problem is that the original COBOL programmers were expecting their stuff to be replaced within a decade, they never realized that companies would never pay to replace something if it was still (barely) functional after years of neglect – now those programs have become the backbone of their company and they’re having to shell out because it takes a special breed to learn COBOL in the day of C, C++, C#, Elixir, Nim, Crystal, JavaScript, Lua, GDScript, Dart, Swift, Kotlin, Clojure, …, …, …