Honestly I feel your heart is in the right place, and your level of principles will produce far better products than those who care little for software quality. But your statements really don't reflect those of someone who has ever actually been in charge of large software projects. Either that or you are one who constantly feels things would have been done better if your bosses simply got out of your way. I could very well be wrong, but I doubt it. My guess would be you are a skilled senior level developer who likes to stay on the technical side of things, but that is mostly just a wild guess.
I don't think it's an unrealistic standard of perfection, the point is that it's done when the code does what it's supposed to do.
Code is not predestined to do anything. It does what humans have it do, either consciously or accidentally. If I say to release a product (with my bosses' approval), then the software is doing what it is supposed to do, which is go into production.
It won't be done because the boss sets a deadline or marketing wants it or the bean counters want to hit the Christmas sales.
This is why I don't feel you are speaking from experience. The fact is getting those Christmas sales are probably far more important than your viewpoints on perfection. And sometimes those deadlines are there for a reason. Developers are not tasked with creating some shining example of what perfect code can be, they are tasked with making a return on investment. I have a project right now which is due at the end of April because it really has to be done by the end of April. The scope of the project has changed frequently as planning and development has taken place to ensure this April 30th deadline will be hit. And it will be hit. It will be hit without 80 hour weeks; it will be hit because proper project management techniques were used throughout the project. And part of those proper project management techniques are respecting that the deadline is there for a reason and there are very few requested features that are more important than this deadline.
They act like the code is in the chain of command, they tell the developer what to do and the developer tells the computer what to do so if they say Thursday it will be done.
Honestly, writing code is the least important and easiest job of a developer. We learn how to write code in the first few years of our career; probably in school. A quality software developer's career is spent learning the more important tasks of software engineering. This includes your senior developers knowing what can be done by Thursday. You can hit deadlines without just pretending your code is working. You hit deadlines by understanding the progress of the project and understanding the priority of its features at all times. This isn't some hobby or college assignment. There may be millions of dollars on the line, and thousands of employees or customers waiting.
I have worked in situations where some features were more important than the project's success. In safety critical situations there are times where you would rather have an entire project scrapped than product an un-safe product. Obtaining FDA approval, for instance, is the type of requirement which must be fulfilled. But these situations are rare; in most software development almost all features can have their priority level be put up for review.
Yes, I know that for planning purposes you have to make some kind of educated guess but I can only estimate what I know needs to be done, not the "unknown unknowns" that can throw your estimates off by an order of magnitude or simply make it unfeasible. This is particularly - but not only - true when you're working with third party software and libraries. I say I have a worst-case estimate but really that's probably the 95th percentile, there's no hard limit where you can guarantee it'll work.
These are all very real concerns for real world projects. These are exactly why project management is so important. You often cannot guarantee something will work, which is why you have plan B. The project I mentioned before has many aspects which we cannot guarantee success. As you mentioned, this is usually because of third party software we aren't experience in yet. But we have contingency plans for each of these risks. For one important risk we even have an employee developing the contingency plan just in case because it is that important. This plan B is missing many desired features, but we have enough experience to know it will absolutely work and can be done fairly quickly.
No one will ever guarantee 100% IT project success rates. But project success is not as rare as you make it out to be if things are done properly. Success does not always mean you get 100% of what stakeholders asked for and sometimes it does not even mean hitting a deadline. And I certainly do not mean by sticking to a motto such as "it will be done when it's right."
Success usually just means the stakeholders are happy with the result.