[exarch]

>If you ever have written a /etc/init.d script for a python non-daemoning script running in a virtualenv, or something to the same effect, you'd notice the improvement over SysV init scripts.

These are solved problems. You're creating a bunch of low level OS spaghetti because you don't want to have to write a shell script for your software? That is not a responsible decision-making process, that's moving the mountain to you.

[digi_owl]

And making it harder for someone to come afterwards and untangle the mess.

I swear a earlier discussion here on this very site had someone honestly suggest the usage of strace to figure out why systemd failed to boot.

[the_why_of_y]

Countless subtle race conditions and security issues in the shell spaghetti that makes up real-world init scripts indicate that it's not a "solved problem".