[mdlug] Completely replacing Windows 98se with Linux!

[Adam Tauno Williams]

On Tue, 2014-08-12 at 02:31 -0400, Aaron Kulkis wrote: 
> Adam Tauno Williams wrote:
> >> That's true. Iv'e notice that since openSuSE switched from the SysV
> >> init system to the SystemD init system that
> >> my computer has become... sluggish.
> > I've seen nothing like that.  And I can't imagine why systemd would have
> > much effect on overall performance.
> >> And that is for software which was SUPPOSED to improve responsiveness,
> >> right down to faster boot times.
> > You are going from a speculation that this is the cause to an argument
> > against the software.  Perhaps something else is causing the slow down?
> >> Personally, I think it's because Sievert and Poettering are couple of
> >> narcissistic gits, based on their track record.
> > But they wrote code.
> Precisely.  They have a habit of writing code that BREAKS everythiing else
> that interacts with it.... they've now forced their misguided vision onto
> every single deamon that is supposed to be started up during boot.
> This violates ALL of the time-tested principles of good software design in
> the Unix domain ... INTERCHANGEABILITY of systems.

No it doesn't.  The previous model had foisted all of its "vision" onto
how things worked.

Systemd utilizes D-Bus - which finally enables real interchangeability
of systems.  Previously you did not have interchangeability - you have
isolation.  A service had no bloody clue what was going on or if other
services it required to operate were, in fact, operational.

This is no different than everything else.  Drivers being upset that my
desire to not drive is foisting bike lanes or bus lanes on them - sorry,
the answers to their demands have already been foisted on me - they are
just accustomed to their demands being met.  Life is all about foisting
onto other people and living with being foisted upon.

> Yes, SysV init was getting moldy in light of how laptops and portables are used,
> but what they have done amounts to no less than vandalism.

Nope.  They coordinated with lots of people - many services integrate
with systemd.

> All this bullshit of moving executables from /bin to /usr/bin, and then
> making it neigh near impossible to configure a system to have a seperate
> /usr filesystem... why?  Because those two assholes moved nearly all of the
> executables needed to boot (such as mount) from /bin to /usr/bin.
> So if you make a system with a separate /usr, it won't boot.  

So don't do that.  Are you really doing in-place upgrades of non-systemd
distribution versions to systemd distribution versions?

> SysV config files were immensely comprehensible -- most of them, I didn't
> even need to dig up documentation -- what was going on was self-evident.

Of course they were comprehensible - they couldn't do anything. 

> Where have I seen this sort of monolithic monstrosity before? Oh yeah, in
> that giant kludge called windows.

A coordinating subsystem using an IPC bus is by definition not
"monolithic".

crond, atd, and xinetd have needed to die for a long time.  There have
been numerous [syslog-ng, rsyslog, ...] replacements for syslog - but we
just need one, that doesn't completely suck [as syslog does].  Logging,
scheduling, etc... are basic services that should work the same on every
box.


-- 
Adam Tauno Williams <mailto:awilliam at whitemice.org> GPG D95ED383
Systems Administrator, Python Developer, LPI / NCLA