Firstly, my apologies to Keith at cncfins.com. It was ignorant, presumptuous and arrogant of me to think I was the only person thinking of fillets.
Second, to my mind, using other people's ideas without acknowledgement/attribution is plagiarism, but if done correctly "We stand on the shoulders of giants to see a liitle further"
Bernard/Newton
Thirdly, Pepe, supercomputers are simply computers that are able to perform things called FLOPS at a much faster rate than normal computers. In reality, they can do lots of calculations faster. When you are modelling fluid flow, they allow for the analysis of each element (a tiny cube of water in this case) to be done at a much finer level and in real time, rather than running your model and waiting overnight to see what happens. It allows for a much more scientific approach to design than the empirical methods we use.
Of course, it still relies heavily on how good your software modelling actually is at predicting real-life environment variables. In this case think, turbulence, chop/introduced air, board surface shape/roughness, weed burgs to name a few. This is supremely difficult for the guys writing the program code. So even when the first Deltas were modelled and produced optimum results in a lab environment, as we all know, they worked ok when the real world environment was similar to the model (smooth water), but were hopeless in chop. Another example, BOM's fancy new supercomputer runs several forecasting models, but they rarely get the wind forecast 'accurate' for Mandurah. In short, producing real-world software models is a very difficult thing to achieve, and much kudos must go to the guys writing the code trying to achieve it!
)
