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

Thu Oct 23 00:36:09 UTC 2014

On 10/23/2014 12:05 AM, Jeffrey Ollie wrote:

> systemd is a tool designed to get the system to a state where "real
> work" can be done.  NTP servers, DHCP clients, consoles, aren't the
> real work of a system, or at least I hope not, because that would be
> boring to me.

That idea sounds interesting. I can see where you're coming from.
Basically, these things are "details", that shouldn't really be all that
important, so systemd is supposed to take care of it all and leave you
to worry about the actual bread and butter. That about it?

Legitimate question, not trying to be sarcastic: would you concede that
the amount to which something is a "detail" may vary significantly per
the use case, and requirements? On my desktop I might not care about
whatever the DHCP client is doing, or the NTP server, but on a server
that may very well be a different story.

For example, I'm very interested in NTP and accurate timekeeping --
mostly as a personal hobby, but it's been useful at work as well. I for
one would definitely not consider NTP one of those "details" that just
"come with the bootup process".

> [regarding the Leatherman]
> Software doesn't have mass so your analogy doesn't quite work.

Analogies can only be taken so far before they cease to make sense.
However, in the software world you could consider a "mass" equivalent:
code size / complexity. Building on that, just as the Leatherman can't
be as good as all the individual tools combined without weighing too
much, one could say systemd can't be as good as all the individual tools
it aims to replace without being too big and complex.

Program "mass", in this case, may impact anything from performance, to
ease of configuration, to complicated dependencies, to security problems.

I did find it interesting, however, what you mentioned in another email,
about systemd implementing certain isolation features such as separate
filesystem namespaces and so on. That may be very useful.

I think the main point that we could hopefully all agree on here, is
that it would be very difficult to have a single "one size fits all"
solution. The requirements and concerns of the desktop, for example, are
simply too different from those of the server or router space. systemd,
for better or for worse, can't be the one magic bullet. Great or
terrible as though it may be, I don't much like the total break in
compatibility (correct me if I'm wrong). I'm not saying SysV is all that
good, but there are other replacements, and new ones can be designed,
but don't make it so that everyone-has-to-use-yours-or-else! I guess we
have GNOME to thank for that... And that's what troubles me the most:
the lack of choice that seems to be creeping up, with everyone just
ganging up and jumping to systemd like the floor is on fire. I'm with
Jay Ashworth on this one: what gives??

Also, for the person who considered this OT for not being about Cisco,
keep in mind all the world is not a VAX^H^H^HCisco ;)