[erikb]

Well, it's always a trade-off. When you inline stuff that is in separate functions you certainly increase the debuggability and readability and maybe even the performance of that one usecase you currently think about. If your goal is rather straightforward and your code is unlikely to be reused in high percentages in later developments (i.e. if you are coding a game in C++) then it might be a really good idea.

But if you work for instance on MS Word your goals will change 2-4x a year, you will have a list of hundreds of goals, your code might be reused in MS Excel without people even telling you, etc. In such kind of situation it is much, MUCH more important to encapsulate everything in classes and methods and functions (methods changing states, functions not) in a way that it can be treated as a black box. So in that kind of situation you are rather building lego blocks and hope that they can be connected easily enough to build the house that the little child, which is your boss, wants. In that case your lego blocks each need to be very testable and very connectable.

And the truth is that most of us live somewhat in the middle. We have a clear, highest goal that needs to achieved asap, while at the same time having loads of competing medium-priority goals that change all the time. So, while John's insides might be awesome by itself if you haven't thought about this topic before, please don't go full steam in that direction for a few years now. The best result is usually a little unclear and in the middle between two extremes.