[sf-lug] Byfield's "Verdict" on systemd

[Rick Moen]

Quoting Alex Kleider (akleider at sonic.net):

> One of the main take home messages I'm getting (and I guess it
> shouldn't come as a surprise) is that perhaps one might be better
> off without Gnome and time would be better spent trying to develop a
> system without it rather than fuss about the presence or absence of
> systemd.

At the risk of overstressing the point, there's nothing wrong with
liking GNOME and holding the view that its fairly high system overhead
and complexity are justified by what it brings to the table.

Many years ago, when I worked at VA Linux Systems, my friend Rob Walker
(who worked there as one of the sysadmins) insisted that I spend some
time at his desk seeing what he could do with a then-cutting-edge KDE
desktop system.  The supporting hardware was (fortunately) a pretty
beefy machine with lots of RAM and a fast hard drive, so the performance
didn't lack for user-responsivement -- and Rob was correct, that KDE'S
central facilities for K-apps made possible some ways of working that
otherwise weren't available.  And so, I don't dismiss people valuing
such things, and thinking 'This is the reason I have a beefy machine, so
it can do desktop-ey things under the direction of heavy-weight
software.'  It's a valid choice.

I just personally prefer something a _lot_ closer to the classic Unix
X11 minimal workstation setup, with a relatively lightweight window
manager that stays out of my hair.  As a side-effect of that, I pretty
much automatically get low system complexity, higher reliability, and 
excellent performance.  But other people would want their integrated
graphical file managers and DE applets and graphical front-ends to
everything, without which they would feel lost.

Linux-based systems can be either thing.