I wonder if this experience had any effect on how the author writes code. He basically backed into making the code testable. The process of identifying the clock as part of the problem would presumably have between much easier if he had adhestrongertronger SOLID principles.