MadPenguin interviews Bruce Perens about the new Debian Common Core Alliance. “As former Debian lead Bruce Perens discusses in this article, Debian has long played a central role in helping people all over the world share the fun, frivolity, fascination, and occasional frustration that is the world of free Open Source software.”
Not a whole lot of good information. I personally do not care who termed the “Open Source” Phrase, i was hoping for more information on DCC and their plans for the enterprise.
I heard about Bruce Wayne and Bruce Lee but who is Bruce Perens?
Well, you’ve heard about Google, haven’t you? So why don’t you use it?
This is a real turn for the better and very good news. Now many more apps will be portable, over time, between Debian-based distros that are part of the DCC.
With this, the future potential for certificates such as, “Linux Certified Engineer,” to have even more meaning is quite higher due to this action.
Nice interview.
Is this going to be something meaningful, or is it going to be another Desktop Linux Consortium and Userlinux? Those really were a waste of time and resources. I know it’s a different plan, but I’m really skeptical.
I haven’t really been following this that closely, so i might be missing someting, but wouldn’t this hold more weight if Ubuntu were involved? Is there a reason why they aren’t?
The way I see it now is Sarge == Server, Hoary == Desktop. They are THE desktop debian distro, so why not have them involved in this? There really is no reason (other than development) to run debian unstable now because we have an up to date debian distro that isn’t hobbled together like unstable can be at times.
Yes it would hold more weight, but as far as I can see, the ubuntu folks are trying to be to debian, as xorg was to xfree: it’s based on it, but it goes a different direction, trying not to make the mistakes debian made. It’s a shame, it’s realy a great distro, but breaking compatibility with all the other distro’s around doesn’t seem the solution, unless they hope the other distro’s are gonna base their distro on ubuntu.
From what I’ve read about the dcca is that they are gonna try to do the same but starting with debian sarge, and keeping compatibility between distros. So that debian itself can profit from dcca. So that not many things are done twice.
And of course, then the ubuntu folks would be less free to do what they want with their distro.
Possibly. Symphony OS is much different than plain Jane progeny, yet we will both be going the CL path. It justs allows for greater cooperation between the various distri’s.
One of the main reasons I run debian unstable on my desktop is not having to upgrade/reinstall twice a year. I know ubuntu users can skip a release and get security upgrade, but I still want to stay on the edge. And then there is all that brown thing…
Back to the FA, what’s all that talk about SUSE and the DCC?
Yeah, I’d really like them to switch
, then I think ubuntu has lost me again 
, maybe they’re waiting for yast for debian to be ported 
, after all, they have put a lot of effort in that.
Well, for one thing you don’t reinstall. What I do on a non-work desktop is just to wait until about halfway through the development cycle and switch over. At that point, things tend to be relatively stable.
But I had been playing around with gobo and foresight earlier and when I went to reinstall Ubuntu, the snapshot installer was completely broken, and even colony2 was broken until i updated everything.
“There really is no reason (other than development) to run debian unstable now because we have an up to date debian distro that isn’t hobbled together like unstable can be at times.”
But Ubuntu doesn’t *remain* up to date – their stable package repository is frozen between releases, whereas Debian Unstable is updated continuously.
“There really is no reason (other than development) to run debian unstable now because we have an up to date debian distro that isn’t hobbled together like unstable can be at times.”
And the term is “cobbled”.
Sometimes major transitions have to be made (like xfree86 to xorg, or KDE or Gnome updates) and sometimes they don’t go smoothly, leaving people who run unstable with breakage. This usually can be fixed, but for people who rely on their computer to get actual work done, the instability can be very damaging.
A production user would rather wait until the newer stuff is reasonably stable, then switch the sources.list and do a dist-upgrade. These upgrades are more likely to go smoothly with Ubuntu than with Debian, because if you wait 2.5 years between major releases, too many things change, so additional fixing is likely to be required.
“Sometimes major transitions have to be made (like xfree86 to xorg, or KDE or Gnome updates) and sometimes they don’t go smoothly, leaving people who run unstable with breakage. This usually can be fixed, but for people who rely on their computer to get actual work done, the instability can be very damaging.”
FWIW, the vast majority of such breakage can be avoided via use of apt-listbugs. Breakage is not a given in those circumstances – just don’t upgrade the affected packages and stick with what’s already working.
“A production user would rather wait until the newer stuff is reasonably stable, then switch the sources.list and do a dist-upgrade. These upgrades are more likely to go smoothly with Ubuntu than with Debian, because if you wait 2.5 years between major releases, too many things change, so additional fixing is likely to be required.”
Perhaps, but the downside is that the user might be waiting 6 months until the next Ubuntu release. Perhaps even longer, since it’s unlikely to be rolled out on-site on the day of release. Additionally, production environments are likely to consider a 6 month upgrade cycle to be too short.
and then they will be able to say without doing any additional work that their application works on Linspire, Xandros, SUSE, Mepis
LSB? Err, they’re all KDE-based distributions Bruce.
the logical conclusion would be for SUSE Novell to join, because all of the packages that we’re working on in the DCC are packages that SUSE Novell maintains single-handedly, and that does not make economic sense.
Not going to happen. They’re not going to rely on another organisations release cycles, quirks and, let’s be honest, silly politics. Given what’s happened over things like UserLinux in the past, they’re not going to trust their future to that. That’s what really wouldn’t be cost effective.
Having had SUSE Novell join, which I think is an achievable goal and not a distant future goal
He’s banging away at Suse and Novell for some reason. They aren’t going to do it now they’ve got Open Suse, their technology is different to Debian’s and they have existing products to support.
Red Hat has always been a little reluctant about the Linux Standard Base, and I think that they’d like to be monolithic
Well yes, because they’re number one. Despite what every one was saying at the time of Novell’s takeover of Suse, that isn’t going to change.
LSB? Err, they’re all KDE-based distributions Bruce.
No. KDE is a desktop environment, not the base of a distribution. KDE runs on top of a distribution’s base.
No. KDE is a desktop environment, not the base of a distribution.
No, and I don’t know where you get that bizarre idea. For those distributions KDE is an integral part of their distribution because it’s the interface for the use and where applications run.
KDE runs on top of a distribution’s base.
As mentioned – KDE is an integral part of these distributions, because it is the desktop.
Without that, you cannot guarantee that one application will work on another as he’s claiming.
Nice try, but it’s all hyperbole anyway.
At last, an informative, sensible comment on OSNews. Thanks for keeping my spirits up 🙂
Read these comments first
http://osnews.com/comment.php?news_id=11543&limit=no&threshold=-1
i looked at some comments here. these should encompass suggestions about debian as foundation for all new user-friendly distros and what can be done to respect debian as back bone.
Why we are discussing Ubuntu as it has not even joined DCC?? if i am correct..
“Why we are discussing Ubuntu as it has not even joined DCC??”
Because someone suggested that there’s no point in using Debian as a base any longer now that Ubuntu’s on the scene.
I was the one that started the Ubuntu conversation. I didn’t quite mean that there is no point as using debian as a base because of ubuntu. Quite the contrary. I love debian as a base system. I don’t think there’s an cleaner, easier to use system out there.
My point about Ubuntu was that I wish that they’d take part in this organization, so that they can contribute all of the great things they’ve been working on. Also I’d would love to see all of the major debian derived distro’s to be able to be compatible with each other to a certain extent.
And to the person who asked why are we talking about ubuntu…I raised the question half knowing that someone would throw that out there. There seem to be a lot of people that are very irritated w/ ubuntu for some reason, but I honestly felt that Ubuntu was a very realavant topic in regards to DCC, because it is now the most popular of the debian derived distros. It also has accomplished a lot in terms of being that “ultimate desktop distro” that many have strived to accpomplish but in the end failed in one way or another.
Ok Flame on…
Actually, there wouldn’t be any point in Ubuntu joining. It’s based on SID snapshots, not Sarge and it’s more focused on moving Debian experimental projects to SID than backporting SID projects to Sarge, which is the DCC focus. Ubuntu is also less focused on business, so the DCC mandate is a distraction.
As for “Stabilized SID”==”Desktop” and “Sarge”==”Server”, it depends on your needs. If you want an enterprise desktop, then you definitely don’t want to upgrade every 6 months. You’d rather upgrade every 3 years and you’d rather thave all versions look and act the same, so Sarge might be just the ticket. On the other hand, Debian 2.2 was *really* old and some packages (like PHP 4.3, Apache2, etc) were only available in SID (which is unreliable) or testing (which gets no bug fixes) or a backport project (assuming one exists and you trust the maintenance of it). I cases where most of your preferred packages as in SID, a “Stabilized SID” might make more sense as a “server”.
.. and I wonder why.
I thought Perens was clever. But there are absolutely no reasons that either SUSE nor RedHat should abondon their codebases for Debian. Should they throw away everything they have built so far – their knowledge about their own distributions, their documentation, their bug databases, their build systems, their certication programs, just to have unpaid Debian developers making individual decisions on each package?
Unrealistic, to say the least
“But there are absolutely no reasons that either SUSE nor RedHat should abondon their codebases for Debian. Should they throw away everything they have built so far – their knowledge about their own distributions, their documentation, their bug databases, their build systems, their certication programs, just to have unpaid Debian developers making individual decisions on each package?”
I don’t think Bruce is talking about RH/Novell switching from RPM to APT or anything along those lines, and he certainly isn’t talking about such distributors ceding control of their package selections. The key idea is to create a common “core,” which would consist of a Linux kernel, libraries, and some other stuff that would be visible to third-party application developers.
Most of the people making decisions (like always these days) would be paid to influence or make those decisions by firms such as Red Hat or Novell.
I guess he does address Ubuntu:
“So what we’re doing now is that we have gotten really all but one now of the leading Debian derivative distributions to gather to collaborate on having a common core which we will certify to the Linux Standard Base 3.0. “
Actually, I think he may be referring to Libranet. Ubuntu isn’t a commericial Debian distribution (which is what the DCC is focusing on) and its not based on Sarge (which is again the DCC focus).
All this press and debate, and still no definition of where the line between “core” and “higher up the stack” will be drawn in the sand? Not even a loosely defined starting point?
If anyone knows where to find some information on what types of animals will live in the core, please enlighten us.
Well, it will be based on the LSB, which defines the following things, roughly:
ELF
libc
libm
libpthread
libgcc_s
libdl
librt
libcrypt
libpam
libz
libncurses
libutil
modest list of basic commands (ls, make, useradd…)
/dev/null, /dev/zero, /dev/tty
/etc/cron*, /etc/init.d, /etc/profile.d
locales
users and groups (id ranges, root, bin, daemon, lp…)
sysv-style init scripts and runlevels
RPM packages format (!!)
libstdcxx
libX11
libSM
libICE
libXt
libXext
libXi
libGL
The interfaces and data structures are defined for libraries, but the implementations are not.
So the question becomes, what else will fall under the specifications of the DCC? Qt/GTK+? udev/hotplug/inotify/hal/dbus? Java/Mono? Kernel specs of some sort?
Will the package format stick from the LSB to the DCC? Will there be a source package standard? Special treatment for non-free packages?
“RPM packages format (!!)”
err, its more that the distro should support the install of rpm packaged software and librarys. not that the whole distro should be based around the rpm as a install media.
this can be done as simply as including the rpm tool in the base install, or like in debian where you can use the alien tool to repackage the rpm as a deb.
only problem is that when you use shared librarys and mix ways of installing binary files you can end up with a mess if your not carefull.
“sysv-style init scripts and runlevels” can be just as big a problem as you cant simply repackage a change to the startup files. and from what i understand, debian dont use sysv-style startup…
from what i understand, debian dont use sysv-style startup…
Debian does use a SysV style init, it just does it slightly differently than Red Hat. The only major distribution I know of that uses a non-SysV init is Slackware, which uses basically a BSD style init.
from what i understand, debian dont use sysv-style startup…
Debian does use a SysV style init, it just does it slightly differently than Red Hat. The only major distribution I know of that uses a non-SysV init is Slackware, which uses basically a BSD style init.
gentoo also uses non SysV still init scripts
Something tells me this is all going to slither into the sand. The project seems to have no clear and realistically achievable aims (trying to get SuSE on board is a complete fantasy, for example), no great measure of agreement (Ubuntu is on hold and the Far East boys are very iffy), no determined leadership (who is actually in charge here?) and lacks even an agreed name (they cannot pinch the Debian name – it is not theirs to take). The whole thing sounds more like a defensive manoeuvre among the big cheeses at Debian who’ve just woken up and realized the world has moved on a little. As a Debian user, I’m much more interested in solid and modern desktop and server distributions (often updated and security-checked in the case of a desktop) instead of the present unsatisfactory mishmash. That’s the real core Debian needs to address if it’s to remain a credible player, imho.
They do have the best idea for Linux yet. The goal to reduce the impact that fractured linux causes. I do agree that they seem to be lost on how best to go about this. Hopefully they get developer support to promote a standard predictable Debian Linux platform through some kind of DCC cert program. Something like designed for DCC. This would either marginalize other distro’s or bring them into play with the standards game.
The BSDs are the way to go. Linux has become way too mainstream to be used by real geeks, with big companies like IBM and SGI behind it and dictating where it should go. In some aspects it’s not better than the proprietary Microsoft Windows XP piece of shit system.
Sir (or Madam),
I would want to mod you troll, but I haven’t got a login now.
If (and only if) you are a geek then you would neither lokk at linux nor at BSD. You would develop GNU/Hurd or even Plan9. These are the playgrounds of geeks at the moment.
But one thing I share with you: The real fear that big companies will take us Linux away. Yes, this will they do not counting what other people say.
They laughed on us when we told them that TCPA will come and slave us.
They laughed on us when we told them that Apple is evil.
They laughed on us when …you say it
Nice to know there are Debian NetBSD and FreeBSD ports in development. Enough geekiness for annoying geeks like you.
… he let Userlinux die before it was born.
I second that. Seems like he is trying to make a come back, because UserLinux is dead and wasn’t able to create the same interrest as Ubuntu has. I don’t think much is going to happen with this ‘new’ project either.
But he could damage seriously the Debian project using its name without permission from SPI.
The Debian project seems to sleep about that fact.
Debian should be cautious and Perens should be cautious.