Why Poettering's Answer Isn't
Lennart was kind enough to condescend to our level and provide his wisdom on the matter of systemd. It was a masterful piece filled with insight on the order of "from my point of view, it is the Jedi who are evil."
You'd have to be a software engineering genius like Poettering in order to understand that a values judgment can be objectively false. Only he is capable of knowing what will be valuable to users and what will be difficult for them. Any of you users out there who state otherwise are just wrong.
My favorite feature of this blog post is that he calls the belief that "systemd is not Unix" a myth and immediately proceeds to brag about how non-Unix it actually is, making it a point to berate actual Unix systems and then proceeds to claim that accusations of him being unkind to actual Unix systems are also mythical.
It's equally important to see how well he addresses non-arguments such as the accusations that petulant contrarians make about Lennart's intentions. Those of you who were actually hoping to have a real argument about the software design don't understand; if you just accept that the intent was not speed, that the intent was not change for change's sake, and that systemd isn't only for desktops, then you will understand the wisdom of its engineering perfectly.
I also love axiomatically false statements. We're told about the myth that goes "systemd could be ported to other kernels if its maintainers just wanted to." He soundly rebuts this by proceeding to list all of the reasons why he, understandably, doesn't want to but never quite explains why he couldn't port cgroups, fanotify, etc. to other systems.
So systemd can't be ported to other kernels. It's unpossible. But distros that use other kernels can still use it if they wanted to! I mean, why wouldn't they? After all the glorious features that systemd offers, they really can't afford not to use this thing that is "not the mafia." But introducing all these glorious features is not feature creep; it's just "empowering the user" with… features.
It's true that there really are a lot of bad arguments against systemd out there. The few of them that are objective arguments really didn't warrant this big long brain-dump; one need only show a shell script running in systemd to disprove the claim that it's not scriptable and that it's a binary configuration. But proving things is for losers; Lennart's time is better spent writing long-winded, vacuous nonsense that he hopes will manage to clear up your misconceptions.