View article

Primary motivations for effective data and process provenance in science are to facilitate validation and reproduction of experiments and to assist in the interpretation of data-analysis outcomes. Central to both these aims is an understanding of the ideas and hypotheses that the data supports, and how those ideas fit into the wider scientific context. Such knowledge consists of the collection of relevant previous ideas and experiments from the body of scientific knowledge, or, more specifically, how those ideas and hypotheses evolved, the steps in that evolution, and the experiments and results used to support those steps. This information we term the provenance of ideas or theory provenance. We propose an integrated approach to scientific knowledge management, combining data, process and theory provenance, providing full transparency for effective verification and review.