The trajectory of an object that moves in a space during a time interval describes its movement. Trajectories can be grouped by similarity according to different criteria, e.g., their shape, speed, distance travelled, duration, and visited sites, among others. These criteria have typically been approached separately, with few works considering the combination of criteria to measure similarity between trajectories. In this paper, we analyse several methods to calculate the similarity between trajectories. The most representative proposals are analysed in detail to identify lines of research. The analysis focuses especially on the proposals that consider semantic aspects for determining similarity, e.g., the categories of visited sites, and the combination of criteria. By means of two motivating examples, we show that the similarity between trajectories: i) is not a trivial notion and ii) can be useful in fields such as marketing, tourism, traffic, and animal monitoring, among others.