That doesn't do anything to reduce the complexity of code, it just means you have a lot more functions to read and the actual logical units distributed across r

That doesn't do anything to reduce the complexity of code, it just means you have a lot more functions to read and the actual logical units distributed across random sets of them to fit the arbitrary line limit. Some of the worst codebases I've ever seen had many small functions because the logical boundaries of the problem domain didn't match the logical boundaries of the source code.