For a PhD in computer science or informatics, are there good reasons **not** to choose an area of research in which very few (3 or 4) known (small- to middle-sized) research groups work in the world?