Visual end-user programming in smart homes: Complexity and performance

End-user programming in smart homes addresses tasks that range from very simple to very complex. In this study we investigate how task complexity impacts performance when using two different visual programing representations: form-filling and data-flow programming. We invited 16 participants to create rules to solve smart-home situations of varying complexity and analyzed their completion rates for the two visual programming representations. We identify the following areas of difficulty for programming novices in our smart home scenario: choosing and connecting triggers and their specifications, and using Boolean operators. Both visual representations enabled users to specify complex rules, with advantages in different areas. They indicate that overall task complexity might be less decisive for performance than complexity of triggers and Boolean operators.