Configuration Versus Programming in User Interfaces for Autonomous Devices

A fundamental issue in designing a user interface for an autonomous device is the level of end-user control. Typically, the designer of the user interface has to make a trade-off between ease of use and full control over the device. In this study we analyze different approaches adopted in user interface design, and identify two basic categories of control: configuration and programming. Currently, most approaches adopt configuration, due to its simplicity, but we observe a shift towards programming, due to the limitations of configuration. We identify several essential differences between the two approaches, and suggest three criteria that can guide a designer in choosing a particular category of control: (1) modified versus added control, (2) unconditional versus conditional control and (3) level of device autonomy. Our criteria help to decide when configuration is not sufficient, and programming should be used instead.