People who have been coding for a while learn not to repeat their mistakes. Their code contains good-ish abstractions and other “clean code” features because they intuitively know what they are doing. The code they just wrote to get something done & shipped is probably somewhat “clean” by most standards.
But communicating those ideas to less experienced developers is where the problem comes in and all the prescriptive dogma arises.
But this happens in all fields. How many times do experts give some advice they don’t always follow themselves? Experts understand when and how to break the rules, and that comes with experience.
Novices need simple rules to follow. The world of a novice is filled with uncertainty, they have nearly zero intuition as to what's good or bad, so simple rules that get them 80% there are essential. Otherwise they'd get lost in the complexity.
However, with time, as they gather experience, learn, and mature, they should be able to figure out the reasoning behind the rules they were once given. That then will allow them to make decisions on whether the rule is appropriate for the given context or not.
Of course, getting to that point requires continuous improvement, which is why many, if not most, programmers don't get there.
Example: "Never use `goto` in C." - great advice for a beginner, they'd just make their code into spaghetti. However, a seasoned coder knows that `goto` is only a problem when used to jump to arbitrary points in the code, so using it to reduce duplicated error handling is perfectly fine.
Same with “premature optimization”. Many things can be fixed later. Some can’t or won’t, and those typically get ignored or relabeled as “good design” when it is still ultimately minmaxing cpu time versus developer brain cells.
> communicating those ideas to less experienced developers is where the problem comes in and all the prescriptive dogma arises.
I think a lot of it is just lack of context. A lot of "best practice" advice is valuable in certain cases but it gets over-applied, and people end up wasting a lot of time on things which don't matter.
But communicating those ideas to less experienced developers is where the problem comes in and all the prescriptive dogma arises.
But this happens in all fields. How many times do experts give some advice they don’t always follow themselves? Experts understand when and how to break the rules, and that comes with experience.