@chris_tarquini @Hacker_Horse @TheTechStewart @CarlCravens The thing is, it demonstrably doesn't do things well - its graph theory is broken/unproven (in the scientific sense), it has bloat (qr writer and httpd), and scope creep outside of a system configuration layer (ip forwarding and masq as well as firewall controls) and .. 1/n
@chris_tarquini @Hacker_Horse @TheTechStewart @CarlCravens changes well-established existing behaviour (KillUserProcesses was switched on by default to fix an issue that's still not fixed) _without proper deprecation notices_, and hardcodes values which are questionable (putting aside whether a configuration layer should handle DNS).
@chris_tarquini @Hacker_Horse @TheTechStewart @CarlCravens SSI was the first example of what's come to be known as a configuration layer, and that comes from distributed computing where provable graph theory is mandatory. SMF/Solars Fault Manager and launchd also managed to implement graph theory thats provable.
@chris_tarquini @Hacker_Horse @TheTechStewart @CarlCravens Configuration layers (which systemd, launchd, SMF/SFM all are) have benefits and downsides which can be discussed, but that's a completely separate discussion to have which has no relation with the specific issues that people have raised with systemd and which get shouted down.
