Comparative analysis of composition paradigms for personalization rules in IoT settings

The rapid pervasive diffusion of Internet of Things technologies has opened up many opportunities for people to directly personalise the behaviour of surrounding objects and devices based on the dynamic events that can occur. To this end, several tailoring environments have been proposed supporting the enduser creation of trigger-action rules. Such tools can support different composition paradigms. In this paper we present a study that analyses three composition paradigms (graphical wizard, block-based, and conversational) to better understand how well they support rule creation activities. In order to make the analysis consistent we considered three implementations of such composition paradigms supporting the same set of triggers and actions. We have carried out a first user study in order to gather empirical feedback for substantiating our analysis, which provides indications of the pros and cons of each approach.