How can cosmological simulations shown to be credible? Perspectives from the literature on validation and verification

How credible are simulation results, in particular in cosmology where data is sparse? The last decade has seen intensified attempts to clarify the evaluation of simulations quite generally. In particular, researchers from disciplines such as fluid dynamics and engineering distinguish between validation and verification and try to make both methods more rigorous. But does their approach make sense in cosmology? The aim of this talk is to compare current practice in cosmology with the demands of simulation assessment in engineering. I argue that these demands define something like an ideal benchmark, which is difficult to attain though in cosmology