For sociologists, studying network dynamics implies more than simply introducing the additional dimension of time as a linear factor. The whole survey process is affected by this research goal. Data collection, database design and data analysis are all involved, and together contribute to introducing new research questions. On the basis of a longitudinal qualitative survey of 60 young people who were interviewed every three years in 4 survey waves, we discuss methodological and theoretical topics related to the study of personal network dynamics. The characteristics of Ego, those of Alter, their degree of similarity and the qualities of their relationship change over time. The structure of each system constituted by Ego, his Alter and the interconnections between them also evolves. These changes over time require us to design different articulated databases that include temporal specifications. Focusing here on structural changes, we can distinguish some sequences of evolution, point out regularities and tendencies, and identify processes in the dynamics of personal networks. Moreover, through including personal and contextual data such as life conditions, life course, expected and unexpected events, bifurcations and so on, we are able to show the strong links between the structural evolution of personal networks and extrinsic dimensions.