Comment by SanjayMehta

There was a paper from Hitachi (I think) decades ago.

Their thesis was one should devise a metric to solve a problem, and discard it once the problem was solved. Then on to the next problem.