The digital twin (DT) is a central idea in the digitization of society. The goal of this position paper is to acknowledge this idea as a paradigm and discuss how current implementations are far from software engineering standards. I propose a consideration of 8 aspects of the DT: the paradigm, its concretization as a component, its variability, its building process, its description with models, type and instance differences, its relation with simulation, and its composability. I conclude that although many digital twins claim to be such, none of them fulfill the definition as expected. This position paper provides food for thought.