Someone told me I speak in metaphors. I recently spoke about how the referees plop down the football where they believe the ball advanced, this is fairly rough, and then they take out a precise chain to measure out the yards. You give up some at one place and are painstakingly accurate elsewhere. Research and researchers tend to be that way too: we are less than angst-ridden in modeling an optimization problem, and then are painstakingly accurate in getting a precise approximation ratio. Sometimes the optimization may not capture the underlying bottleneck in the system in practice. Which brings me to another metaphor I am tempted to use as I go through the day and almost always manage to trap it before it tumbles past my lips: you don't go to war with the army you want, you go to war with the army you got.