The concept of computational thinking is timely. It is widely agreed that in order to benefit and thrive in a digital world, more than digital literacy is required. Computational thinking promises understanding of information technology and its utilization without necessarily knowing how to write code. According to many authors, skills in computational thinking include the ability to evaluate and design solutions to problems. However, the vision for computational thinking, laid out by a number of scholars, is even bigger than this. It proposes that thinking habits from computer science are beneficial for any kind of problem solving and that those thinking habits form a central part of modern sciences. It might be that components of computational thinking are indeed useful in problem solving and that they play an important part in other sciences. However, computer science combines three types of thinking traditions: theoretical (mathematical) tradition, engineering tradition and scientific tradition, and it is important to differentiate between what is an aspect in one or more of these traditions and what is unique for computer science. This clarification will support the application and teaching of computational thinking. This paper explores the uniqueness of computational thinking and its roots in theoretical, engineering and scientific traditions of computing.