Re: Deadline for jessie init system choice

On Sun, Dec 15, 2013 at 1:19 PM, Brian <ad44@cityscape.co.uk> wrote:

On Sun 15 Dec 2013 at 14:03:36 +0000, Tom H wrote:

The goal to have native systemd support in every package with sysv
scripts (if accepted) and a decision on a new init system may be
related, but only the first is linked to the time of the freeze. My
'couple of years' might have been spurious but, considering the stated
goal might only be realised a day before the freeze (or not at all),
the merit of putting a focus on a decision date as pre-freeze isn't
clear.

Personally, I don't see why they feel the need to change. I'm not a big fan of systemd, with it's combined bin and sbin directory trees, it's binary log formats.

And if I am honest, it seems to put entirely too much control in systemd's hands, and if I may be so bold, it gets away from the Unix philosophy. Instead of a bunch of small apps that do one thing extraordinarily well (e.g. grep, sed, awk, sort, uniq, etc.), we now have a large overseeer application which seems to take the "one ring to rule them all" approach. (And yes, I may be generalizing here, since I am just starting my exploration of systemd, but what I have seen so far, I'm not enamored with.)

What's wrong with sysvinit? It's not broken. Hell, the BSDs and Slackware still use a BSD-style startup. I feel like Debian is looking for a solution to a problem that doesn't really exist.

Even if the decision was made in February/March/April would this imply
going into Jessie with a new init system is a realistic possibility?

I'm not even going in to the debate about the political pros and cons for going with systemd vs. upstart. Just suffice it to say that there are an awful lot of distros that depend on Debian for this to be a decision to be rushed in to...It's not like sysvinit is going to turn into a pumpkin...

Just my 2 cents.

--b