The simulation Curmudgeon

Computer simulation is a true success story of modern analysis, so it is hard to be anything but positive about it. But the Simulation Curmudgeon∗ thinks maybe we should question some of the standard practices of simulation. In particular, he complains “Why do we teach people to build simulation models as if they will never change? Why do we treat simulation like poor man's queueing theory? Why do we fit input distributions like it's 1922? And why, in 2013, can't we talk to our simulations?” In this presentation the Simulation Curmudgeon will expound (in a lighthearted way) on these complaints, plead guilty to most of them, and offer some thoughts about what we might be doing instead.