As someone used to problems of software engineering, chemical engineering, and electrical engineering, I feel it necessary to defend the phrase, "If it ain't broke, don't fix it." Taken out of context, it sounds like someone wants to rest on their laurels. But the real problem is that taken in proper context, it is excellent advice for an optimization of results over a larger set of distinct problems. If you have limited resources and want to maximize results, you often throw resources at the things that need the most work. You shore up the worst thing first. Then stop and see if anything else isn't working.
If what you are currently fixing is now working adequately, you need to redirect attention to that which is still broken, or at least that which is more broken than the thing you last worked on. In a world of infinite resources, you can fix everything. But in a world of limited resources, that which isn't broken doesn't need to be fixed NOW. Later? Maybe. In that larger context, the phrase makes sense. In isolation, you ignore the origin of that concept and the reality of the situation that led to coining that phrase.
There is also the 80/20 rule. The first 80% of a problem takes up 80% of the time you had available. The last 20% takes the other 80% of the time. (No wait, the math ain''t right...) But the truth is that life doesn't linearly approach a problem. It does so logarithmically, and thus the apparent math problem. If the 80% fix can get you past whatever disaster is looming, the "ain't broke, don't fix" rule is actually quite practical. I have seen the 80/20 rule in force many times and know that when the math disagrees with reality, the math is wrong.
Having said that, I also understand that human endeavor is often flawed and in need of tweaking. NO solution is likely to be perfect unless you have derived an ab initio mathematical model to conclusively prove what needs to be done. But that's so rare that I would never bet on one popping up in any field of human endeavor. So yes, anything CAN be improved, and therefore if it ain't broke, see if you CAN fix it anyway - but not if something else has higher priority.
Jon, specific to your post #16 of this thread: Change is usually an enemy for creatures who learn by rote. Learning isn't always based on reasoning. It is sometimes based on someone surviving an incident long enough to tell the tale. Imagine Grug the caveman: "Me big caveman, take sharp stick. Tiger come, I poke stick in tiger chest, tiger die." Then everyone sees the dead tiger and living Grug, at which point marketing is born, because now, thanks to Grug, everyone needs a sharp stick.
Here's the catch: As long as poking the tiger hard in the chest with a sharp stick preserves the caveman's life, nobody is there to contradict the result - and the tribe has learned by rote, not by reason. Those who fail for some reason just don't make it back to disprove that rule learned by rote. And thus that rule persists, with many cavemen bringing in many dead tigers ... until the day that Prok the caveman uses the sharp stack on a tiger but misses any vital structure. So the understandably angry tiger claws poor Prok before escaping in pain. Now Prok's story and Grug's story are in conflict. Having only learned by rote, not by science or reasoning, the tribe has to reconcile stories, both of which are incomplete. How do they resolve this? Most likely, Prok's tiger was a demon tiger who can't be killed by a sharp stick. And thus religion is born and the first demon is created.
For people who don't think so well as they just remember, rote learning is better than no learning at all - but thinking outside of the box takes them places they don't want to go; places where demons dwell.