Strategies for Intuitive Interaction in Public Urban Spaces

feedback and visual feedback versus audio feedback. 4.1. Directional versus scattered feedback Directional feedback is the mechanism whereby some interfaces yield clear and unambiguous responses to specific participants, creating the illusion of a strong mapping between physical movements (enacted with their bodies or remotely with mobile devices) and their digital counterparts. As a consequence, such digital representations are also intuitively mapped to each participant’s personal space, thus being effectively perceived as extensions of their physical bodies. Allotted and responsive ambient interfaces, being capable of accommodating a larger number of simultaneous participants, naturally respond to those at various different locations at the same time. In accordance with Principle 3 (internal consistency) in Blackler et al.’s (2006) framework, participants should be able to interact with the environment equally wherever they decide to do so within the wide interactive zone, with responsive features uniformly distributed across the interface. Communication about the overall state of the environment ends up being targeted to observers outside of the zone, who can appreciate the ‘big picture’ conveyed by the broader space. For example, in allotted interfaces, the feedback is scattered at the level of the whole environment; however, directional for each individual participant, who can clearly and unequivocally make sense of their input.Ambient interfaces, on the other hand, appear to evolve independently of participants, even when they are, in fact, responding to their input. For the latter to work effectively, people should therefore be kept oblivious of their level of participation, with little or no direct feedback provided in that regard. 4.2. Immediate and concrete versus delayed and abstract feedback Coupled to the notion of directional feedback discussed above, interfaces that are highly responsive and issue feedback immediately after an action performed by the participant contribute to the notion of control over the interface. Intuitively, participants interpret the feedback as a consequence of their acts. Contrarily to immediate feedback, a delayed response relaxes the perception of control over the interface—and, consequently, responsibility over the content generated. This is a feature particularly sought by ambient interfaces (e.g. Chromapollination (Hespanhol and Tomitsch, 2014)), which seeks to avoid synchronous interaction. Worthy of note is that the level of abstractness in the feedback is also often associated with its immediacy. Immediate feedback is often concrete and meaningful to the participant (e.g. silhouettes and shadows), while also contributing to an Interacting with Computers, 2015 by gest on Jauary 9, 2015 hp://iw c.oxfournals.org/ D ow naded rom Strategies for Intuitive Interaction 13 increased sense of agency. Delayed feedback, in turn, holds a more abstract nature (e.g. change of colours, ambient sounds and lighting conditions), causing the participant to feel little responsibility over the interactive interface behaviour. 4.3. Visual versus audio feedback Arguably, visual and audio are the two main modalities employed by interactive applications in general. In the particular field of interactive public spaces, visual feedback seems to be largely preferred for direct and immediate responses, making it the tool of choice for digital representation of individuals (Brynskov et al., 2009; Hespanhol et al., 2011; Hespanhol and Tomitsch, 2014; Jacucci et al., 2010; Lozano-Hemmer, 2001; Mueller et al., 2012; Wiethoff and Gehring, 2012). It is also the most effective modality for indication of progress regarding commands (e.g. selection) issued by the participants (Hespanhol et al., 2012; Microsoft Corporation, 2013). Studies have indicated that audio feedback, on the other hand, can be more effective as cueing mechanisms for upcoming events affecting the whole crowd in a space (Hespanhol et al., 2013). This is largely due to the nature of each modality: not everybody can see every section of the environment, but a sound played loud enough will be heard by all participants and intuitively interpreted as a general alert. Such inherent intuitiveness of distinctive audio feedback can be traced to population stereotypes (Blackler et al., 2006), given that similar audio mechanisms have been traditionally used to catch people’s attention towards an upcoming event (e.g. fire alarms). Generally speaking, however, audio feedback addresses our very innate behaviour, closely related to ‘reflexes, e.g. the startle response: an involuntary reaction to a sudden unexpected stimulus (especially a loud noise) which involves flexion of most skeletal muscles and a variety of visceral reactions’ (Blackler and Hurtienne, 2007). Naturally, a lot of audio feedback can also be disruptive and confusing in a public space; for that reason, cueing and warning sounds must be easily distinguishable from other ambient noise. Moreover, cueing and warning sounds must also be easily distinguishable from each other (e.g. audio used to indicate a green pedestrian crossing must be remarkably different from that adopted for fire alarms). It is important to point out that such strategies are not comprehensive. We sought to select case studies that are representative of common types of interactive public spaces, which mostly resort to audio-visual effects for giving feedback to the user. There are, of course, other modalities that can be potentially used for feedback—such as touch, for instance— that have not been covered by the present analysis. Also the identified strategies are not mutually exclusive—rather, they refer to particular trade-offs on how feedback is given to the public and the purpose of providing such feedback. For example, performative public spaces generally make large use of feedback mechanisms that are directional, immediate and predominantly visual. In contrast, responsive ambient environments tend to favour scattered and delayed visual feedback, coupled with audio cues or alerts when it is necessary to address the whole crowd at once.Allotted interfaces generally reflect performative design goals, although distributed over a larger interface; such larger scale, however, often causes them to occasionally employ ambient features, like the use of audio feedback to address environment-wide events. This analysis can be further elaborated by grouping the strategies described above with respect to their primary goals. For instance, interface elements (particularly visual ones) that are mapped directly to the movements of an individual’s body tend to be taken as a literal representation of that person in the digital realm (body reflectors). Therefore, feedback being directed to individuals or scattered across the interface causes participants to perceive their identity within the interactive environment differently. Similarly, whether the feedback is immediate or delayed changes the perception of the extent participants can control what they can see or hear. Finally, if the purpose of the feedback is to respond to specific individuals, using visual elements will likely work more effectively; conversely, if the intention is to cue or warn about conditions that may affect the whole crowd, then audio effects would be more recommended, given the pervasiveness of sound and its visceral appeal. Based on this analysis we therefore suggest three dimensions, each with two opposite strategies to achieve the design goal: identity; control and scope. Identity refers to how much the feedback is associated with a specific individual: directional feedback therefore delimits an individual control zone, while scattered feedback promotes equality among participants. Control refers to how immediately an action from an individual is perceived to cause as a consequence a clear impact on the interface: on one hand, immediate and concrete visual effects convey a strong agency; delayed and abstract feedback, on the other hand, indicate weak or absent agency. Finally, scope refers to the use of different media modalities to promote individual or collective awareness: visual feedback for individually driven actions; audio effects for cueing environment-wide events. Figure 7 illustrates the identified strategies, their corresponding design goals, and how their utilization would contribute for the intuitive perception of the space interactive nature by passers-by and participants. It consists of a matrix mapping the three design goals we previously identified—identity, control and scope— to the three types of interactive public space discussed— performative, allotted or responsive ambient. Each cell in the matrix indicates the type and level of feedback we have identified as ideal to promote intuitive interaction on each of the scenarios. As discussed in the previous section, some interactive public spaces clearly favour particular strategies—in fact, we would argue that to a large extent it is the very utilization of such strategies that forges the interactive space as being of one type or another. For example, if the interaction is designed around directional feedback immediately mapped to people’s Interacting with Computers, 2015 by gest on Jauary 9, 2015 hp://iw c.oxfournals.org/ D ow naded rom 14 Luke Hespanhol and Martin Tomitsch Figure 7. Feedback strategies for intuitive interaction in public spaces. Table 2. Feedback strategies mapped to type of feature familiarity and knowledge levels. Type of public space Design goal Feature familiarity (Blackler et al. continuum) Knowledge level (IUUI continuum) Performative Identity Body reflector Sensorimotor Control Body reflector Sensorimotor Scope Population stereotype Sensorimotor/culture Allotted Identity Body reflector Sensorimotor Control Body reflector Sensorimotor Scope Familiar feature from other domain Culture Responsive ambient Identity Population stereotype Culture Control Population stereotype Sensorimotor/culture Scope Familiar feature from other domain Culture body movements and predominantly visual, the responsive environment is likely to be intuitively perceived by the crowd as performative. This is a direct consequence of the high utilization of physical affordances (or body reflectors) (Blackler and Hurtienne, 2007) in the full-body interaction design. On the other end of the scale, if the feedback is scattered through the space, delayed in relation to people’s body movements, abstract in nature and largely audio-based, the environment is likely to be intuitively interpreted as ambient, albeit responsive. The abstract nature of the latter also encourages visual and audio feedback drawn from cultural conventions (familiar features from the same or other environments, as well as metaphors) to highlight intuitiveness. Table 2 maps each scenario to the type of feature familiarity and knowledge levels as presented in the intuitive interaction continua developed by Blacler et al. and the IUUI research group, respectively. It is precisely the balance in the utilization of each of the feedback strategies that will ultimately tip the intuitive perception of an interactive public space towards either the performative or ambient extremes; likewise, an even balanced distribution of all feedback strategies will result in an interface intuitively perceived as allotted by participants, with feedback scattered at the level of the whole environment (a feature of responsive ambient interfaces); however, directional at the level of each individual participant (as typical of performative interfaces). As a consequence, designers starting a new interactive urban space project should consider where their brief sits in regards to each of the design goals displayed in Fig. 7 (i.e. how much emphasis on the individuals as opposed to the broader environment). That will, in turn, determine the type of interactive environment they should design to match their audiences’ expectations so that, when people approach those environments, the interaction can unfold intuitively. For example, if the concept prescribes individually driven actions, high sense of agency over the interface and a strongly perceived identification with particular media artefacts in the environment, Interacting with Computers, 2015 by gest on Jauary 9, 2015 hp://iw c.oxfournals.org/ D ow naded rom Strategies for Intuitive Interaction 15 a performative interface should be considered. That, in turn, will have impacts on the spatial distribution of technology over the architecture for guiding the crowd movement towards that goal (e.g. a central interactive display, delimiting a clear stage area). Conversely, if the opposite is expected, an emphasis to audio should be given and the feedback mechanisms should be scattered across the environment as well as delayed in relation to movements from the crowd.