I saw this fascinating graphic in an online presentation called “Programming and Scaling” by Alan Kay:
Kay used this to illustrate the role tinkering plays in mathematics, science, engineering, and also in making computer software. He uses the Empire State Building as a marvelous example of good engineering (and indeed it is; the building was completed ahead of schedule in 1 year and 45 days at a rate of over 7 floors per month), and contrasts that with how (even today) nothing this dependable can be “engineered” by the disciple of software engineering.
Notice how disproportionate the tinkering circle elapses the computer software bubble. The premise is that we largely build software today by tinkering along with the related “make and fix paradigm” (which he also describes).