Re: [dev] suckless distro
On 06/24/2014 04:20 PM, Dimitris Papastamos wrote:
>
> Systemd is not the only issue.
>
Specifically, maintaining a stable platform is something of an
impossibility. Upstream fixes are too-rarely backported
(and what the **** else would I use a distro for?), and so
when Heartbleed_2.0 or whatever comes out, rather than
just having it fixed with a security update in my existing
stack, I end up having to install the upstream version
bump that fixes it, which pulls in libfoo3.0,
foo-gobject-introspection, and of course systemd.
There was one point a few years ago (back when I used
Arch) where one morning I ran pacman -Syu (or whatever
it is) and _it removed HAL_. (No possible chance I could
still need _that_, right?)
Arch has what I call a "Heraclitus problem" (from his
aphorism "you can't step in the same river twice").
It moves too fast and explicitly doesn't have a goal
of being a stable platform. When you're an admin, bleeding
edge just leaves you bloody.
It's like running Sid -- and I knew Ubuntu was going
to be a disaster when I read Shuttleworth talking about
"the joys of Sid"... oy vey... Ruby has that too. I'm
not much of a programmer, but I'm an admin who manages
the platforms programmers use. I want to know that a
system I install today will be useable in 2 or 3 years.
Back when I was in a Ruby shop, that got me laughed out
of meetings, because rails or rind or radish or whatever
the devs were using released incompatible major version
bumps _every 6 months_ in some cases. (Yes, I know, always
roll and freeze your own development stack, but still...)
Frankly the least suckfull distro I am familiar with
is the venerable Slackware, which is still full of suck,
but full of vanilla suck that I'm familiar with and is
predictable across upgrades. YMMV, of course.
Received on Wed Jun 25 2014 - 13:12:19 CEST
This archive was generated by hypermail 2.3.0
: Tue Jun 24 2014 - 13:24:06 CEST