Mailinglist Archive: opensuse (1658 mails)

< Previous Next >
Re: [opensuse] Re: [opensuse-factory] equivalent to boot.local for systemd
  • From: "Brian K. White" <brian@xxxxxxxxx>
  • Date: Thu, 08 Dec 2011 18:30:57 -0500
  • Message-id: <4EE148B1.607@aljex.com>
On 12/8/2011 5:04 PM, John Andersen wrote:
On 12/8/2011 1:58 PM, Brian K. White wrote:
Yes it's more complicated to reproduce certain actions that used to be simple,
in trade, it makes other actions that used to be impossible,
possible and even relatively simple.

There never were actions that were impossible. There was always a way.

As someone who has had to write start scripts that handle situations that were never imagined when the basic scheme sysv init was imagined.


It might have taken a bit of hacking but it could always be done, and hacking
was seldom if ever actually
necessary with sysv. It did 99.44% of things automatically the right way.


What systemd seems to do is to turn that upside down, making everything complex
so that
the very few difficult things seem easy, even when those difficult things were
rarely ever done.

sysv init is like DOS. It "worked fine" "for years and years". And in fact it did. But eventually the world channged beyond the point of such a simple system to remain useful. People just plain needed to use more than one program at a time and TSR's were'nt enough and even multi-user dr-dos/novell (hey we're on topic by two vectors!) wasn't enough.

Although sysv init is almost infinitely flexible by the simple virtue of being mostly made out of shell scripts that you can write anything you want in to, there still are various basic assumptions that are no longer always true and they do get in the way and they will only get worse not better nor even stay the same. sysv init is a simple concept that is because of that very simplicity not able to handle the kinds of situations that exist today and tomorrow.

At some point, the net payoff in backwards compatibility from torturing the old system is less than the net payoff in functionality usability from abandoning the old system.

Volume control analogy = Fail.

I disagree.

Don't get me wrong. I don't think systemd is quite ready to be shoving down everyone's throats just yet either.

I particularly think opensuse is in a crap state right now with the systemd integration being half-baked, documentation being wrong or incomplete, backwards compatibility shims being incomplete or non existant, making ?? how many years? of google results all broken overnight and breaking countless already written software over night, much of it old no-longer supported but still used and necessary commercial software, all broken, and with the new forced-march release schedule and short support life, the option to just use an older version that won't break everything will go away in mere months. But these are all suse distribution and implementation problems not really systemd problems.

I just find all the arguments I've heard against systemd as a concept kind of stupid so far. People who don't happen to perceive a problem are just assuming that this actually means there IS no problem. How nice it must be to be so sure about everything by pure inference and assumption.

http://25.media.tumblr.com/tumblr_lu3s15NBmF1qm4zvuo1_400.jpg

--
bkw
--
To unsubscribe, e-mail: opensuse+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: opensuse+owner@xxxxxxxxxxxx

< Previous Next >