Ah, well, there you have it! One certainly cannot argue with such an overwhelming barrage of evidence and nuanced analysis.

Sorry. But, imho, you're wrong :)

1. Have you ever written a proper SysV script? It was a nightmare.. The last Systemd unit I wrote was 4 lines I think
2. SysV had lots of race conditions and was a nightmare to troubleshoot
3. The old scripts were entirely distribution specific.
4. The old Init scripts depended on all kinds of hacks to work properly.
5. Dependencies in init scripts? not really a thing.. You basically just fudged everything all together in a specific order and hope they loaded in sequence as required.. In SystemD, they build dependency handling in (and, built in a lot of prerequisites to handle it)
6. Because Init scripts don't do real dependency management, they also ain't fast.

Init scripts are the equivalent of using a perl script as the kernel programming language. They were crappy, hard to read, hard to modify, hard to maintain.

The biggest objection against Systemd as it stands today was because Linus had an opinion about it (but, admits something like it is needed), and the assumption that redhat == bad, so therefore Systemd == bad

It's one of those silly arguments that is repeated, by the "OMG, but wayland doesn't have network transparency" crowd. Everyone else just gets shit done (like the pulseaudio crowd) which is why this stuff was heavily adopted