On Thu, 26 Sep 2019 12:59:28 +0200
"Dr. Nikolaus Klepp" <dr.klepp@???> wrote:

> Anno domini 2019 Thu, 26 Sep 10:13:56 +0200
> Arnt Karlsen scripsit:
> > https://www.theregister.co.uk/2019/09/25/systemd_inventor_home_directories/
> >
>
> "those who do not understand Unix have to reinvet it poorly"
>

Show of hands: Who ever perceived a problem with using /etc/passwd,
/etc/shadow, and /etc/group? They just worked .

If this is mainly for laptops, why the attention to home directories?
On a one-person computer, why have data mixed up with config, cache,
and who knows what hosted in the home directory? First thing I do with
a laptop is create a /d tree, owned slitt.slitt, to hold my real data.
Easy to back up, easy to LUKS, survives upgrades and entire OS changes.
This just isn't needed.

But of course, the real problem is complexification. All sorts of
databases to keep all sorts of metadata: Lose a piece and you never get
your drive back. Years before Poettering became a public idiot, I wrote
the following:

http://www.troubleshooters.com/lpm/201202/201202.htm#_editors_desk

Basically, Kmail became dependent on a huge database called Akonadi and
an always dragging lookup facility called Nepomuk, to the point that
you couldn't fix things by moving files around or changing a config. I
bailed out of Kmail, and six months later out of KDE. Today I have not
one KDE app or library on my system.

Poettering is making yet another mess. I wonder if IBM gave their
blessing to this. IBM all these years has made good money from Linux
without obfuscating it. I wonder if wunderkind Poettering is making
IBM's new purchase a problem instead of an asset.

I wonder what the effect of a few thousand letters to IBM, stating that
we don't appreciate the complexity of SystemD, would have. After all,
IBM weren't the fools that hired Poettering, and IBM is perfectly
capable of making money from Linux without complexifying it to create a
monopoly.

Here's a nice bumper sticker sentence:

"Don't let your computer be Lennart's hobby: Use Devuan!"

I think this is a pivot point. If Poettering succeeds in foisting this
on the Linux universe, we all might as well buy Apple Macs, because
we'll be increasingly isolated until hardware isn't even compatible to
us. But if we can convince folks like IBM and Debian that systemd is a
messy piece of crap that will make costly problems forever, things
might turn our way.

And this time, in 2019, they can't pull that "we're using a 40 year old
init" crap anymore, because s6 is maintained, top quality, and easy to
install.

Does anyone know of people in IBM who would be appropriate receivers of
letters pointing out that systemd is an unnecessarily convoluted
problem?

SteveT

Steve Litt
Author: The Key to Everyday Excellence
http://www.troubleshooters.com/key
Twitter: http://www.twitter.com/stevelitt