks/Cole Publishing Company.

J. Redstrom, T. Skog, and L. Hallnas. Informative Art: Using Amplified Artworks as Information Displays.  In: Proceedings of DARE, Designing Augmented Reality Environments, ACM Press, New York, pp103-114, 2000.

X. Shen, A.V. Moere, and P. Eades. An Intrusive Evaluation of Peripheral Display. In: Conference on Computer graphics and Interactive Techniques in Australasia and South East Asia (GRAPHITE'05). 2005. Dundedin, New Zealand.

T. Skog, S. Ljungblad, and L.E. Holmquist, Between Aesthetics and Utility: Designing Ambient Information Visualizations. Proceedings of InfoVis 2003, IEEE, 2003.

N. Streitz and C. Magerkurth, From information design to experience design: Smart artefacts and the disappearing computer. Interactions, 2005. 12(4): p. 21-25.

M. Theofanos and J. Scholtz. A Framework for Evaluation of Ubicomp Applications. First International Workshop on Social Implications of Ubiquitous Computing, CHI2005. 2005.

B. Ullmer, Musicbox in tangible bits. 2006. url: http://tangible.media.mit.edu/projects/musicbox/.

M. Weiser, The computer for the 21st century. Scientific American, 1991. 165(3): p. 94-104.

P. Wellner, W. Mackay and R. Gold. Computer Augmented Environments: Back to the Real World. Communications of the ACM, Vol 36, No.7, July 1993.

C. Wisneski, H. Ishii, A. Dahley, M. Gorbet, S. Brave, B. Ulmer and P. Yarin. Ambient Displays: Turning Architectural Space into an Interface between People and Digital Information. Proceedings of the First International Workshop on Cooperative Buildings, 1998.