20 Nov 2013
The lessons learned from past optimizations aren’t useless. They can hint very strongly at where to look. But we are never justified in trusting them blindly the way we trust, say, gravity. If someone gives you a performance tip, ask for the data. If they show you a benchmark, dig into the methodology. If they tell you “X is slow” or “Y is fast”, take it under advisement. Then measure it yourself, every time.