I get it now. The distro ecosystem has been moving towards the other half of Sutherland's wheel of reincarnation - in this case, from modular systems towards mo

I get it now. The distro ecosystem has been moving towards the other half of Sutherland's wheel of reincarnation - in this case, from modular systems towards monolithic ones. First it was stuff in the desktop space - now it's creeping into other areas. You can state the particular point of inflection as a terrible thing, or a good thing, but seen as a technology learning cycle, it becomes _necessary_ even if the implementation chosen is shit and you hate everything about it. At the other end of it, it still gets broken up again into a different set of modular components whose design wouldn't have been apparent at the beginning. Since the initial discussion stage is over, I would say, if you really think this is the wrong solution(and I don't consider myself expert enough on inits to cast much judgment here), your work is best spent on the code environment immediately surrounding systemd - not direct competitors, but stuff that it has to deal with at some point. That will put pressure on the design to either break entirely or evolve to support whatever you've come up with - and when that happens you get to walk in with a patch to show and kick off another firestorm.