Rules-of-thumb are used in the design's first cut. Then engineering approximations are used in the second cut. Then sophisticated codes are used for the fluid and structural modeling. IMHO computers play too big a role. There is too much reliance on computer modelling and not enough upon actual experiment. Also there is not enough flight data. Many of the computer models have not been fully validated by flight data. I actually have far more trust in a well thought out engineering approximation than a fancy 3-D color animation generated by an unknown computer model (they're deceptive and deadly).

Gary