[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Another system management tool to disappear.

On 08/30/2015 17:49:31,Doug wrote:
> On 08/30/2015 03:33 PM, T.J. Duchene wrote:
> > I know this because I have actually built 95% of base Linux from
> > source by hand - multiple times  -  over the last 17 years. In my
> > opinion, if you don't want to take the effort to do the work, then
> > you simply have to accept other's decisions regarding what they
> > compiled in.
> >
> >  
> /snip/
> That's easy for you to say, since you are obviously a programmer. The
> rest of us may never have programmed anything, and C just looks like
> some foreign language --which it is! What we would like is stability,
> and until Poettering started messing with Linux, we pretty much had
> it--at least in any given distro.

Don't fret, I program in C and I'm only running funtoo and I'm already
way over my head. I don't mind having to read the manual or compile
things, it's the bugs, I must have a collection of at least 100 and the
devs want all non mission critical ones reported upstream. I've
tried to report 6 and most of the time I can't reach them or they don't
respond. I've gotten 1 reply, and I've decided to help him with his
project, menumaker.
In short, there is no way you could have enough time to maintain your
own system that you create from scratch unless you have a serious
amount of education, time, and working config files. Even then it would
be a nightmare.

On 09/01/2015 2015 23:11:51, Joel Rees wrote:
> The issue is not whether they should be leading their projects or not,
> the issue is whether what they build is really appropriate for
> becoming a necessary part of all major distributions of Linux kernel
> OSses.

This is where Lennart goes wrong, you see, sysvinit was, and still is,
AFAIK, easily replaced with another tool. Unlike Linus, Lennart is more
brutal with his software's design such that it is extremely difficult
to replace parts or the whole of it. He also does not let himself be
proven wrong, as it says you can do in the linux kernel mailing list
I keep hoping he'll learn...
You don't have to take my word for it, there have been bug reports
which have not been fixed because the systemd team does not want them
fixed. You want evidence perhaps? I don't have a link on me but there
is the trouble with an encryption package that was ignored, and there
is the issue with you having to reboot every time you attach "new
hardware" which means that I had to reboot my machine every time I
(re)attached my USB to SATA adapter, but not a USB flash
drive, which is really strange. I would have to enter my password
and go into my BIOS to check that the card was properly attached. And
then I must be careful not to bump the adapter. Who cares if the system
boots faster if I have to reboot all the time?
I just wish Lennart was more gentle and benevolent. As opposed to
saying the the future is systemd (I think that was a literal quote,
though it may have been the conclusion I drew after I read all his
writings on the subject.)
Incidentally, he only considers three init systems out of all the
possibilities during his review, systemd, upstart, and sysvinit. We all
know which would win given that comparison.

> If you will excuse me, that's not our decision to make. 
> FOSS is not about making everyone happy or whether something is
> "really appropriate". It is about empowering you to make your own
> decisions. Should you desire, you have the ability to fork code, and
> part company with the herd.
> Distributors have every right to build their Linux as they see fit.
> You have the right to reject that and build your own. That's your
> freedom, and that is all that was ever promised - nothing more. 
> If you ask me, Linux users are far too dependent on binary
> distributions to start with.  They fill a need for those who have time
> constraints, but that's all.  You are willingly allowing others to
> make decisions for you, in return for using the fruits of their labor.
> As such, IMHO, binary distro users have no standing to complain. 

<strong flame_retardant_underwear="on">
That's a rather cold view, even if is is the FOSS view, people depend on
others to make good decisions, and yes, we do have a right to question
their decisions, that's how people learn about what is a good one and
what is not, then they can make better ones. It was one of the
recommendations of St. Paul, "Question all things, retain that which is
true." 1 Thessalonians 5:21. If a faith based religious leader can
recommend that, surely we who are (think we are), more into the
scientific method can also.
As systemd makes both testing and changing much harder, I think you can
understand why it's a poor choice. This is not an opinion, try to get
your system free of systemd, may are the posts I've read where the devs
say that it is impossible. And, as systemd continues to take on the
functionality and often times in place of what used to be modular
software, I think you can see how something that's free is not
necessary something that's what you need. If I sent the starving people
in Africa sugarless bubble gum should they thank me? If they don't want
it they can fly over here and return it to the store. This may be a
drastic example, but the point remains that many, many, three billion++
many people are in the same situation when it comes to software. They
can't help themselves, like Doug. And I can't spend my whole life
maintaining my system even though I might be able too. Look at RMS, he
is a prolific software writer and a master hacker, and yet for all his
skill he still asks for help, all the time. And he now has carpal,
tunnel syndrome from what I've read, so all that work has taken a nasty
toll. He probably uses a distro like you too.
I don't think you have a clue what your proposing for Doug, I do. If
you really think that it's possible to create your own fork, then I
challenge you to do it, don't forget KDE and GNOME support. Systemd is
integrated with udev which is tied to most of the graphical tools on
the system, including the xorg-server, vlc, KDE, GNOME, and many of
their utils, just to name a few. If you want a full list try to
uninstalling udev. That's just one tool systemd is into. It's also into
login, GNOME, kdeinit (or so I've heard), and there may be more as
systemd supports actions that xinetd, cron, su, and syslog would
I think I might make several more really good arguments, but I think I
should rather encourage you gentlemen to start reading systemd's code
base, as I did, and all the documentation, as I did, and all of
Lennart's systemd postings, like I did (actually I missed one), I don't
mean the systemd dev's mailing list, and how about all the intelligent
criticism his systemd has drawn, like I did, and some of his mailing
list postings, LIKE I DID. The ignorance brought into this debate is a
problem. FOSS != FOSS unless they are copies or one is drop in
replacement for the other. There are always trade-offs, and almost
always, queer features.
init system's don't need to do more then boot the system, systemd does
much, much more then they do and requires much more then they do, thus
exposing us to a rare kind of bug known as mono-FOSS-lock-in (mono does
not just refer to the init system).

DOUG, There is a fork of Debian called Devuan, I talked on their IRC
channel and people think it's stable, I'm going to try it. Would you
or anybody else be interested in the results of my test?

BTW: I'm not familiar with why sysvinit was a problem, links, anyone?

Taking a humerus approach to this debate, if systemd continues the way
it is going then we will have to irk RMS by calling our systems,
systemd/Linux. Even more interesting is that systemd does not support
Hurd and Linux does not support a micro kernel and thus lacks
functionality, like su, so systemd's devs will have to create a new
micro kernel and then your systems will just simply be called Lenux.

Sincerely, David

Reply to: