Of course the one thing most people don’t consider is maintainability, which means not only how readable the code is, but how easy it is to find people who know how to maintain the code. That’s why Java/C# are the winners, over languages that while better, don’t have nearly as many people who know it. All Java and C# has to be is “good enough.”
Maintainability is bunkum. IMO. Yes, I am a friggin code purist, and write my code to be maintainable most of the time, yet I know my own sins, and that is one.
Far better code be totally unreadable! Then folks would push the information about what the code is supposed to do — the specification — to the outside.
Yet in development you have to read your own stuff. and so you have to use structure and good style. Still. Trillions has been wasted in trying to get old code to do new tricks, and in trying to dope out functional specification from massive programs because that’s the only place it exists. Or worse — trillions wasted in having no specification except that of the code, because management believes code source is readable. It creates fear of “reinventing the wheel”. Bunk!
Reinvent the wheel every time! That is evolution.
How would the world save trillions in software development? By requiring that every source deliverable be obfuscated, if there was any source deliverable at all. That way no one would ever think, ever for one second consider, re-engineering old code.
I’m not against code hornbooks — examples of code. That’s fine. But that’s the limit of readable source code, or should be, imo.
Write-once programming is salvation! That’s like why as humans, we only get one life. No redos. Thus we have to take it seriously.