But that's not the point, that's not an example to be picked apart. The point is if there's a less complex way to solve the problem, then it's still causing non-essential (aka accidental) complexity. Even if the simpler solution is ugly, the more complex one is still introducing non-essential complexity. If we claim that every single thing we do with the code is part of the "real world problem", then we're throwing up our hands in the air and giving up on the problem of complexity of software. It's a naive and simple justification that's very comfortable for us, but doesn't amount to much. I really recommend reading the No Silver Bullet paper and Mythical Man Month. Fred Brooks explains those things very well.
