I also highly respect the engineering skill of the designers of the Vasa:
That being said, Poettering created software that
A) Is much better than I could make.
B) Is used by many, many, many, people in production, without showstopper issues.
I will not deny that he deserves at least some amount of respect for what he has accomplished, including from an engineering perspective.
That's not exactly the same as the Vasa, but there too - fundamental principles were ignored while other aspects of the design were impressive.
Mostly hacky stuff that does what I need, and not much more. Definitely nothing professional,at least that's relevant to this.
I agree with what you are saying. I am by no means an expert, but I have read/studied enough to know that I disagree with almost all of the philosophy behind SystemD, and have used it enough that I know it does not work properly all the time.
Everything above also applies to Windows. That doesn't change the fact that both are impressive, though flawed (more flawed than most software I use, imho) pieces of software.
Despite all of it's flaws, SystemD has been proven through use. I don't like it, but I have to acknowledge it's capabilities.
The networking daemons were what used to cause it when I had Fedora installed, iirc. I would start looking there.
As for why it's so vague and self contradictory, I couldn't say.
Side note: While I joke about this, I respect Poetterings determination and engineering skill, even if I disagree with him.