I don't know the intricate details, I do know that updating Raspbian on a Pi2 to use systemd reduces boot time from 30 seconds to 15... that's pretty cool in my book.
Let's be clear about what is going on: some of that time reduction is because systemd doesn't do things like checking that an IP address is not already in use on the network when bringing up an ehternet interface (at least, that's the default setting in systemd)
I can see how shortcuts like that might work better in some cases than in others. Personally, I like being able to power down my computers - especially now that they don't have spinning hard drives that temperature cycle when you do that. When I power them back up, I'm not thrilled by 5-7 minute wait times before I can use them (think I'm exaggerating? try a corporate secured windows image...)
So, if systemd tries something like reusing the last known IP address and hoping for the best, that definitely works in my home network, and I hope that they do something sensible like detect the conflict after initial boot and deal with it if/when it is found? I doubt they go so far as to tailor the boot process to check before enabling the interface when conflicts were found in 2 or more of the past 10 boot cycles, but they could.
Anyway, if I have a point, I guess it is that competition is good, haters gonna hate on systemd - and I understand some of the reasons, but if systemd isn't out there mixing it up and showing what can be done to improve the user experience, the established competition won't have as much incentive to improve.
And, when a tech like the OP devolves into something that just doesn't have relevance in the modern world, letting it go obsolete is the right thing to do.