If the owner of a first generation mobile phone pulled his device out of his pocket, it was clear what he wanted to do: He wanted to conduct a phone call. Over the course of the evolution of mobile phones this has changed: If the owner of a mobile phone pulls out his device nowadays, we cannot know in advance if he wants to make a phone call, read news, check the weather forecast or play a game.
A Thought Experiment. With the increased variety of functionalities that mobile phones provide to their users, the uncertainty of anticipating which function a user would use next also increased. However, a theoretical optimum for the design of current smartphones would be a phone that instantly has the correct application open when the owner wants to use it — as old phones always had the “phone call application” immediately available. Whereas the dial pad was directly available for conducting a phone call on old mobile phones, nowadays a user first has to tell her phone that she wants to make a phone call by clicking on a telephone icon, before she can dial a number. An optimal smartphone would remove the additional effort that users need to make, which have resulted from the variety of applications that became available for mobile phones. Reasons for this optimal smartphone being impossible mainly come down to the uncertainty in modeling human behavior and people’s changing interests; sensor-based approaches bear an a-priori failure that leads to ambiguity.
So, given that we cannot design a perfect multifunctional phone, what can we do to come such an idea as close as possible? You can find more thoughts on this in my thesis.