[Big Snip] While I'm sympathetic to Linus and his concerns about destabilization, I would not be so 'hissy fit' about it. And anyway, poor execution is not an argument against what systemd is doing, only against poor execution. I've been in the position of a manager responsible for releases of updates [for an international bank] and know that having a stable PRODUCTION system is paramount. I've fought against developers releasing untested code to production, inadequately tested code and on one occasion of trying to do development on the production system. One and only once, when I was dealing with a similar case in another country [I said it was an international bank] did they get one by and it was a major embarrassment for the bank. Stricter controls were subsequently implemented. But Linux isn't MS, HP or IBM, or even Apple. They all have professional in-house pre-release testing (and we know that's often still not enough) and large bankrolls to commit to avoiding embarrassment. We have FOSS and we have just had Heartbleed. As some articles have pointed out, this critical function was done without a proper development budget and without aggressive testing. Where were the "Million Eyes"? We have FOSS. We don't have a testing budget. The "release early, release often" principle applies and hopefully that reflects a progression, but as new features and capabilities are introduced its clearly not a strictly monotonic progression. So we have a dichotomy here. We need a stable platform since Linux is used in so many commercial settings and those are "Business Critical". But at the same time Linux has to deal with virtualization, containerization and more, which is putting demands on system functions that are not met by sysvinit and are met by either systemd or the systemd design. We can't stay concurrent with the demands of technology and business without progressing and letting go of the past. The problem doesn't like with systemd or with Linus maintaining the integrity of the kernel and kernel API. If anything the problem lies, as Heartbleed demonstrates, with a flawed and inadequate testing/release model that demands the only way to test is to release. -- An unstable pilot steers a leaking ship, and the blind is leading the blind straight to the pit. The ruler is like the ruled. Saint Jerome, Letter -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org