Evolving Needs in Iot control and Accountability

Consider levels of detail required: Trust in the system was a major concern for users and getting used to the remote control of devices in the home was not always easy. To support transparency, we found that households wanted more detail regarding automatically triggered rules, user-triggered commands and system events. Right from the start, our first prototype enabled users to see the complete history of their logs. In addition, participants were not looking for a reduced information load. Instead, when they wanted to understand the system’s behavior they wanted full details about past statuses and the performance of their smart home. In the re-designed widget, next to an overview, a tooltip was provided for each sensor event, showing the exact time and value set for the component.

Use time as a structure: As a counterpart to detail, time was a structuring metaphor in our improved design. In the exploration phase, we observed that many rules were triggered by time, brightness and movement (with about 40% of all events occurring regularly [24]). We therefore used one day as the default timespan for viewing all system data at once. A zoom-function also allowed for detailed views.

Present home activities instead of raw data: To cope with existing and increasing amounts of data, we provided levels of abstraction and contextual cues that were meaningful to households. Next to providing the triggering (“if this”) sensor state and the executed action (“then that”), or the initiated singular actions when triggering a defined scene, we directly included the rules and scenes into the awareness widget. The goal was to prevent reported situations in which sensor events (e.g., movement sensors) jammed the home log with a plethora of events. To reduce clutter, movement detected in frequent intervals was interpreted as an ongoing movement activity detected by the sensor, which we represented with one starting point and an end point.