This reads like the ultimate goal is writing computationally efficient software. The key tool to measuring non-performant code would be asymptotics and specific

This reads like the ultimate goal is writing computationally efficient software. The key tool to measuring non-performant code would be asymptotics and specifically Big Omega and Big Theta, because you need to shrink the total window size as well as shift it down. Engineers hate being evaluated on whether they know Big O, much less we hardly test on Big Theta. Secondary to that, it's increasingly difficult to find engineers who even have a general mastery of asymptotics and algorithmic efficiency.