I think this poster was talking about a rapid example to see if a concept works.

In that stuation, reducing complexity is the correct thing to do. You can achieve this by stuffing all your code in one fine, not checking if malloc failed, etc.

When you decide the concept works, THEN you go back and do it properly.

This sort of "prototyping for speed" is discussed much better here: https://gameprogrammingpatterns.com/architecture-performance...