An application developer spending hours refining a method/class/function to it's most optimized form is usually not worth the time. It's only going to be used i

An application developer spending hours refining a method/class/function to it's most optimized form is usually not worth the time. It's only going to be used in the application so the ROI is much lower. The author of a library with wide usage spending hours refining a method/class/function to it's most optimzied form may actually be worth the time because that effort is amortized over hundreds or even thousands of applications. The ROI is much higher in that case. Determining when it's correct to make that tradeoff is an important part of the job. Boost probably makes the correct tradeoff. But your average C++ application dev might need to make a completely different one.