Abstract

The article presents a conceptual framework for distinguishing different sorts of heterogeneous digital materials. The hypothesis is that a wide range of heterogeneous data resources can be characterized and classified due to their particular configurations of hypertext features such as scripts, links, interactive processes, and time scalings, and that the hypertext configuration is a major but not sole source of the messiness of big data. The notion of hypertext will be revalidated, placed at the center of the interpretation of networked digital media, and used in the analysis of the fast-growing amounts of heterogeneous digital collections, assemblages, and corpora. The introduction summarizes the wider background of a fast-changing data landscape.