internetnews.com, Sarge vs. The Hoary Hedgehog?: “Ian Murdoch [sic], Debian’s founding father, does not believe Ubuntu’s popularity bodes well for Debian-based distros. ‘If anything, Ubuntu’s popularity is a net negative for Debian,’ Murdoch told internetnews.com. ‘It’s diverged so far from Sarge that packages built for Ubuntu often don’t work on Sarge. And given the momentum behind Ubuntu, more and more packages are being built like this. The result is a potential compatibility nightmare.'”
Mark, this doesn’t end well. If you want a glimpse of what will happen, take a look at the RPM world, where software developers and ISVs have to build a different RPM for every RPM-based distro (either that, or the ISVs have to choose the one or two most popular RPM-based distros to the exclusion of all others–or perhaps that’s what you have in mind?).
Here’s a suggestion on how we can avert the crisis before it becomes one: Provide a Debian compatibility runtime and development environment for Ubuntu, and make the development environment the default environment. That way, when developers build packages on Ubuntu, they can be installed as-is on Debian as well. Provide a Ubuntu-specific development environment too, so developers can take advantage of Ubuntu-specific features that aren’t in Debian yet, but only use those features when you absolutely must. Everyone wins.
If you’re really interested in joining forces with UserLinux, this would be a great start. I’m sure Bruce would agree. You’d have my interest as well.
Related link: The Debian Delay: Is Sarge MIA? Or Simply a POW of Process?
“Can’t we all just get along?”
yes! but debian is losing the desktop. get over it.
“There is no need to use their demands to drive distributions like debian who are much more than x86 desktop.
And it gets ridiculous to think of ubuntu as a basis of debian — x86 only is a very narrow world…”
You’re absolutely right. Who cares if the software is five years old, if you’re running it on ten-year-old hardware.
Debian should just drop support for x86 and PPC (plenty of distributions have that space covered already) and focus on the Amiga and 68k Macintosh users that are their core constituency.
(And if you don’t think that’s their core constituency, ask yourself why sarge has been held up to fix the production pipeline for these architectures.)
I find Stable pretty useless on servers – just as on the desktop, there is too much truly ancient software in it, software that in some cases is no longer supported upstream.
A while ago I upgraded a server to Sarge to oversome these problems – I was forever backporting Sarge packages to Woody – thinking Debian had in place security updates.
I’ve since discovered Sarge is pretty hopless on servers too: I connect over a VPN (using OpenVPN), and I find the maintainers for OpenVPN and Shorewall regularly make changes incompatible with previous versions and so I run the risk of undesired visits to the sites I administer because of a broken VPN or a broken firewall.
We are not amused.
> I bet the Ubuntu maintainers that actually do work wouldn’t say Debian “sux.”
Considering that the Ubuntu maintainers are (mostly) DDs too, I’d say the major reason they’re with Ubuntu is that they think Debian sux.
> Sarge has been in testing… who else would test an OS for that long before releasing it?
Sarge has been _called_ testing (or vice versa) for a long time, but it’s not yet being tested in any formal sense. That doesn’t happen until package-freeze, when new releases of KDE, Gnome, shorewall etc are banned and only bug-fixes permitted.
Whether that will happen this year or next I’ll not hazard a guess.
> Maintaining compatibility with Debian will at least probably mean that it will work for all these other distros
Which Debian?
Maintaining compatibility with anything older than Sid is insane; even trying to keep back with Sarge prohibits the use of newer packages – Ubuntu had Gnome 2.8 while Debian/Sarge was still pondering 2.6.
Clearly this isn’t the beginning of the end for debian, but maybe the leadership needs to ask themselves why this topic has generated so much strong and vehement discussion both here, on slashdot, and on a number of other forums. There is an old saying: “where there is smome, there is fire.” Something is wrong and needs to be fixed. The signs point to this being both a governance (ie leaders not reflecting the will of the people) and a a structural issue (cf user/desktop vs server debian proposal above).
Why not fix the Debian-developers instead, by making them less arrogant pricks?
Ubuntu just works and does not smell like 1990, something that can’t be said about Debanal.
Maybe a better way to have Debian “everywhere” is to stop trying to release anything after Sarge.
There can be some great distros for specifical targets (from ordinary users to geologists, from musicians to enterprise admins), ALL Debian-based, with software of great quality, an enourmous community, etc.
Just divide Debian branches in “stable packages”, “testing packages” and “experimental packages” (a fusion of present unstable and experimental repositories). Packages, not versions of Debian!
Ian: A compatiblity layer is the LAST THING you want. A compatibility layer means you’re not compatible, and it’s that incompatibility what you want to kill, not workaround it and leave it there.
The One True Fix ™ would be to join debian and ubuntu developers. Debian developers could comaintain packages in ubuntu and viceversa. Ubuntu should be a sort of “debian desktop distro” – with different FSG maybe, it DOES NOT have sense to provide a desktop distro which has a non-free driver out of the installer. Debian and Ubuntu need to agree in that
Looking at the comments: good Lord.
Are Ubuntu users the new Gentoo users?
Ubuntu users *are* the new Gentoo users. OMG-everything-binary (but-don’t-make-me-fight-dependency-hell I-didn’t-have-to-think-on-Windows) is the new OMG-everything-from-scratch (but-not-actually-from-scratch because-that’s-really-hard).
Hooray for fashion.
Now, anyhow, someone above said:
“Sounds like someone doesn’t like loosing control. Where someone see failure another sees opportunity. I predict Ubuntu will become the defacto Debian and Debian will follow Ubuntu’s lead.”
Losing control isn’t a problem, except that Ubuntu itself relies on Debian to provide a lot of *documentation* for the underlying system. Want something organized more authoritatively than an incomplete FAQ or posts on a phpBB? You’ll be heading back to Debian.org to find out what Ubuntu built on.
If the two projects don’t find a way to complement each other, Ubuntu will fall out of sync with the resource that makes their platform seem saner than RPM-land, and Debian will… well, miss out on all the cool stuff Ubuntu is doing, which means even less movement in the famously conservative Debian tree. Recognizing the issue is the first step to doing something about it, there’s no reason to call it a holy war.
Debian -was- a good distribution. The release schedule is not practical in an environment that moves as quickly as the open source world. There are very real differences in kernels and libc that make it important to keep up.
If Debian is interested in staying relevant it might be a good idea to commit to a more regular release schedule and adhere to it religiously. The OpenBSD developers have managed to combine regular releases with rock-solid stability without any problem. That’s why I moved by old Debian systems to OpenBSD.
When I first started using linux (1993) and throughout the early/mid 90’s, I tried EVERY distro that was around, to see which one I wanted to use. All the people I found using it for workstations were using redhat, so I gave it a good go with that. I tried and tried to put up with the frustration of yet another broken rpm from somebody who wrote a piece of software, and packaged it too specifically for their system. After about 6 months, I was left with the decision to either just compile everything from source, or switch distros again. Haven’t used anything but debian since. The package management is the best in the linux industry (let’s be frank, if you’re thinking “but what about gentoo?!” – that’s not a package, it’s a source tarball. But yes, I like emerge, too… It reminds me of what I like about FreeBSD :), and never, not once have I had to deal with a package someone else put out not working for some silly reason.
Yes, debian’s an awesome starting point. Yes, they’re slower on releases than [insert the slowest thing you can think of here]. No, it’s not a good idea to start making debs that don’t work with debian. Yes, it’s snobbery on some level, but at the same time – how would you like to find this great program that does EXACTLY what you need to do, and even better it’s got a deb to install it from(!) – …as long as you’re using Ubuntu, not Debian.
If Ubuntu goes off and makes their own package filename extension (I mean as simple as mv *.deb *.udeb, I’m not talking about making a new package format), then frankly I couldn’t care less…Anyone with the will to care that they’re deb’s, and use them if they want that version of something would already have installed the required packages to allow them to work. Heck, I’m pretty certain I can see a way to even make it apt-able, so it wouldn’t require more than adding an apt-source to install anything. I’m fully in support of the compatibility runtime idea…debian would be debian, and ubuntu would be ubuntu, and if you wanted to cross pollenate, you’d first install the appropriate runtime. I don’t think it’s appropriate for Debian to suggest that Ubuntu use the debian runtime by default, though I don’t think that’s what Ian was talking about. But a compatibility package on both sides of the fence, that’s tied via dependancies/equivs when it’s built to the packages from the other side of the fence. Apt for something, it fetches the compatibility requirement. If you want to make sure you don’t mess things up you need at debian’s level, apt-pin them.
I’m running SuSE (after giving Debian the flick years ago, and when I feel like booting up Linux) and was rather impressed by the last Ubuntu, but it seems to me that using the file extension of .deb for packages is not future-proof. I mean, there are *mdk.rpms and *.rpms, and probably others that I just can’t think of.
Why not use *.pkg as the file extension so that it, at least, appears more portable (even if it isn’t necessarily so). If everybody is used to seeing something as logical as *.pkg for a Linux package, then a) they might begin to expect ALL files with that extension to be portable and get very mad when –insert evil forked distro here– decides to be naughty, and b) perhaps even begin to subconsciously see *.pkg as THE Linux package format, and start to say: “why does this distro use these multitudinous *.rpm files – these *.pkg files seem more cross-platform, I might find a disto that uses these *.pkg packages.”
Of course, I’m rambling, and probably won’t switch to Linux on the desktop until the horrific excuses for GUI file managers are put on a diet, but the change to *.pkg just seems so logical. You’d be silly not to pass up an opportunity to lock in a file extension like this and make it a standard (or else, usurp it if some crappy FOSS project uses it).
Just imagine *.pkg being the new *.zip!
“here are way TOO MANY distros… there needs to be some consolidation”…hey, you’re right! I’m gonna start a new *consolidated* distro.
Joe: “here are way TOO MANY distros… there needs to be some consolidation”…hey, you’re right! I’m gonna start a new *consolidated* distro.”
Man, stop making me laugh so much :-) If only the well-intentioned masses could hold themselves back and stop creating their own distros which inevitably are abandoned after realising the *true* amount of work involved, the world would be a better place!
“I wish debian (and various other projects) could have a “core-release” with apps being more ‘floating’ on top. The core-release would consist of the installer, kernel, main libraries, main system utilities, and compiler environment(s).”
This is the approach that FreeBSD has used with great success for years. The core freezes, packages/ports continue to evolve and stay current. Linux distros need to learn from this.
Why not just change to Windows as a desktop. It Actually works, it’s not that expensive and all fun stuff is there. I have wasted 2 days getting xmms to work in latest Ubuntu… and I’m still mad! There is no packet hassle or anything like this in the Real Desktop Systems.
The reason I like Linux is that it keeps Redmond at work. If there is an (un-)narural enemy like Linux then everybody Wins in the end. Ubuntu is a corporation and it clearly shows that charity is just a charity: it keeps people thinking about the developing third-world systems. Commercially based Linux distros like Ubuntu are a real match for MS in the near future.
Debian coders are being payed for developing Ubuntu. I don’t think Ubuntu is ever going to go too far from it’s roots. It wouldn’t be wise for the system and certainly not for the business.
What ever happens the competition is going to be more even when money is in the picture. Only pirates live free.
When I first switched from RedHat to Debian 6 or 7 years ago, I loved it! I still remember my delight after my first apt-get install downloaded a whole slew of dependancies for me and installed them.
However, in recent years I’ve become more and more disillusioned with debian – I picked debian because it gave me an easy way to get the most up-to-date packages… But more and more often it was taking *years* for new software to make it into even debian unstable. I don’t want to have to wait forever to get the latest version of Gnome, or the X.org X server. From the outside this looks like a sign of a distribution in bad shape, buried in red tape.
I switched to Ubuntu lately and I haven’t looked back.
I currently run 3 Linux distros on my laptop, Ubuntu Hoary, RHEL4 (CentOS), and Debian Sarge. I have messed with a lot of Distros, and these three are here to stay on my laptop for a while (I like to test and tinker). Having said that, yes, I agree that Ubuntu Hoary is quite a bit more polished than Debian Sarge (which I am using as I write this).
I think the Debian developers should concentrate their work. By that I mean, perhaps instead of ‘Stable’, ‘Testing’ and ‘Unstable’, they should only have two branches such as ‘Server’ and ‘Workstation’. On the ‘Server’ side, get things up to speed (so to speak) so that it is roughly equivalent to RHEL4 (ie: 2.6 kernel, Xorg, etc). On the ‘Workstation’ side of things, go ‘bleeding edge’. This, I believe, would put Debian on the top of Distrowatch dot com! :)
Dave, I was thinking the same thing!
Here’s to reading post after post about how everything else in the world sucks.
You know, this really wouldn’t be a problem if the Linux community could get its act straight in regards to package management. There needs to be base set of API’s that package managers use that can be configured on a distro by distro basis… something like “getuserapplocation()” that would return “/usr/bin” on some distro’s, “/usr/local/bin” on others, etc…
Of course this is only part of the problem. You can also build a set of API’s for runlevel management, desktop menu management, etc.. Then all the package management apps can build on those tools to provide truly patform independant package management no matter what distro. The package manager can always find the locations for it’s files and always find the mechanism to configure them properly.
Yeah, I know, it’s a simplistic view of the situation, but it would be a good start.
“This is the approach that FreeBSD has used with great success for years. The core freezes, packages/ports continue to evolve and stay current. Linux distros need to learn from this.”
I have heard this suggestion before. Seems like the obvious solution. Is there some technical reason why Debian doesn’t do this? Certainly seems like it would solve a lot of problems. There’s nothing all that inherently weird about a 3-year release cycle, as long as you can get up-to-date applications.
What you want already exists. It is called “The Linux Registry” or “Registery”.
http://freshmeat.net/projects/reg/
I tried out Debain Sarge on my home desktop. Everything worked great except the reliable unmounting/mounting of any external USB device. So, I patiently waited for Ubuntu Hoary to arrive and within 2 hours, I had my SSH/Samba home file-server set up which also acts as an MP3 player for my stereo unit. I chose Kubuntu in particular. I would recommend any Debian-based distro…which one you choose depends on your needs.
I was using FreeBSD 4.9 when I added a Debian machine to my collection, and I have been very happy with it. Now I have moved the FreeBSD machine over to Debian.
It seems to me that the dividing line is really dial-up vs broadband. Personally, I wouldn’t bother downloading any ISOs, being on broadband. If I were stuck on dial-up, I would want to get ISO’s somewhere, download them at work, at a friend’s house, local computer shop, whatever.
In my mind, ISOs are for dial-up, and netinstalls are for broadband. What bugs me about installing Debian is not having a choice of JFS or XFS or ReiserFS, although I have been able to get around that by moving partitions around – so it’s not so much of a problem – you only have to set up your partitions once after having compiled a new kernel, and you are good to go. I have been using the Gentoo install CD of all things to boot from, mount different partitions, move them around and recreate the filesystems I want. Has worked very well so far. I can install on ext3 and then change things around later.
I have been very happy with Debian, all you have to really do is remember to upgrade your system every now and then. Install what you need, when you need it, keep it up to date often, and look for 3rd party websites with .deb files when you can’t find them on the debian repository itself.
I go out of may way to find netinstalls – FreeBSD has a good one, looks like Dragonfly BSD also is getting better, and Debian is very easy to install and keep upgraded over the net. If I had to deal with dial-up, however, I would probably be taking a completely different approach to things.
I have started migrating my systems to ubuntu. I’m sad to leave debian but there are two issues compelling me to make the move. Both are critical for me. One is amd64 support, and the other is the lack of security updates for sarge. Keeping servers on woody is just not an option for me anymore.
I can live with sarge being a bit behind the cutting edge, but security updates are more important to me than either being on the cutting edge or stability. I would happily trade the nth degree of stability for good security updates.
I really like the philosophy of debian but I left debian a few years ago because the packages were so out-of-date even the ones in unstable. I converted to gentoo because it gave me the flexibility of debian albeit with a downside of recompiling everything on an update.
instead of providing a debian compatibility layer, drop the current unstable on the floor, and make the new unstable the current Ubuntu and work from there.
At the same time, try to understand why the fork happened and try to bring the two groups back into alignment.
I really don’t get the whole Ubuntu thing. It seems trendy and overrated, and I dislike having to tweak it in order to get what I consider a recognizable Linux filesystem, complete with su, etc. As a home Linux user whose background is not technical, I find that at this point, I can get more out of Sarge than I can from any other distro. I realize that may be because of my limitations– but I’m a human being, and I’m always gonna have limitations.
Compatability is a good thing, but is it essential? Since most users of U and D get their software packages from their respective repositories, I don’t see that as nearly as annoying as having to hunt down a different RPM for each distro. I think that Debian and Ubuntu will both remain vital as Linux continues to grow.
damn, the answer is _really_ simple :
Componentized Linux is _really_ close. Is has been said before:
1 ) make a stable core that is released at a convenient pace. Update it in the currecnt Debian stable manner.
2 ) start compiling all the apps starting from that core. This part shuld be called dynamic or something like that. All new updates are based on the core.
3 ) Ubuntu starts from this core.
4 ) All inovations can be brought back to Debian.
This has been said maqny times, but I felt the need to repeat it once more.
Maybe Branden Robinson will read this and he will gather some other influential DDs to support this idea.
Periodic releases were a hot promise in every candiate’s platform.
All Debian developers are aware of the problem, but communication barriers always arise.
Perhaps you’re thinking about this the wrong way? If a developer wants a package to work on Debian & Ubuntu, then they shouldn’t be building on Ubuntu in the first place. They should be building on Debian. Debian IS the base Ubuntu builds upon, and if you want to support both, you should be building on Debian.
So perhaps the question should be: Do packages built for Sarge work on Ubuntu Hoary (their latest stable release)? In most cases, the answer to this is yes. In some cases, there are library issues. And guess what? Most of the time it’s because there is in fact a newer version of a library in Sarge than exists in the stable version of Ubuntu, and in particular an ABI has been broken along the way.
Is this an issue? Yes. And until Sarge actually totally freezes (I think it’s actually happened now) and gets out the door, it’s going to keep happening. But it’s Ubuntu’s problem, NOT Debian’s. And the way they can fix that, is by creating a update repository that actually ports these newer version of such packages to Ubuntu. In most cases, it’s simply taking the Debian package and rebuilding it under Ubuntu.
I’ve not seen many commercial packages that actually ship .deb’s. I know that Win4Lin does, and I see upstream packages that have .deb’s on their web pages, but they’re all built on Debian. Once again, if anything, Ubuntu is the one with the problem here.
The real issue IMO is that Debian is taking way too long to get a stable release out the door. I’m not accusing anyone here, as everyone knows it’s a fact, and arguing over the specifics will only waste time that could be spent on doing stuff to get Sarge out the door. Sure we need to analyse what went wrong, but lets get Sarge out the door first. And Ubuntu is trying, where it can, to help in that regard, by paying the wages of developers, to work on the same projects that many are Debian Developers for.
Enough for now.. Sheesh, I need my own blog to rant about this sort of stuff. *grin*
Debian has brought this on itself. Two major reasons Ubuntu is so popular are that the installer sucks less and releases happen more often than ice ages.
Folks, use what you want to use, and forget about this Debian vs. Ubuntu crap. The whole point was about freedom right? The Purist can mull over every bit in Debian they want, and the Pragmatic can use Ubuntu. There’s a choice! Awesome. Everybody is going to have their reason to use this or that. Forget about having to merge this or that, forget about Microsoft. Use what you want to use, and perfect it the way you see fit. That’s the whole point! Forget about the whining, the complaints, the opinions – do your own thing and to h@ll with the rest.
I tried hoary but switched back to sarge, didn’t see any particular advantages other than the read ahead which make sysinit much faster. However, that is lots of things not as well tested as debian(mostly server packages).
I would say ubuntu is a very good desktop system but I don’t need that. I run XP which is still far superior in terms of hardware support. Under XP, I have colinux with sarge as my “server”. I also have some headless servers which can’t be benefited from ubuntu.
BTW, it seems that security updates of debian sid is almost in sync with woody(which then would flow to sarge) so I wonder why people keep on saying as if sid/sarge don’t have it.
Also many packages(in debian) in ubuntu is under “universe” which is not supported(or have security update).
ubuntu has many “uh” factors to compete with say Windows but it is not as versatile as debian. I can see this in their forum too, 95% is geared towards desktop users(like KNOPPIX) and most of the server oriented questions have no responses, sharp contrast with debian.
I used Debian Sid for about half a year, after I switched to Ubuntu. Debian Sid broke a bit too much for my liking (even as a desktop system). I’m using Ubuntu since the Warty release, upgraded to Hoary, and I can’t see a reason to switch back to Debian.
Please note that I appreciate the efforts of the Debian team, and their fantastic tools (dpkg, apt, …). I remember that there was news that Sarge would be released in September, to which I was pretty excited. But we’re 7 months further now, and Sarge is still not released. This gives the impression that things won’t get started. People contribute a lot to the unstable and testing distributions, but making a release won’t happen.
I can see the problem with Debian and Ubuntu diverging. I for one, would love to keep using Debian’s packages on Ubuntu. But the Debian project should also get it stuff together. Time is running out. Make a release, consider a new release strategy for the future.
Ubuntu is a good distribution, but it also gets a lot of momentum because it is the answer Debian desktop users have been waiting for. People have been frustrated at Debian’s slow releases for a long time. Ubuntu gives them an up-to-date, nice-looking, “just working” desktop system, with Debian’s power under the hood.
We maintain a student’s GNU/Linux server, and were planning to switch to Debian for a while now. We used Slackware, and wanted a system that was more maintainable. We used to say, we’ll wait for the Sarge release. That would make things easier. And we would have a more up-to-date system, since our users we’re asking for software such as Subversion.
But since Sarge was still not released, we eventually decided to use a Ubuntu server install. It provides up-to-date software, which is still stable enough for our use. I must say we are all very content with this switch.
I read about other people using Ubuntu as a server system. If this becomes more common, where will Debian be in a few years? Now we are all confident that Debian still has it’s use as a server system. But if the project still fails to release for some time, the situation could suddenly change. Desktop users are more demanding, but servers users can get frustrated as well.
All in all, I want to express my gratitude to the Debian hackers. They have done a great job (and still are). But it is time to be more concrete now if we want the project to still have a future.
I agree with the basic idea of changing names for Debian stable, testing etc. Enterprise & Desktop are my suggestions as these would focus developers more on what they were aiming for.
And I believe both sides (U & D) could make some changes to help all. Is it so hard for the Ubuntu team to have their packages compatible with Debian?
And for their part, is it so difficult for the Debian developers to release more often? Surely a shorter cycle along with a smaller jump in package versions is possible, even with the number of architectures Debian supports.
And on that, perhaps there’s a way for different groups to handle different architectures, meaning a smaller team for each architecture & one that can move more quickly.
Or maybe a separate team for stable/Enterprise & testing/Desktop, with two separate Debian distro’s that work together when it makes sense & agree on basic stuff in order to remain compatible.
I think it would be very sad if .deb distributions went the way of .rpm distro’s & all the incompatibilities they have.
just a comment on the Ubuntu installer “sucking less”. It is the new Debian installer.
The development of this installer has been one of the principal delays in Sarge. If more people had been willing to give a little time to help by trying the installer and giving the developers feedback, _etch_ would be out by now.
Pingback: eyou.com 开发者
Pingback: CVD's Weblog
Pingback: Lars-Schenk.com » Blog Archiv » Spannungen zwischen Debian und Ubuntu
Pingback: Diamonds, Diamond Rings, Diamond Jewelry » Comment on Can t we all just get along? by Topic Gold rings … (Diamonds)
Pingback: attic » Blog Archive » Kāpēc Debian nav sajūsmā par Ubuntu?
Pingback: Dunkelangst » Und Tschüß Ubuntu