Sociologists are arguably the first, among humanists and social scientists, who have built a consistent canon of social network analysis. This includes often cited pioneering papers, the concepts and indicators that these had introduced, software allowing to calculate such indicators and to produce more and more standardized visualizations, and textbooks summing up and coagulating these elements (e.g. Scott/Carrington 2011). 2 In this canon, standard network data typically include a dozen to a few hundreds individuals or organizations, whose ties are described thanks to sociometric surveys, i.e. by asking individuals about one or a few specific types of relationships at a given moment. Non-standard data of course have always exi-1 Previous versions of this text have been presented in various conferences and seminars and have very much benefited from collective discussions. I am especially grateful to Fabien Accominotti, Claire Bidart, Ainhoa de Federico, Michel Grossetti, Karim Hammou, Linda Reschke, Isabelle Rosé and Tom Snijders, who in various ways made me take time more seriously, and even more to Marten Düring, who made invaluable comments on a first draft. Figures are reproduced with permissions from the authors and the journal. 2 In fact, the sociologists who built this specific canon of quantitative and structural network analysis had borrowed their main ideas to a previous generation of anthropologists and social psychologists: see Freeman (2004).