The cycle of observation, hypothesis formulation, experimentation, and falsification that has driven scientific and technical progress is lately becoming automated in all its separate components. However, integration between these automated components is lacking. Theories are not placed in the same formal context as the (coded) protocols that are supposed to test them: neither description knows about the other, although they both aim to describe the same process. We develop integrated descriptions from which we can extract both the model of a phenomenon (for possibly automated mathematical analysis), and the steps carried out to test it (for automated execution by lab equipment). This is essential if we want to carry out automated model synthesis, falsification, and inference, by taking into account uncertainties in both the model structure and in the equipment tolerances that may jointly affect the results of experiments.