[mdlug] Completely replacing Windows 98se with Linux!

[Aaron Kulkis]

Jonathan Billings wrote:
> On Mon, Aug 18, 2014 at 12:46:37AM -0400, Aaron Kulkis wrote:
>> Distros are choosing it because once one deamon has been written using
>> a systemd library call, then ALL of systemD gets pulled in, which then
>> makes it EXTREMELY difficult to use any other startup system.
>>
>> It's not that it's superior, it's that Poettering has deliberately used
>> poor programming practices (such as breaking the principle of
>> interchangeability)
>> to lock-in systemd, regardless of whether it is actually wanted or chosen
>> on the basis of merit, or not.
>
> That's not a particularly good argument for why distros would choose
> it.

Oh, REALLY.  If half of the standard deamons have systemd code in them,
then EACH of those deamons' packages will call in the systemd package.

WITHOUT calling in the systemd package, the deamons will fail when they
hit the function calls that are in the systemd package.

POetering has written "Linux isn't like Unix, you can do whatever you
want because there are no rules."

The man has no respect the rest of the community.  This behavior is much
like a 3-year old having a tantrum on the floor "I *WANT it!!!!"

He justifies all of this with 200-page design documents which he knows
NOBODY is going to read... then uses the fact that nobody has read his
ionterminably long sagas as a way to disqualify anybody from criticizing
what he's doing.

The fact of the matter is, an init program SHOULD BE SO SIMPLE THAT IT
DOESN'T NEED A 200-PAGE DESIGN DOCUMENT!... nor should it EVER need to
be stopped and restarted, because it should be so simple that it never
needs to be updated.

This all reminds me of the way Microsoft puts out things that are fundamentally
broken, in the same way that if they were in the automobile supply business,
we would be seeing ads like "New Microsoft Tires -- now with 64 sides!"
because so many of the ideas they get from the Unix world, there is some
essential quality (such as roundness in tires) that they just never get,
for whatever reason.  Example -- Windows "shortcuts" which work perfectly
for executable files and directories but are completely broken if used
to point to a data file -- instead of making a file-reference that just
says "this isnt' the real file, the real file you're looking for is
over there." they instead built a couple of special cases (directories,
.bat, .com, and .exe) and said to hell with everything else.


Poor programming is poor programming.  Replacing code which is long in
the tooth is generally a good thing.  Replacing code which is long in the
tooth which creates a whole bunch of new problems, some of which impact
system stability (example: updating systemd REQUIRES a reboot!).

Eric S. Raymond wrote some comments a few months ago about systemd.
While admitting that he had not read all of the code, he noted that
the init process should be one of the absolutely simplest pieces of
code on the system, so that it can't have bungs