The anticipated usefulness of a new technology is one of the critical factors in user acceptance of the technology (e.g., [29]). In the context of consumer-oriented applications and voluntary use, where achieving certain results in an effective manner is less important than at work, this factor has been extended beyond rational usefulness, for example to “value” [32] or “benefit” [33]. The first comprises the consumer’s appreciation of and motivation for the technology, and the latter the user’s current need but also the future benefits the user expects to obtain from using the system.

In an interview study by Lee and Yoon [26], the users expected the value of ubiquitous technology to be in helping them in effective decision-making, increasing personal and family safety, helping personal management, increasing personal freedom, improving ease of use, increasing the portability of services and decreasing the cost of living. For elderly people, the value of intelligent applications is especially in increased safety and more effortless health care activities at home [21].

In a mass survey study of more than a thousand respondents, Allouch and colleagues [38] discovered that people expected the benefits of an intelligent home to be entertainment and facilitation in everyday routines. In addition, intelligent applications would help the users to be “ahead of their time”. The users valued the feeling of being involved in technological development, being modern and acquiring higher social status with the intelligent applications at home.

Outside the home, usefulness or benefit means something different. For instance, efficiency is specifically related to work but less often to home or leisure time. Röcker [39] asked users to evaluate scenarios for several typical smart office applications, including content adaptation, speech interface, ambient displays and personal reminders. Röcker was surprised that the level of expected usefulness and ease of use of the applications was only moderate. This indicates that the users did not expect that the intelligent office would help them a great deal to be more efficient at work, or otherwise support them in their tasks. In a museum environment, the value of an intelligent service was found to be in enriching the visitor’s museum experience with additional and multimodal information, in addition to assisting the user in finding interesting objects [40].

Another significant feature that people expect from intelligent environments is that the applications and services in them are easy to use and control. For instance, Ziefle and Röcker [41] studied the acceptance of smart medical systems among the middle-aged and elderly. The system, presented in scenarios, monitored the biosignals of the user, communicated automatically with health care personnel and if necessary activated an alarm. The most critical characteristics of the system were ease of use, controllability and communication with the device. The users were less interested in the design or appearance of the system, or whether the system could be recognized as medical by outsiders.

Sometimes the assumption is that an automatic system is the most easy to use. However, an automatic system may cause the user to feel that (s)he is not in control. Misker and colleagues [42] studied experimentally different user interaction styles of combining various resources of an environment, i.e., input and output interaction devices, ad hoc. The test subjects preferred a style in which they had control over the device selection, even if they had to use extra effort, compared to an automated selection. Zaad and Allouch [22] conducted a study of a motion sensor system for elderly people and compared two versions of the system: a fully automatic system and one in which the system carried out checks with the user on whether the information detected by the system was correct. It was found that those elderly people who perceived having more control over the system and thus over their wellbeing had a greater intention of using it.

The intelligence—context-adaptive, learning, and proactive behavior—of a system may be a reason for the loss of user control. If the user does not understand the logic by which the system behaves, that is, the system lacks intelligibility, the result will be a loss of user trust, satisfaction and acceptance [43]. For instance, Cheverst and colleagues [44] tested this with an intelligent, learning-able office system prototype. The system allowed the user to scrutinize and control the rules that the system followed in action. In the questionnaire study, more than 90% of 30 participants expressed the need for controlling the system. Independent system action was sometimes found useful, as two third of the participants would have liked to switch between the two system modes, automatically or by prompt. The participants who had more expertise in computer knowledge expressed more interest in controlling the system.

Lim and colleagues [43] studied if different types of explanations about a system's decision process would add intelligibility of a system for the user. They carried out a comparison experiment with over 200 users, and found that explanations clarifying why the system behaved in a certain way resulted in increased user understanding and trust.

Therefore, to prevent the threat of intelligent environments taking control over the user, the user needs to be provided with enough information and tools to control system actions. For this purpose, Bellotti and Edwards [45] propose a design framework for context-aware systems. Their framework declares the following four principles to support intelligibility and accountability in design: 1) Inform the user of current contextual system capabilities and understandings, 2) Provide feedback (both feedforward and confirmation for actions), 3) Enforce identity and action disclosure particularly with sharing nonpublic (restricted) information, and 4) Provide control to the user. Also, the framework by Assad and colleagues [46] for creating ubiquitous applications ensures that the applications will be such that the user has control over what information is modeled (especially about people) and how it is used.

Sometimes, users are willing to give up control for benefits. In a study by Barkhuus and Dey [47], users felt a lack of control when using fully or partly autonomous mobile interactive services, but still preferred them over applications that required manual personalization. The usefulness of the autonomous applications overrode the diminished sense of control. However, Barkhuus and Dey warn that loss of control may still result in user frustration and turning off services.

Multimodality may increase the usability of intelligent environments. Blumendorf and colleagues [48] found that multimodal interaction (voice, mouse, keyboard, and touchscreen) was preferred and more positively evaluated by users carrying out interactive tasks, compared to single modality interaction. Multimodality enables a more personal use and creates flexibility; for instance, speech-based interaction with office applications is avoided in public places but may be the most preferred interaction method elsewhere [39].

However, multimodal interaction can be implemented in a more or less usable way. Badia and colleagues [49] found that, if the interaction is as hidden and as natural as possible, the results and consequences of the interaction may actually be more difficult to understand.

Furthermore, some user groups such as the elderly may have difficulties in accepting an intelligent environment despite easy and flexible control and interaction. For the elderly, it is important that the intelligent applications and services integrate to and support the existing practices and routines of the user—who may never have used a computer [21].

Trust is related to control, privacy and monitoring, but also to security and the data protection of intelligent systems [32,50]. Trust is a multi-dimensional, complex issue (for analytic definition, see e.g., [51]) that may grow slowly but is easily lost [32]. In a long-term experiment, the inhabitants of an intelligent home did not gain complete trust in the system even in six months [17]. Lack of people’s trust and belief is a critical shortcoming regarding the acceptance of intelligent environments. One of the main challenges in gaining trust seems to be users’ fear of losing their privacy.

Intelligent environments face a privacy dilemma: the user’s personal data, current and historical, needs to be collected and stored to be able to provide intelligent, contextual and pro-active services for the user [52]. Indeed, in Allouch and colleagues’ [38] study of an intelligent home, users did not accept a smart mirror sending private health data to a doctor. Nor did they accept automated home services functioning without the user’s intervention. In other words, user acceptance is low if users cannot trust the system to protect their privacy and to be fully in control. Privacy and control are especially important at home. In Ziefle and Röcker’s study [41], a medical system, which monitored users and automatically communicated data to health care staff was least accepted when integrated at home, but better accepted when implemented as a mobile or wearable system.

Privacy is also an issue outside the home. Röcker [39] found that users were willing to use smart office applications which may reveal something personal about the user more in private places (a private office room) than in public offices. Because of privacy worries, gaining user acceptance for one of the core functions of intelligence environments, namely continuous monitoring, is a challenge. Moran and Nakata [53] reviewed several studies of the effects of monitoring (as in data collection, with no surveillance) and concluded that in general, awareness of monitoring has a negative effect on user behavior. However, the studies were often conducted under laboratory conditions (“smart home labs”) which may cause different behavior compared to the real world. The researchers analyzed several factors that may impact behavior: the number of devices, disturbance of the devices in looks and positioning, control functions such as deletion, deactivation, inhibition and avoidance, the frequency of data collection, informing the user, etc. Systematic studies of the effects of these factors are still lacking due to the rarity of real intelligent systems.

There are some research results showing that nations or cultures may differ in how people accept intelligent environments. Forest and Arhippainen [54] studied the cultural aspects of the social acceptance of proactive mobile services. They discovered that the concept of labor was different in France and in Finland, which caused Finnish users to more readily accept solutions that organized the entire life of the user in order to improve his work efficiency. French users did not accept the approach where life was devoted to work, as it threatened their identity and the meaning of their life. Also, the roles of women and men in society were different, and this caused differences in acceptance of services for controlling the activities of other people. For Finnish people, it seemed to be more natural to accept living in symbiosis with a device ecosystem, whereas French people would have wanted a device environment where the human being can keep control of what is happening all the time. Röcker [39] found that willingness to use intelligent office applications in private or public places differed more clearly with American users and less with German users. However, the difference was not put to a statistical test.

Individuals clearly differ in their willingness to adopt technologies and intelligent services. For example, gender, region of residence, job and usage of the Internet predict an acceptance of digital home care services [18]. Other individual characteristics might be age, education, social status, wealth, etc. Even personality characteristics such as conscientiousness, extraversion, neuroticism and agreeableness may matter [55]. In the field of mobile Internet services, it seems that socially well networked and “personally innovative users” (those having a tendency to take risks) may find the services more useful and easy to use, and this way are also more willing to take them into use [56].

A widely known description of how different types of individuals adopt new technologies is Rogers’ [57] categories of adopters: innovators, early adopters, early majority, late majority and laggards. As Punie [52] notes, it is difficult to believe that intelligent environments would be accepted by everyone. There will always be early adopters as well as late adopters and those who totally resist. However, it is important to develop intelligent technologies for a variety of user groups, not just those who appear to be the most ready and willing, to avoid the growth of a digital divide between those who can and those who cannot use intelligent environments.

First, based on a review of relevant literature, the following highlights various aspects related to user experience of haptic, tangible and tactile interaction. Because of the fact that not much research literature exists on the user-centric aspects of haptics, various elements of experience are discussed based on papers on both actual experiences and expectations. This is crucial with regard to both expectations and experiences, as these two interplay closely with and influence each other; expectations affect the actual experiences, and understanding actual experiences can indicate aspects that would also be central in the users’ expectations.

To clarify the terminology first, in describing touch-based interaction, a widely accepted term is haptic interaction, which pertains to sensory information derived from both tactile (i.e., cutaneous, skin sensor -based) and kinesthetic (limb movement -based) receptors [72]. Examples of tactile interaction are vibration, pressure- or temperature-based output, and smoothing, pressing or squeezing as input. Kinesthetic gestures have been categorized as, for example, manipulative, conversational, iconic, metaphoric, or semaphoric [73,74]. In day-to-day interaction, touch and gestures are essential in a wide range of daily ad hoc tasks, such as in providing a feedback loop for manipulating objects, guiding our movements in three-dimensional space and various social interaction settings. Considering this and the fact of haptic being an underused modality, TUIs can also be considered to be a potential technology in the development of new ambient intelligence services. A system with haptic output provides the users with heads-up interaction—that is, letting the user visually scan their surroundings instead of reserving them for interaction with a handheld device. Incorporating the sense of touch into remote mobile communication has also been proposed as one solution in enriching multimodality in mobile communication [75].

Touch and gestures could allow more accurate interaction with intelligent environments, for example initiating object-based services, downloading further embedded information, user registering to a service/location, selecting augmented information in an AR browsing view (see 4.2), etc. In addition, the haptic modality serves well in multimodal interaction, supporting and enriching interaction based mainly on other sensory modalities, such as vision or audition.

To start with, several studies indicate that tangible interaction can enhance creativity, communication and collaboration between people. Africano et al. [76], Ryokai et al. [77] and Zuckerman et al. [78] have tested tangible interaction concepts in the learning environment with children using hi-fidelity prototypes. All of the studies report that the tangible interfaces promoted collaboration and communication; the children were engaged in the interaction and considered it fun. Hornecker & Buur [79] show examples of three case studies of tangible interfaces where the same applies to adults.

When tangible interfaces clearly enhance the social aspects of interaction, virtual and augmented reality tend to serve more solitary experiences, since the environment is generally viewed through a personal device such as goggles or a hand-held display device. These technologies have the potential to significantly change people’s perception and experiential relationship with the environment by immersing them in the interaction. Hornecker [80] has compared experiences with two different kinds of exhibition installations in a museum environment. One installation consisted of a multi-touch interactive table and the other provided an augmented reality view of the museum environment with a telescope device. The ethnographic study indicated that, while the telescope installation provided a more immersive experience, the multi-touch table enabled more social and shared experiences.

With regard to tangible interaction, one of the earliest examples is by Ullmer and Ishii [81], where user’s hands manipulate physical objects via physical gestures; a computer system detects this, alters its state, and provides feedback accordingly. More recently, RFID-based NFC (Near-field communication) ‘touching’ has generated greater interest in the research community. For example, Riekki et al. [82] and Välkkynen et al. [83] have built systems that allow the requesting of ubiquitous services by touching RFID tags. In the studies by Välkkynen et al. [83], touching and pointing turned out to be a useful method for selecting an object for interaction with a mobile device. Kaasinen et al. [84] studied physical selection as an interaction tool, both for selecting an object for further interaction, and for the device also scanning the environment and proposing objects for interaction. They found that in an environment with many tags, it will be hard to select the correct one from the list presented after scanning. Distance and directionality in tag interaction need to be precise, because indefiniteness in them can cause a confusing user experience. The results by Kaasinen et al. [84] indicated that the best technique for touching would probably be having a ‘touch mode’ that can be activated and deactivated either manually or automatically.

More recently, Kaasinen et al. [24] have studied the acceptance and user experience of touchable memory tags. Ease of upload and fun interaction were identified as the main strengths of such an interaction metaphor. The main concerns related to control over the interaction between the mobile phone and the memory tag as well as fears regarding security and reliability. Users should have ways to assess and ensure the trustworthiness of memory tag contents, and the user-generated content may require moderation. The studies by Kallinen [85] also support these findings: touch-based interaction with NFC-enabled mobile phones was regarded as an efficient, intuitive and novel interaction technique that is suitable for a variety of domains and applications. Such interaction could be considered in, for example, smart home environments (see Figure 2). It was found to be the most efficient, effective, practical, and simple method compared with speech and gesture-based interaction in both subjective and objective measures. In addition, touch-based interaction was considered to be socially very acceptable, exciting, and innovative. The biggest user concerns were device dependency and security, especially in use cases with payment and access control features. Furthermore, Isomursu [86] identified a need for tag management. They propose that if NFC technology becomes common and if management practices are not used or available, tags can create “tag litter” that ruins the user experience by corrupting the trust in tags and tag-based services.

Also the potential users’ expectations of haptic interaction in interpersonal communication have been studied [75,87]. In these papers, the main benefits of tactile communication were seen to be the added richness and immersion it can bring to interpersonal communication. The interaction could become more holistic and thus richer information could be conveyed. The haptic modality was seen to add the currently missing contextual nuances and thus ease the use of mobile technology. For example, to convey an emotion, a tactile input could either enrich the multimodal message as tactile output or be transformed into another sense modality (e.g., represented in a certain type of sound). Another clear benefit was expected to be the spontaneity and speed of communication. Haptic communication was seen to be an easy, simple and fun way to convey sudden emotions. Furthermore, the added privacy in the unobtrusiveness of tactile interaction was seen as beneficial for remote communication. On the other hand, touch was regarded as such a private sense that the user must be in control of who can communicate with them via the haptic modality. The authors conclude that, to make it fluent, the interaction itself should resemble or have similar metaphors to our everyday use of the touch.

To summarize, the above-mentioned studies clearly confirm even the earliest researcher-originated expectations of touch-based interaction—intuitiveness and efficiency in interaction with real world objects. Publications about user research in this area are diverse and do not currently allow a holistic view of UX and expectations of haptic interaction. Nevertheless, the following provides an early consolidation of the UX-related aspects discussed above, mainly to be used as early guidance in the design of touch- and movement-based interactions in general and especially in AmI services:

The different submodalities in the haptic modality enable extensive ways of inputting information into a system and providing immersive and intuitive output for the user in a multi-modal, holistic way. Thus, touch-based interaction also has great potential to help make ambient intelligence a well-accepted technology and a central part of the everyday life. Currently, such interaction technologies becoming more general might mostly depend on how the possibilities of touch can be mapped into the input and output for a system in a universally understandable way, and what the first widely accepted services utilizing these technologies are to be.

Augmented Reality (AR) is another central interaction tool that can be seen as a promising interface and interaction tool for intelligent environments, and has recently received considerable interest also from the user-centered perspective. Generally, AR can be considered to be one key technology in catalyzing the paradigm shift towards mobile and ubiquitous computing, happening “anytime, anywhere” [1,66,88]. The general approach of AR is to combine real and computer-generated digital information into the user’s view of the physical and interactive real world in such a way that they appear to be one environment [89,90]. This allows integrating the realities of the physical world and the digital domain in meaningful ways: AR enables a novel and multimodal interface to the digital information in and related to the physical world. This new reality of mixed information can be interacted with in real time, enabling people to take advantage both of their own senses and skills and the power of networked computing, while naturally interacting in the everyday physical world [91]. Therefore, it also serves as a versatile user interface for intelligent environments [92]: for example, browsing the digital affordances embedded in the real world as well as manipulating the ambient information in a natural and immersive way. This aspect emphasizes that objects themselves including digital information can create ‘smart’ services, in which AR could be used as a paradigmatic interface metaphor (see, e.g., [93]). On an abstract level, this is closely related to the concept of physical browsing, i.e., accessing information and services by physically selecting common objects or surroundings [94], but approaches the paradigm from a visual perspective.

Augmented reality has been applied in diverse practical day-to-day use cases, in entertainment and gaming, as well as in tourism (Figure 3) and navigation. Examples of augmentable targets have varied from surrounding buildings, machinery or other static objects to vehicles, people, and other dynamic objects.

The diversity of AR as a concept provides a powerful toolkit to design new services, but at the same time creates great challenges for the design of pleasurable user experiences. In our earlier research [65,66,95] we have identified various categories of user experience that potential users expect to arise in interaction with mobile AR, considering it not only as an interaction technology but, rather, as a holistic service entity encompassing AR content and various ubicomp- and AmI-related functionalities. The most central experiential expectations reported in the papers are summarized in what follows (see [66] for more detailed descriptions).

The expectations can be categorized in six main classes of experiences. First, various instrumental experiences were often expected, demonstrating sense of accomplishment, feeling of one’s activities being supported and enhanced by technology, and other pragmatic and utilitarian perceptions of service use. Second, cognitive and epistemic experiences relate to thoughts, conceptualization and rationality. Especially experiences of increased awareness of the digital content related to the environment and intuitiveness of the interaction were found to be much anticipated. Third, emotional experiences relate to the subjective, primarily emotional responses in the user: for example, amazement, positive surprises and playfulness were often expected from services with AR interaction. Figure 4 illustrates two examples of AR applications focusing especially on playful experiences. [66]

Fourth, sensory experiences relate to sensory-perceptual experiences that are conceptually processed, originating from the service’s multimodal AR stimuli and influences on the user’s perception of the world around them. Fifth, social experiences relate to human-to-human interactions that are intermediated by technology and features providing a channel for self-expression or otherwise support social user values. For example, connectedness and collectivity relate to the expectation of mobile AR offering novel ways for reality-based mediated social interaction and communication. Social experiences would develop from the collective use and creation of the AR content, in utilizing the socially aggregated AR information in personal use, in co-located and collaborative use of the services, as well as in using the AR as a tool for communication. Finally, motivational and behavioral experiences are created when the use or ownership of an AR service causes a certain behavior in the users [66]. Overall, the more detailed experiences under each class are summarized in Table 2.

All in all, potential users attributed a diverse range of expectations to services based on mobile AR. Clearly, the end user’s perception of a service entity or an interaction technology is not limited by theoretical or technical definitions, but driven by their own needs and practices. For example, social experiences, community-created content or proactivity are perhaps not specific to mobile AR –based interaction per se but were expected nevertheless. In addition, many expected mobile AR services to include features like context-awareness and proactivity, which could also be seen as characteristics of intelligent environments. Hence, the expectations related to one aspect or embodiment of intelligent environments (here, mobile AR) can easily highlight issues that can be seen as equally important in other technologies. These anticipated experiences can be considered to be positive and satisfying experiences, that is, something for the user to pursue and the designer to target in their design, and something that could also be utilized in design activities around systems that more fundamentally represent intelligent environments.

To summarize, this section has introduced various expectations and experiential aspects that can be considered important in designing tangible and embodied interaction for intelligent environments. The results on augmented reality in particular imply that people’s expectations of new interaction tools and the experiences they can evoke display a very extensive spectrum of facets that demonstrate minor aspects of the overall UX—from instrumental experiences to emotional and epistemic, for example. The relevance and importance of the various aspects listed here most probably vary from case to case and from context to context. Consequently, the section aims simply to identify and highlight different aspects that should be considered, but the questions of “which of them”, “to what extent”, and “how to incorporate them” should be considered in each development case based on its context and other boundaries.

Overall, user experience includes several dimensions on different levels from instrumental needs for motivation. Even if the dimensions of user experience were identified with the selected interaction tools, the same framework can serve as a starting point when studying other interaction tools and the entities of intelligent environments. The overall user experience is partly based on the interaction tools and the whole interaction experience when acting in an intelligent environment, which makes this point of view on interaction-based experiences a relevant one.

In the next section, we will enhance our focus from mere interaction to the design of intelligent environments and the role of users as crafters of intelligent environments. User role as a crafter of intelligent environments in particular serves the emotional and motivational user experience goals, such as playfulness, inspiration and creativity.

Home was seen to provide the perfect settings to consider the user expectations relating to DIY creation culture. According to Rodden and Benford, the essential property of home is its evolutionary nature and receptibility to continual change [100]. Beckmann, Consolvo and LaMarca see that the need to support this change is critical to the successful uptake of ubiquitous devices in domestic spaces [99]. They emphasize that the aim of using DIY methodology in the assembly of the ubiquitous devices (which share the environment) is to allow users to have more control in the space they live in.

Kawsar et al. continue that, ideally, the inhabitants should carry out the installing and configuring of the devices [98]. They have in-depth knowledge of the structure of their home and their activities, resulting in a better understanding of where and which physical artifacts and applications to deploy [98]. System issues for involving end users in constructing and enhancing a smart home entail two requirements: 1) a general infrastructure for building plug and play augmented artifacts and pervasive applications (the architecture) and 2) simple, easy to use tools that allows end users to deploy the artifacts and the applications [98]. To study these two approaches and consequent user expectations in the DiYSE program, we developed and carried out user research into a Home Control System [101]. The use and expectations of pervasive applications and augmented artifacts were studied in Life Story Creation [102] and Music Creation Tool [103] research projects. Figure 5 illustrates these projects.

The three projects were studied from August 2010 to January 2012. The Home Control System was developed for supporting elderly people in living more independently in their own homes and in a (smart) nursing home. The system was expected to enable the creation and combination of interactive operations in a sensor network-enabled intelligent environment. The user expectations were considered with a group of nurses (four users: 4 female and 1 male, varying in age between 28 and 42). The Life Story Creation -project was about constructing an easy do-it-yourself application for elderly people to create their personal retrospections. The acceptance and user experience factors were studied first with focus groups involving senior citizens (15 users: with an average age of 70), and later the application was co-designed with expert-amateur writers (five users: 4 female and 1 male, varying in age between 55 and 69 years). The aim of the Music Creation Tool was to allow disabled people to play music together. The tangible, interactive system was considered to provide an easy means of composing and configuring digital instruments. User expectations were studied with end-user observations (four users: all male, varying in age between 26 and 58) that focused on learning, independent action, supporting creativity and managing confusing situations.

User research, focusing on user acceptance and user expectations, was considered to help in constructing an ideal DIY smart system. The results from all user studies of the DiYSE program were refined into manifesto statements that were published after the project for opening up the technology-driven DIY creation to a wider audience [104]. Table 3 presents the 13 topics of the manifesto for constructing an ideal do-it-yourself smart system.

To illustrate these manifesto statements and to consider them in the terms of user expectations of intelligent environments, in what follows they will be reflected to the three DiYSE -projects: Home Control System, Life Story Creation and Music Creation Tool.

At the core of the do-it-yourself culture is the creativity and craftsmanship that is expected to be provided from the part of users. When DIY is positioned in the intelligent environment context, support for self-actuality and creativity is equally essential. Two case projects, Life Story Creation and Music Creation Tool, provided an interactive writing application and tangible musical instruments that were explicitly intended be creativity supporting tools. User studies of the Life Story Creation focused on defining experiences relating to writing as a creative activity, but discovered user expectations also led to study the issues around the processes—the flexibility to extend and modify the writing process according to the flow of creativity; the different types of writing persons and how various kinds of creativity should be supported, and the difficult issues relating to sharing an intimate piece of writing.

Several different expectations were made of all the case DiYSE systems, as in each case users were different and had various levels of computational skills. The extreme case was the Music Creation Tool, in which there were many variations in the intellectual and adaptive skills of the people it was studied with. The Music Creation Tool designed for a special group, people with a mild or moderate (Diagnosis ICD-10) intellectual learning disability, had to be flexible and provide options according to various skill levels. The anticipated use of the system was considered at first with a music therapist; then observation studies confirmed the usage of the instruments and by analyzing the results the user expectations could be determined. Life Story Creation and Home Control System focused on finding ‘warm’ experts who would be the support persons for the DIY ecosystems. A ‘warm’ expert is a person who has some degree of emotional ties to the (end-) users of the application they help to design and/or implement [105].

The principal expectations of the Home Control System user research related to the component creation—what kind of components users wanted to create and what they expected to be ready-made. The studies revealed that the nurses, the ‘warm’ expert of the ecosystem in question, preferred ready-made templates for constructing the setups, and their task would only be to fine-tune the templates. Also according to the results from Mackay [106], when studying triggers and barriers to customizing software, simply providing a set of customization features does not ensure that users will take advantage of them. The process involves a trade-off—the choice between activities that accomplish work directly and activities that may increase future satisfaction or productivity. This was also our important finding relating to user expectations of the component creation.

According to Gershenfeld [107], digital recycling is about a digital processes being reversible, based on an assumption that the means to make something are distinct from the thing itself. And further, the construction of digital materials can contain the information needed for their deconstruction. In all DiYSE projects the possibility of recycling templates and projects were presented as an option, but the developed proof-of-concept prototypes concentrated more on speculating about meaningful opportunities for reuse. In the contemporary digital world it should be easy to organize sustainable conditions in which things are not thrown away, but rather (scrap) materials and older projects are kept in store for recycling.

As regards the culture of collaboration, Kuznetsov and Paulos [96] clarify that DIY communities invite individuals across all backgrounds and skill levels to contribute, resulting in: 1) rapid interdisciplinary skill building as people contribute and pollinate ideas across communities, and 2) increased participation supported by informal (“anything goes”) contributions such a comments, questions and answers. The study by Fischer et al. [108] also reveals that question asking and answering is the core process behind the propagation of methods and ideas, i.e., participants tend to “learn more by teaching and sharing with others”. From the three DIYSE projects, Life Story Creation was the only project in which the users actually shared their work with each other during the evaluations, by providing instructions and comments or by just reading each other’s stories. According to our studies, collaboration has a central role when defining the efficiency and usefulness of the DIY system; collaboration provides most value for individual users. According to Kuznetsov and Paulos, DIY is a culture that strives to share together while working alone [96]. Social media was thought to be the entryway for publishing music in the Music Creation Tool and sharing memories in the Life Story Creation, and the people who were studied appreciated this opportunity. However, in both cases, the sharing of unfinished projects was seen to be unpleasant.

An important part of the user experience of the intelligent system is that it should motivate and support users to create the projects by means of subtle, non-harassing coaching. As a result, all users, regardless of their level of expertise, are able to use the system to the fullest [104]. In the Home Control system, for subtle tuition, there were various different attempts to employ diverse methods for the subtle tuition. The attempts comprised, e.g., step by step wizards that would guide the users through processes; a puzzle metaphor for connecting the components representing tasks and devices together; use of a sentence structure for supporting to create the setups; enlarging and diminishing grids, palette type of selection methods and ultimately, voice feedback for detecting the objects of the environment [101]. It appeared that the level of expertise set most critical requirements for the tuition method.

Smith et al. [109] have claimed that any textual computer language represents an inherent barrier to user understanding. According to them, a programming language is an artificial language that deals with the arcane world of algorithms and data structures—people do not want to think like a computer, but they do want to use computers to accomplish tasks they consider meaningful. The above-mentioned numerous tuition methods illustrate how in the development of the Home Control System special attention was given to how the system would adapt to user’s terminology.

To better make use of what is offered by the physical environment, DIY environment should be facilitated with multimodal system inputs. The Music Creation Tool consisted of tangible instruments that provided musical experiences, and one of the key objectives was to observe how users interacted with them in their environment. Observations surprisingly revealed that there were several (learned) expectations of corresponding analogue, tangible objects. A fine example of the use of physical modality is one of the considered alternatives for configuring the Home Control System. The research group developed the idea of taking a snapshot of the environment by physically selecting devices: a switch could first be turned on, and then the state of it would become visible in the user interface. In this way the user could have a better cognition of the control environment. According to Fischer et al. [108], this responds to the aim of reducing the cognitive burden of learning by shrinking the conceptual distance between actions in the real world and programming.

At best, the Do-it-yourself approach should respond to the challenge of many different expectations of the different characteristics of users. The main value of the Home Control System, for the nurses, was the pre-configured tasks that they expected to ease their workload. An unexpected value, that was not thought initially, was that the system could also assist patients who cannot move from their beds. The Music Creation Tool observation period confirmed that the unexpected end-users’ needs were related to publishing music and performing for an audience. In the Life Story Creation, the interviewees anticipated that the system would respond to the need to trace back one’s personal history and learn about lost family ties.

In addition to the topics we identified in our DIYSE studies, other topics may also influence the value of DIY intelligent environments. One potential topic is context-awareness. Context-aware applications are applications that implicitly take their context of use into account by adapting to changes in a user's activities and environments [110]. Context-awareness has been successfully introduced in projects such as iCAP [111] and CAPella [110]. iCAP is a tool that allows users to quickly define input devices that sense context, and output devices that support response, create application rules with them, and test the rules by interacting with the devices in the environment. CAPpella empowers users to build context-aware applications that depend on intelligence—making inferences based on sensed information about the environment.