Abstract

A theoretical astrophysicist discusses the principles and rules-of-thumb underlying the construction of models and simulations from the perspective of an active practitioner, where it is emphasised that they are designed to address specific scientific questions. That models are valid only within a restricted space of parameters and degenerate combinations of parameter values produce the same observable outcome are features, and not bugs, of competent practice that fit naturally within a Bayesian framework of inference. Idealisations within a model or simulation are strongly tied to the questions they are designed to address and the precision at which they are confronted by data. If the practitioner visualises a hierarchy of models of varying sophistication (which is standard practice in astrophysics and climate science), then de-idealisation becomes an irrelevant concept. Opportunities for future collaborations between astrophysicists and philosophers of science are suggested.