Empowering People Rather Than Connecting Them

Interaction As Closed-Loop Design

In this article, one of the basic assumptions about the fundamentals of interaction is that a user’s behaviour is about them controlling their perceptions (Powers, 1973). We view the closed loop between user and phone as a dynamic system, where designers can alter the feedback mechanisms in the phone, and where, to an extent, the human user can adapt, in order to create and appropriate closed-loop behaviour. (See Figure 1).

Figure 1.

There is already a wide variety of sensing and display technologies that can be used to construct the physical aspects of a human-computer interface, and much recent research has been dedicated to expanding the sensing and display capabilities of typical devices. Rich sensors, from accelerometers, smart clothing and GPS units, to pressure sensors, create the potential for whole new ways of interacting with computational devices in a range of contexts. Each of these has different information capacities, noise properties, delays, power demands, frequency responses, and other modality-specific characteristics. Sensors will continue to get cheaper and smaller, and new ones will create as yet unimagined interaction possibilities. Building interfaces that make use of possibly high-dimensional, noisy, intermittently available senses to create usable communication media is a significant challenge for the current HCI framework. We need general frameworks which are not tied to specific sensing or display devices, but generalise to wider classes of devices.

The display in any human-controlled control system is to provide the user with information needed to exercise control, i.e. to predict the consequences of control alternatives, to evaluate the current status and plan future control actions, or better understand the consequences of recent actions. Current examples of basic feedback loops include: Visual, audio, or vibrotactile display of the states of phone, or of distant events, people or systems.

In a mobile context users are subject to significant levels of disturbances and tend to have a lower attention span, leading to fragmentary or intermittent interaction. Because of this, in many cases it can be advisable to use modality scheduling, where the order of presentation of information in different feedback channels can be controlled as a function of the context, and the user’s control behaviour. Perception is commonly seen as process of receiving information from the world, which is typically followed by cognitive processes and then action. However, in reality, perception is tangled up with specific possibilities of action, so perception is not mediated by action-neutral models. Inner states are ‘action-centred’. Gibson called this detecting ‘affordances’. Such affordances are nothing other than the possibility for use, interpretation and action offered by the local environment to a specific type of embodied agent.

Perception is not passive reception of information—it is geared to tracking possibilities for action. Traditional actions on a mobile phone consisted of button pushing, but in a modern phone the action might also include gesturing with the phone, tapping the phone, walking or driving to a new location, changing the phone’s compass bearing. Some of these are essentially discrete actions, but many others are actions which are spread out over time, and which may be difficult to classify consistently. This perspective changes the design of feedback for users, the software engineering platforms devices are developed on, and has implications for the users’ models of how to engage with their mobile devices.