Posted by Aaron on Sunday, December 21, 2008, at 8:24 am.
I have a love/hate relationship with Debian.
It is a solid operating system, with a lot of potential. When initially
released in 1993, then following the subsequent years, Debian wan
innovative. It was one of the first distributions to introduce package
management that resolves dependencies, and it did so well. Leaving
Debian decoupled, so it could also run with the HURD, FreeBSD and
NetBSD kernels, albeit still very much in alpha stage currently, make
it a universal operating system. Nevermind the fact that it is compiled
for fifteen CPU architectures, and has over 20,000 packages. Lastly,
it's core philosophy or removing all firmwares and non-free software
from the core operating system make it a viable operating system for
GNU.
There are plenty of other advantages that just make Debian one of
the best Linux distributions out there. But, then there are
disadvantages, one of those being decentralization. Debian isn't backed
by any major corporation. As such, Debian has a massive world-wide
community surrounding it that oversee the future of its success. This
community is beginning to tear itself at the limbs. But this is nothing
new. Debian developers, over the past few years, have earned a
reputation for griping, pissing and moaning about each and every thing
that comes across the -devel mailing list. Many thought Sarge would not
get released, and the Debian project was dying a slow death, due to
these unfortunate events. Yet, Sarge made it. Then came Etch. Etch
didn't see quite the drama Sarge did, but it had its fair share. Now
it's Lenny, and the flames are white hot. There is more tension in the
project than I've ever seen.
This is the hate relationship I have with Debian. It's lost it's
focus. It's gotten overly concerned about the nitty-gritty, and can't
see two feet in front of itself before tripping. Will Debian die? No,
probably not, but it won't be a happy road for the traveler who wishes
to join and commit. Lenny will release, and Squeeze probably will as
well, but I can just imagine the amount of flame wars and personal
attacks going on to make it happen.
Now, I'm no Debian developer, and it's not my goal to become one.
However, here's what I would like to see from the project: innovation.
When I was a Linux instructor for Guru Labs,
I had the opportunity to teach RHEL, Fedora, SLES and OpenSUSE. Not
only teach these operating systems, but write courseware for them as
well. I was subscribed (still am at the moment) to several mailing
lists on each of the operating systems to see what Red Hat, Novell as well as Canonical are pushing out the gate. What was interesting to me was to watch some innovative products come out of the door:
Red Hat:
First on the list is Spacewalk.
This is the Red Hat Network that once used to be proprietary, and
controlled on Red Hat servers. Now open source, any company can deploy
their own RHN in-house. Next would be oVirt- a new way of managing virtualization. oVirt is a GUI fontend to managing KVM virtual machines. It has plans to support managing Xen guests in the future. Take a walk through the screenshots
to get an idea of what the application is all about. Of course, we
can't forget KVM. With acquiring Qumranet, Red Hat is now positioned to
offer a a serious alternative to VMWare, for virtualization using oVirt
with it. KVM is a kernel module that make the Linux kernel a
hypervisor. No one can dispute the lines of code that Red Hat
developers have put into this module, or the kernel for that matter,
either. Lastly, Red Hat has made other innovative moves in the field.
JBoss, Red Hat Directory, RPM (which is superior to DPKG), the Anaconda
installer (more powerful and flexible than the Debian installer),
Kickstart (superior to preseed), the SELinux targeted policy, many of
the system-config-* tools, and plenty of others. Red Hat innovates, has
a rock solid operating system, and a large community surrounding the Fedora project.
Novell
The most obvious coming out of the Novell camp would be AppArmor. AppArmor is a MAC that could be compared to SELinux. With Miguel de Icaza on board, Mono
has become a serious development platform for .NET on Linux. It's
stable and implements nearly all of Microsoft's .NET core features.
Also, Novell has released the OpenSUSE Build Service
as a way for developers to get their software packaged for many
distributions on different architectures. This makes it easy for them
to reach the largest Linux user base they can with little effort.
Lastly, Novell has really made the desktop something to work with. It
has changed the way the Linux desktop looks, making it far more
pleasing for the average employee and consumer. XGL is default, the
panels and menus reconfigured, and all around, the SUSE Linux
Enterprise Desktop yields a very familiar Microsoft Windows look.
Canonical
Here's the camp that I stand in. The first major innovation could be
called Ubuntu itself. Many say this is Debian done right. It has
regular 6 month releases, that are synchronized with the GNOME
releases, with every fourth release termed "LTS" for long term support.
It has reached more Linux user than any other distribution. It has gone
places other distros have only dreamed of. It comes with the best
hardware support out of any distro, hands down. The single CD install
could also be called innovative, as requiring users to download an
entire CD set, or DVD could be asking a bit much, even with todays
bandwidth. Canonical has introduced upstart,
a replacement for the standard System V Init system that we have been
using for ages. It's gained popularity, and has now found it's way in
Fedora, which means we'll see it in RHEL for sure. Canonical has
introduced its own version control system bazaar. It could be compared to git or mercurial as a distributed VCS. It's used on the other innovative product Launchpad, for PPAs, or "personal package archives". Launchpad could be thought as the SourceForge
killer, as it allows developers to host their code, integrate a bug
tracking system, provide documentation and encourage collaboration.
Launchpad features translations through Rosetta and has the ability to
manage team memberships Launchpad is open source through the Storm ORM, written in Python. The only thing it lacks so far, IMHO, is a wiki.
Now, I ask the Debian team- "What innovation have you pushed out the
door recently?". Honestly, I can't think of anything, aside from
wanting to remove every last binay blob from the system, and have The
One Free System. Now, maybe this isn't your focus. Maybe your focus is
to get Debian on all the CPU architectures and kernels you possibly
can. Maybe the focus is to take the innovations from others, and make
it available as a Debian package. I don't know, but from where I stand,
I see Debian walking down the same road as Gentoo. All the internal
turmoil and problems could tear you apart, make users switch to another
distribution, and leave the state of the project in question. I hope
this isn't the case, but a re-evaluation of your goals and future
should be outlined.
I titled this post what Debian means to me, but haven't gotten to that in full. Let me state outright what Debian means to me:
Debian to me means Freedom. Debian is the foundation of which Ubuntu
is built. Being an Ubuntu advocate and user, I want to see the tightest
relationship possible between these two projects, as I realize Ubuntu
would not be in existence, if it weren't for Debian. However, Debian
runs one of my servers in my basement. It also runs on a virtual
machine I need for testing and development. It's been on my laptop a
couple of times, and it's on a 50MB business card CD in my wallet,
should I ever need to do an install, or rescue a Debian system. Because
of the Free Software stance Debian takes towards software, it will
always be very close to me.
Debian means Ubuntu will be successful. Ubuntu just can't
exist without Debian. We realize this, and hopefully, we are doing
everything we can to contribute back to the Debian project, so it can
see the same success Ubuntu is seeing. However, if Debian were to rip
itself apart at the seams, then this means the future of Ubuntu hangs
in the balance. Many Debian developers probably don't care much about
this, but losing that symbiotic relationship could be for more
disastrous and more reaching. It's not just Ubuntu, but Knoppix, Damn
Small Linux, and so many of the other Debian-based distributions.
Debian is key to a large portion of the Linux ecosystem.
Debian means being the universal distribution. I don't know
of another operating system that is compiled for four kernels and
fifteen CPU architectures. Gentoo is close, as are the BSDs, but aside
from that, there just is no other distribution that has that wide range
of hardware support. When I install Linux on a SPARC, for example, I
can rest assured that Debian will install on it. What if I'm faced with
an IA32 or IA64? Debian will install. MIPS? Debian. ARM? Debian. Alpha,
PowerPC, S/390, and many others all Debian supported. EEEPC,
PlayStation 3, XBox, servers, desktops, laptops, PDA, phones, etc.
Debian is far more reaching than any other operating system out there.
It truly is the Universal Operating System.
I hope the Debian project can get itself out of the mess it's
currently facing. I hope it can re-evaluate its focus, and get back on
track. I would love to see the success of Debian. I would love to see
innovations come out of the Debian project. I would like to see more of
the developers involved, and not package maintainers, although they are
key as well. Let's see Debian reach new heights that it hasn't reached
before.
Link to original story
Useful Links