On Monday 22 October 2012 13:57, GangGreene conveyed the following to
comp.os.linux.misc...
> On Mon, 22 Oct 2012 06:55:54 +0200, Aragorn wrote:
>
>> On Monday 22 October 2012 01:05, GangGreene conveyed the following to
>> comp.os.linux.misc...
>>
>>> What linux is UNIX like? You have to be kidding!
>>
>> Slackware, Arch, Gentoo, LFS... :-)
>
> I have my own LFS distro, based on my needs and in the future it will
> not be what I what and need because of this insane feature creep.
>
> Linux is going too far away and is pursuing change for the sake of
> change, like the systemd hyperbole, all of linux will be systemd in
> the future. You can see that once the change is made there will be
> nothing else, it is even creeping into the kernel.
Hmm... The thing with systemd is that it is developed at RedHat, by
Lennart Poettering, who is - to put it mildly - rather arrogant, and
disdainful of other distributions. Also developed at RedHat, by Kay
Sievers, is udev.
Now, Sievers and Poettering are working on streamlining and integrating
systemd and udev in such a way that you will not be able to use udev
without systemd anymore in the foreseeable future, because only RedHat
and derivatives matter, and all other distributions are irrelevant, in
their opinion.
However, these other and irrelevant distributions - such as Gentoo - are
already working on alternatives. For one, Gentoo offers the option of
using an alternative /dev handler, mdev, and the other solution is to
fork udev. Part of what brought this about was Sievers's and
Poettering's screwup on account of having the both of them invoke
scripts which were living under /usr before /usr was mounted, which
Sievers and Poettering then "made more consistent" - so as to have it
appear as if it was intentional - by installing systemd under /usr/sbin,
rather than /sbin, and which was retroactively justified as being an
intended merge of /usr with the /bin, /sbin and /lib{64} directories,
"for greater compatibility with UNIX, and specifically Solaris".
The merge of /bin, /lib and /sbin with their counterparts under /usr may
indeed be as certain versions of proprietary UNIX have already been
doing it for quite some time, but it does break the ability to boot a
Linux kernel without using an initramfs if you happen to have /usr on a
filesystem separate from the root filesystem itself. This did not make
a number of people happy in the Gentoo community, where people are
accustomed to rolling their own kernels and foregoing the use of an
initramfs or initrd.
So Sievers and Poettering may think that they are deities and that they
can control the future of GNU/Linux development, and indeed, a number of
distributions chooses to rather comply with the mainstream rather than
go against it, but Sievers and Poettering have obviously overlooked that
very specific property of the FLOSS community called "forking".
Whenever something evolves into a direction that generates rejection by
a sufficiently critical mass, the project will get forked.
As an example, look at KDE4. The distributions themselves were pushing
KDE4 long before it was production-ready, and as a result, several
developers chose to fork the old, trusted and reliable KDE3, and this
has then resulted in what is now called Trinity Desktop, a highly
stabilized and modernized version of KDE3, currently at version 3.5.13
or whereabouts.
Similarly, not everyone was happy about GNOME3, with as a result that
alternative GNOME Shells were developed such as MATE and Cinnamon, and
they are quite popular.
The bottom line is that, yes, the mainstream will follow the course set
by a few narrow-minded individuals, but there will always be resistance
to that course from developers who do not agree with the mainstream.
On account of the BSDs, I can only say that, first of all and in light
of the above, BSD follows the Cathedral model, rather than the Bazaar,
so it's not evident to fork something over there, unless you're forking
the entire distribution, as was the reason why NetBSD and OpenBSD came
to be.
Secondly on account of the BSDs, while they generally may be reliable -
and OpenBSD is probably the most secure one - hardware support in BSD
lags behind on the Linux kernel by several years. Just think about the
proprietary video drivers from nVidia or ATi, but that's just the tip of
the iceberg. There's a lot more hardware which is supported in Linux
but not yet in the BSDs.
I'm not trying to advocate in favor of GNU/Linux and/or against BSD
specifically. I'm just listing some facts to consider. In the end, the
choice is yours, but suffice to say that, if you do stick with
GNU/Linux, you are probably better off with a non-RPM distribution,
because most of those follow the template outlined by RedHat.
I am running Mageia 1 on this machine here, with Mageia 2 already being
the mainstream now and Mageia 3 awaiting release, simply because Mageia
1 still doesn't use systemd.
Like somebody else said in this thread, systemd is a solution looking
for a problem, and it attempts to do much more than simply replace the
traditional System V init. For one, it also attempts to replace
(x)inetd, and quite frankly, it isn't even half as transparent to the
user as either one of those.
Ultimately, it all depends on your own planning. If you're someone who
favors a long term support platform, then you could already bite the
bullet and check out Gentoo now, but if you're someone who regularly
installs new distributions, then I would advise you to use Debian or
Slackware for now, and the same advice applies to the OP. ;-)