The need to replace sysvinit was/is valid but systemd is truly a horrible piece of software. I would argue that systemd has (greatly) increased overall complexity and simply camouflaged it by moving most of the logic out of init scripts (unit files) and into program logic. And then added an http server and binary logs.

The original submission title, while editorialized, was pretty apt in my opinion. There really does seem to be a core group of people involved with the GNOME, RedHat, and systemd projects that seem to want to add features just for the sake of adding features, or perhaps so that they have something to claim ownership of. Gnome basically decided that text-basd configuration was too sane and opted for a mix of config files and the Windows Registry (gconf, dconf) which is a hilariously poor design choice. One-size-fits-all doesn't work unless you are Apple and your users will happily swallow your marketing BS and wait in line to buy a new iphone every 2 years.

The startup world is rife with the notions of "out with the old, in with the new" and "change for change's sake" so it's not surprising to me that a great deal of HN readers are Poettering apologists.

But to me, it sure seems like systemd is just a chance for Poettering and some others to pad their resumés, or else perhaps simply an ego trip. I honestly can't find anything about it that makes me go "hmm, that's a good feature that outweighs all the other negative parts".