This paper considers three important research programmes in the philosophy of science: structuralism, truthlikeness, and belief revision. They all emerged some 30 years ago as formal or logical accounts of the dynamics of science. By analyzing the role of the notions of truth and similarity in these programmes, it turns out that structuralism and truthlikeness can be reconciled at least to some extent, but the difference or even the opposition between truthlikeness and belief revision is highlighted.