Linux: concerns over systemd adoption and Debian's decision to switch

Wed Oct 22 23:07:01 UTC 2014

On 10/22/2014 02:43 PM, Rich Kulawiec wrote:
> A Leatherman pocket multitool is highly useful: I've had one for years.
> It's great.  Until you need two screwdrivers at the same time...at which
> point it becomes obvious why serious mechanics/craftsmen carry around a
> toolbox with dozens of tools and why glomming all of those into one
> supertool would be A Very Bad Move.

I find this analogy very, very interesting.  I'm looking at my
Leatherman tool, and see that every single tool is based on a standard;
nothing "invented here just for the Leatherman."  The Phillips
screwdriver is the standard P2 shape, not some off-the-wall grinding.
And so forth.

> Similarly, the monolithic (and ever-expanding) nature of systemd is a
> strategic design error.  That's probably not obvious to people who
> measure their experience in years instead of decades -- it wouldn't
> have been obvious to me back in the day either.  But it's pretty clear
> from here, and dismissing it as "hey you kids get off my lawn" geezer
> whining is not going to advance the discussion.  

It's one thing to integrate a bridge to existing functions into a
system.  It's another to re-invent things for the sake of re-inventing
them.  If you have a perfectly good DHCP package, work *with* it instead
of going through the trial of building and debugging your own version.
Why duplicate effort?  Why have two of something that has to be
maintained instead of concentrating on improving the one you have?

Now, sometimes you do have to sit down and say "this is hopeless, let's
start from scratch" -- particularly when it was done to a bad spec and
tried to do too much without step-wise refinement.  (healthcare.gov,
anyone?)  But even when HeartBleed-SSH was forked, it wasn't abandoned
and started from scratch; the BSD people just pruned all the non-POSIX
guck and stripped it down to essentials.  And did a proper security audit.

We don't need systemd to turn into a hydra.  If there is an issue with
NTP or DHCP or whatever, systemd developers should go to the upstream
and fix the issue in that package, not clone its own version and,
perhaps, make the same mistakes all over again.