Summary
The always-on connectivity of smart mobile devices (such as phones, watches, and tablets, etc.) have made them a unique platform to push information in real time and, thus, they represent a medium for receiving information in an effortless way. However, due to the mobile nature of these devices the inevitable notifications often arrive at inopportune moments. In this book we discussed the key studies in the area of interruptibility management, which demonstrated that such disruptive notifications can adversely affect the ongoing task and affective state of the user. Moreover, with the advent of advanced technologies in mobile phones, this tension is exacerbated as individuals have to deal with a plethora of mobile notifications everyday, some of which are disruptive, and on multiple devices at the same time. Overall, these findings provide evidence for the need of a smart notification mechanism that can reduce the level of disruption by filtering out the irrelevant notifications and delivering the information at opportune moments through the right medium.
We have also devoted special attention to the discussion of the state-of-the-art in modeling and anticipating interruptible moments. Desktop interruptibility management mechanisms have mainly focused on task transition phases, where tasks are related to interaction with the device itself. On the other hand, in existing mobile interruptibility studies, anticipatory models predicts users’ interruptibility by relying on their physical contextual information sampled through mobile sensors and the content of notifications. Furthermore, we note that apart from handful of interruptibility management mechanisms most models are evaluated only in the offline settings. This is due to the fact that performance of these intelligent mechanisms do not always meet users’ expectations in terms of usability and correctness. Indeed, learning the behavior of users for interacting with notifications is not an easy task as it relies on various contextual modalities, both physical and cognitive. Furthermore, we also discussed some recent studies focused on intelligent notification delivery in multi-device environments in order to reduce interruptions by predicting the right medium to deliver notification.
Moreover, we note that all the proposed interruptibility management mechanisms aim at learning the observed user behavior associated with the sensed contextual information and adapting the notification delivery process accordingly. Ideally, such mechanisms should also possess the knowledge about good behavior that can be exploited to improve the behavior of users. More in general, intelligent notification systems can be used to deliver intelligent mobile systems.
We believe that further improvements in smart devices’ contextual inference capabilities would also enable researchers to enhance our understanding of users’ interaction with notifications on these devices. We hope that the critical discussion of interruptibility studies and the overview of the key open challenges presented in this book will be of valuable help for researchers and practitioners working in this exciting and important area.