The model aspect is a good addition, because that's one of the things that differentiates engineers from pure scientists, or mathematicians. Engineers design to the model, which in some cases may not represent reality with complete accuracy, but will in those cases be much easier to work with. So long as that model has a demonstrated track-record of producing good results, and the same assumptions it makes have been proven valid for the current design, then it's usually good enough for design purposes.