Ian Murdoch, 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.“
Ubuntu harmful for Debian
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
As Ubuntu is based on a sid snaptshot, saying that it diverged to far from Sarge isn’t really a very valid point, is it?
So, someone has built the better mousetrap, and is gaining popularity. Although it’s bad to see that things are compatible between distros, we’ve been dealing with that for years under the RPM (SuSE, Mandrake, Fedora are all reasonably different from what I understand). I guess it would be nice if there was 100% compatability between Ubuntu and Debian packages, but I’m not heart broken. This is Darwanism, either be better than your competition or get out of the way.
my sentiments exactly
I love(d) debian for years, but even I eventually had to switch off of it. Stability is a good thing – being stuck in the very distant past package-wise is a very bad thing.
While Ubuntu is bad for getting sarge released, I don’t think it’s bad for Debian as a whole. None (or very little) of Ubuntu’s work has gone into sarge because it has been using sid. Once sarge gets out the door, all the Ubuntu work will be in etch and sid and we can all be one happy family again. I think it’ll be far easier to get a stable release out when Ubuntu has already had a release based on the same packages. The question is why will people want to use actual Debian releases? Anyone who can bear 18 month upgrade cycles will probably use Ubuntu instead. It’ll be interesting to see what happens.
…their distro is better than mine…
Seriously, this can only be bad for Debian if they keep operating the way they have so far. This can only be good for Debian if the project takes it as the sorely needed kick in the pants that it is.
It seems to me that Debian has become very, very, very stagnant in the last few years. So suddenly a distro comes along and does what debian does better, and is more friendly to the community, and is built on the concept of being totally free for everyone, and the Debian people get pissed? I could understand a bit of jealousy, but Debian hasn’t seen a real release in years, and Ubuntu already has 2 good ones under its belt.
I say kudos to the Ubuntu folks(my new linux distro of choice) and shame on the Debian people for muddying the waters of a great and mighty distro. The Ubuntu people have always been kind and generous to the Debian community, so it would be nice to see Debian folks reciprocate.
Just my thoughts.
As I already said, I really don’t thing Ian Murdock has a valid point here, but bashing debian now, or claiming it has totally stagnated simply isn’t accurate and fair.
After all, Ubuntu is based on debian, wouldn’t exist without debian and heavily uses debian sid, which is anything but stagnating btw.
I recommend reading Ubuntu website regarding Debian and Ubuntu.
http://www.ubuntulinux.org/ubuntu/relationship/document_view
Some highlights:-
– First, Ubuntu contributes patches directly to Debian as bugs are fixed during the Ubuntu release process, not just when the release is actually made. With other debian-style distributions, the source code and patches are made available in a “big bang” at release time, which makes them difficult to integrate into the upstream HEAD
– Second, Ubuntu includes a number of full-time contributors who are also debian developers.
– As Ubuntu prepares for release, we “freeze” a snapshot of debian’s development archive (‘sid’). We start from ‘sid’ in order to give ourselves the freedom to make our own decisions with regard to release management, independent of Debian’s release-in-preparation. This is necessary because our release criteria are very different from Debian’s.
I use Ubuntu out of necessity. I love Debian and I would continue to be a Debian user if they would commit to a regular release schedule. “It’s ready when it’s ready” isn’t really going to work for those that need to maintain current machines and doesn’t bode well for marketing an OS (Longhorn anyone?).
“Use Debian, we won’t release regulary and you may have to wait a couple years for a new update… not to mention the fact that we will tell you that it’s going to be released then push that date further and further back.”
Then there is the old argument “Use testing/unstable… it’s really very stable.” (Then call it something different and support it with security updates)
Anyways, that’s my beef and I really do appreciate Debian but I can find a niche for them at the moment.
I don’t know personaly noone who use stable (or even test) debian. I mix sid/experimental/hoard – it works for me and that’s it.
I cant beleive these people become so entrenched in the money world they start saying stupid things like this.I would just say get a grip. If it was not for the background behind ubuntu i would feel maybe differently. But i just cant project of a money hungry corporation on canonical. What i have seen so far tells me what progeny was supposed to do long ago.
I think what Ian means is that Ubuntu is bad for his Progeny/Componentized Linux. You see lots of projects, Guadalinex and Linex were using Componentized Linux as their base and have now made Ubuntu their base distribution.
I run Debian on servers and would not change it for anything in the world. I would also like to see each Ubuntu release come out once a year and with 5 years of support, but I doubt that they have the infrastructure to support this. If they did, they would clean up the enterprise, home and SMB markets in a short two to three years.
It surprises me that they don’t see the huge market that’s awaiting anyone willing to provide security distributions for a distribution very cheaply for more than eighteen months.
KDE and Gnome are now at the stage that they are as usable as their proprietary counterparts if not more. All the ducks are nicely lined up waiting for Canonical to say we are going to continue to ship CDs and if you want security updates beyond the first 18 months, we will charge $50 a year per machine for the service.
That would rock!
to different os’s for to different groups of people. mandrake and redhat arent even largely compatible any more why should debian an anything that springs off of it be compatible? (note, in actuallity they r very compatible, i hav a mishmash desktop setup- with ubuntu and debian and it works fine)
“The Ubuntu people have always been kind and generous to the Debian community, so it would be nice to see Debian folks reciprocate. “
90% of the work is being done by the Debian devs, and Ubuntu just uses that to create a snapshot of Sid, and release it every six months. So there is absolutely no need for the “Debian folks” to reciprocate – they’ve already given the best gift of all to Ubuntu and other Debian derivatives – the huge, stable base upon which to build one’s distro.
If anything, Ubuntu, and all the other Debian derivatives, owe Debian huge reciprocity.
Anyway, Debian does seriously need to get it’s act together in terms of releases. Debian, as a stand alone distro, is edging towards irrelevance – it’s the derivatives that are getting the attention.
“After all, Ubuntu is based on debian, wouldn’t exist without debian and heavily uses debian sid, which is anything but stagnating btw.”
Yes, but the thing people, myself included, are complaining about is not Sid, but rather Woody. If debian issued security patches with the same priority for all 3 releases, then Sid might be a viable option for more people. As for me, I found it to be buggy (well, duh), and breaky (again, duh). Ubuntu is far from perfect – there are many small easily, and not-so-easily overcome problems with it still, but it is a move in the right direction IMO. Libranet and Mepis are both fine distros based on debian for people who don’t want to do the Ubuntu thing.
Debian, while pioneer, needs to be moved past. Finally, we have a fast development model that also has debian’s strengths.
…open source is all about competition. Spreading fud about a popular distro because its based on yours isn’t really a reasonable way to behave.
foreach(Toy t in Pram)
Pram.ThrowOut(t);
Ubuntu is doing really well based on the hard work of its own team. Discrediting it in such a way is just silly – learn from their strengths and incorporate things back into the Debian tree as necessary.
I like the product. I think its popularity helps Linux. Anything that helps Linux helps free software. Debian has no rights of entitlement to Linux – even though some people believe that.
Sarge was delayed long before Ubuntu was launched — in fact I think it’s fair to say that Ubuntu was created in part because of Debian’s poor track record with creating stable releases. Really, if the Debian project could commit to a predictable release schedule I doubt that Ubuntu would have been as immediately popular as it was. And if Debian itself does embrace predictable or at least regular releases I’d expect that there would be some bleeding from Ubuntu back into Debian.
As it stands now though, you can hardly blame Ubuntu for responding to a problem that it did not create. Get sarge out the door already — and if you can’t, try to address the fundamental problems with doing so instead of blaming competitors for responding to your problems.
<k>Murdoch argues that if Ubuntu were truly compatible with Debian, all of the energy going into it could be directed at Sarge and toward getting it released, which is what would really benefit the Debian developer ecosystem as a whole.[/i]
Why don’t they just fork from Ubuntu? </sarcasm>
Seriously though, Ubuntu is contributing back upstream, and the do a lot. If Debian is making no progress, that’s their own fault, and definitely not Ubuntu’s. Maybe they could concentrate more on getting Sarge out if they stopped whining and bitching about Ubuntu. Sarge is the free (as in speech) Longhorn – they never release it.
That’s the world of free software: If something better comes along, people will choose the better.
He wouldn’t need to worry about this if Linux incorporated a universal installer. But oh no, we can’t have that is what all Linux developers are saying. So suffer!
ubuntu does what debian does? Someone needs to drop the crack pipe. ubuntu is a dynamic, small, desktop focused distribution with limited portability. debian is a slow-moving, gigantic general-purpose distribution which runs on toasters. get a grip.
Update Debian more quickly then!! That is one of the reasons people are going for Ubuntu. Ubuntu is all the happiness of Debian but quicker.
you would hope they would be beyond this by now.
#1) ubuntu is not alone as a debian derivitive.
#2) ubuntu fixes are rolled back into sid, making the sid trunk more secure
none of the other debian derivitives have harmed debian in the least. ubuntu will never supplant debian if it builds on a new sarge every six months. ubuntu will alwas benefit debian by making sarge more stable.
the only way you could say ubuntu is harmful to debian is if 3rd party debian repos start dropping debian support and only going ubuntu, or if debian contributers start jumping ship. i dont see either happening, and the “harm” this would cause debian would really be minimal, compared to the huge amount of people who will be using sarge based binaries.
90% of the work is being done by the Debian devs, and Ubuntu just uses that to create a snapshot of Sid, and release it every six months. So there is absolutely no need for the “Debian folks” to reciprocate – they’ve already given the best gift of all to Ubuntu and other Debian derivatives – the huge, stable base upon which to build one’s distro.
There’s nothing going actually.Debian is continuing to have it’s strong silent momentum.I think it’s good for them to stay calm and keep everything at the same high standards as the are now,there’s nothing more damaging as a bad reputation from misbehaving or crashing apps .
Where ever you go,go..,no serious,a lot of professional applications for linux mostly come with installers for RHEL,SuSE and Debian,especially server based ones.I think most server admins won’t be impressed by those fast release cycles.Release something when it’s realy ready and audited.
Ubuntu is a nice and refreshing distro though,the community seems friendly also.
Well Debian people, then fix your product. Releasing sarge is good start. After that, plan (yes as in schedule) an annual stable release.
Ubuntu is nice, since they focussed on a small target (desktop users), it’s quite polished. BUT. Debian is more than that. Debian is the jack of all trades and will run on every piece of hardware you might think of.
Ubuntu is not a replacement for Debian, and never will be, unless you are the target market.
I installed Kubuntu 5.04 last night.. It was good, but aside from detecting my wireless card (which still didn’t work, but it’s a start), it offers me nothing over standard Debian unstable. My Debian install does everything that Ubuntu does, without the hype, and with more flexibility. Ubuntu is great, but Debian has been great for years, so don’t act like this is a revolution.
Ian is not whining about Ubuntu, he is merely saying that introducing incompatibilities (yes there are some) is the wrong way to go. This is undisputably true.
I’m always amazed at how readily people will kick their first gift horse in the balls as soon as they get a new, slightly shinier one. (Turns out the joke’s on them because the new horse is just the old one dressed up nicely).
Debian runs on my toasters just fine!
I’ve been using Debian (testing) at home and work for some time now and have had no problems with it.
Ubuntu is a happy eye candy desktop (in a good way) — and from what I gather, most people use Debian for stability and compatibility and not for bleeding edge eye candy desktops.
Sure, a few people may migrate, but I don’t think Ubuntu’s poularity threatens Debian’s core audience.
Debian is dead, frozen, halted. Ubuntu is alive and well and moving forward, so of course it is going to be incompatable. if Debian would get off it its butt and do something you would find that it could easily take advantage of ubuntu’s momentum and remain compatable, but the debian folks would rather whine instead of working on something…
David
Heh, heh! Ubuntu’s popularity is now delaying Sarge’s release? Either Murdoch is smoking too much pot or his Debian developer chums are running out of excuses. Next time it will be UFOs seen over Texas that delay Sarge’s release.
Just get the friggin’ next stable Debian release out, so we can all move on to more important things.
Puh-lease. I’m using Debian testing and I’m quite satisfied with it. It has relatively new packages. It takes a few weeks after a new version of a software is released to reach testing. So what?
Just because Debian has a stilted version (‘stable’) doesn’t mean ‘testing’ is stilted, or even unstable. It’s just as current as all the other modern Linux distro’s out there. The main difference is Debian/testing has tons more packages built for it than most others. And it’s still one of the easiest distros to upgrade between versions.
ps. So a few packages are old (eg. Xfree86). Big deal. Get over it.
They just need to stop making sure they release the same version for 11 different archs.
Just focus on the main three,Intel,AMD,PPC, and do the rest later.
I think this is part of the problem with their slow releases.
Just get the friggin’ next stable Debian release out, so we can all move on to more important things.
Sarge should be released when they think it’s ready.Or are all servers going to melt because of to much delay of Sarge?Why upgrading when things run just fine?
Ian is not whining about Ubuntu, he is merely saying that introducing incompatibilities (yes there are some) is the wrong way to go. This is undisputably true.
What would be the point of being 100% compatible with the parent distro? Since it takes forever for Debian to make a proper release, a major modification would probably take the same time before getting accepted.
I prefer the flexibility of Debian, but it’s getting more and more irrelevant everyday…
…and fix your distro. Simple as that.
“I prefer the flexibility of Debian, but it’s getting more and more irrelevant everyday…”
I don’t see why. So we have to wait a little longer for Sarge, so what? It’s up to date and I definitely have a need for a stable Sarge so I don’t have to upgrade for a few years.
Personally I hate having to be a sysadmin on a daily basis when apps change their configurations. Sarge will be on my servers and my desktop for quit a while.
Bleeding edge is nice if you have time to spare. I prefer to actually do some work. Debian’s APT saves me lots of time (no compiling, clear documentation) and it’s stable version saves me even more time.
I agree with Murdoch, because is true! Ububtu developers are not interested in develop debian community, but only in your own, popularity and so on (to what?)…this is not a spirit of the linux…linux is not about pride.If they are interested in develop why not join with debian community ?
“If anything, Ubuntu’s popularity is a net negative for Debian,” Murdoch says. “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.”
How could they-
A. Sarge isn’t released
B. Ubuntu is based on Sid.
“I understand what the Ubuntu folks are trying to do, and they’re doing lots of good work that will eventually find its way into Debian,” Murdoch said. “But what we really need right now as a community is for Sarge to be released.”
There. He admits that Ubuntu is now more about helping out Etch (the release after Sarge) then helping out Sarge. But Warty should have helped Sarge a bunch, and Sarge has problems even millionaire Mark can’t fix quickly.
“In that respect, Ubuntu’s popularity is more harmful than helpful.”
How is it harmful that Etch is going to kick ass because of Ubuntu’s work?
I’ll tell you how- each Ubuntu release is an embaressment to the Debian people. Two Ubuntus have been released before a Sarge. And if they don’t watch out, it will be three. Businesses don’t like that many upgrades usually, so a slow Sarge is good many say. But from the words of of Debian’s founder is obvious that Sarge not being released it is turning into a bit of a joke…not good for Debian’s image.
Thats the only way Ubuntu hurts Debian.
…Ian’s comment. Debian has been designed with derivative distributions in mind and it has the infrastructure in place to make it easy to distribute specialized distros based on Debian which retain a high degree of compatibility with the Debian trunk as well as with other independent distro branches.
If everything were working according to plan, Debian derivatives should ideally be branched off of testing, unlike Ubuntu, who branch their releases off of unstable. So that is how it is supposed to work, but right now Debian is broken, and by Debian I don’t mean the OS but the organization.
Testing is a mess and no amount of elegent engineering is going to fix the problem, because it’s a organization problem. Policies need to change, not code. I have a few suggestions, but they aren’t new and they’ve been roundly rejected many times in the past as not being the “Debian way.” IMHO, the Debian way is no longer working and the organization has to confront that fact. Anyhow, my suggestions are:
1. Two different stable branches: A server distro and a workstation distro have different requirements and it’s foolish to continue servering both masters with a single distro. Debian should release both a server oriented stable branch and a workstation oriented stable branch from the testing pool.
2. Core architectures vs. secondary achitectures: Only x86 variants and PPC should be considered core supported architectures. This would mean that releases will depend only on getting everything working for core architectures. It’s great that you can run Debian on your toaster, but the fact that the new version of X won’t compile on your toaster shouldn’t interfere with the other 99% of Debian users. All other achitectures should remain supported as second class citizens.
3. Core packages vs. periphereal packages: As with architectures, it’s great that the Debian repostitories have every piece of F/OSS software under the sun, but that doesn’t change the fact that there a core group of applications which cover the needs of the overwhelming majority of Debian users (and they aren’t necessarily the same packages when talking about servers and workstations).
Murdock’s CL idea is a really good one. If you can carve Debian into smaller, more manageable chunks that someone could hope to test and integrate, you’ve got a winner.
Debian has too many packages, and too many maintainers who can make a package but not support it. And maintainers who can support a package are sometimes too far out of the loop of the development of the program being packaged.
Ubuntu doesn’t seem to be doing anything to address this; it seems to be making things worse. Short term, it’s snazzy, but long term it just makes a mess.
With due respect to Ian Murdoch (founder of Debian), if Ubuntu didn’t exist, the energy that went into it would not go into Sarge, it would go into Fedora (my former desktop), Gentoo, and other distributions that have rich communities.
Like it or not, there was/is a need for Ubuntu, just as there was a need for Fedora. (I know that if Ubuntu didn’t exists, I’d probably still be on Fedora or moved to Gentoo.) RedHat eventually learned to integrate Fedora (which originally was an unofficial third party repository that sometimes went out of sync with Red Hat) with Red Hat Enterprise Linux, and both have benefitted greatly. Ubuntu makes Debian stronger.
From what I understand, contributions from Ubuntu are synced back into Debian SID every release, so Ubuntu is essentially helping to stabilize SID. Anything that stabilizes SID will ultimately stabilize Sarge.
IMO, Debian will eventually move towards a predictable time-based release structure (perhaps once every 2 years) since it’s been proven to work for may other successful projects (Ubuntu, Fedora, GNOME, ….). This will ensure that Debian Stable never gets too stale or out of sync with Ubuntu.
A structure like the following would work out:
Experimental/Grumpy (bleeding), Ubuntu (desktop stable), SID (.5 year lag), Testing (1 year lag), Stable (1.5 year lag), Frozen (2 year lag)
Co-incidentally, RedHat follows a similar structure with RHEL:
Rawhide (bleeding), Fedora, Fedora maintenance (.5 year lag), Fedora Legacy (1+ year lag), Red Hat Enterprise Linux (approx 2 year lag).
and it allows RedHat to support a few previous releases quite comfortably.
Debian should not complain because Ubuntu is more popular. Debian should rather pull up their socks and try to keep up. Maybe Ubuntu need to come along to wake Debian up. Competition is always good even if the players are on the same side.
Hoary is doing thing its own way (because of the prolonged freeze status of Debian)
– Gnome 2.10: rolled their own packages
– X.org: same
– KDE 3.4: from semi-official (experimental) repositories
– XFCE: from experimental (not os-works.com)
– APT: experimental
I don’t know if part of this work will ever go back to Debian I hope so, it would speed up things for Debian)
Yes, there are a handful of custom rolled packages and backports from experimental included in Hoary, but that doesn’t change the fact the Hoary is most definitely a brach off of sid.
Ububtu developers are not interested in develop debian community, but only in your own, popularity and so on (to what?).
Bullshit…
http://www.ubuntulinux.org/ubuntu/relationship/document_view
Maybe I am not actually Ubuntu’s target group (people who would be happily running Mandrake or Fedora Core). Because there are 2 things that I hate about it:
– I had to report 4 times more bugs in 3 months running Ubuntu than 2 years with Debian testing
– Packages in universe have very poor support/testing (except selected few).And some of them are 50% of the reason I run Linux.
– I want the latest and greatest of EVERYTHING in Debian, every time. Not just Gnome.
– Ubuntu fanboism is very annoying and totally ignorant (did you know Ubuntu invented that “apt-get” thing, and Ubuntu makes a very useful program called “Gnome 2.10”, where would the world be without Ubuntu)
/me likes Debian more than Ubuntu. Debian developers, this is your “wake up” call.
This handful of custom packages is one important reason why people run Ubuntu insted of Debian (other than hype). Gnome 2.10? X.org?
It does have “ls” and “gcc” from Sid, though.
Will those packages (especially x.org ones) end up in Debian? Don’t know, but somehow I doubt it.
For Ubuntu to hang around forever waiting for Debian to finally release something? Woody has been out for what, three years or more now, and they’ve been making noises about Sarge the whole time.
Full kudos to Ubuntu for building an up-to-date distribution on Debian. I’ve heard little but good things about it; a couple of my friends have tried it and are loving it.
This article sounds a bit like sour grapes to me; Ubuntu’s doing better than it’s parent project. I think Ian could calm down a bit and accept Ubuntu’s success, and possibly take it as a bit of a warning that Debian isn’t providing what everyone wants. Which is fine, so long as they don’t expect everyone to stick with it!
I had to report 4 times more bugs in 3 months running Ubuntu than 2 years with Debian testing
Big surprise in case you were running Hoary…
Debian has been left behind. Sarge was scheduled to be out last fall. We’re about 2 months from the 3-year mark for Woody. Even when Sarge comes out, it is going to feel old compared to the other distributions out there. Of course, I will continue to use Debian for servers, but in terms of the desktop Ubuntu is way outpacing it and I don’t think Debian should try to catch up. I mean, a distribution that changes as quickly as Ubuntu isn’t going to have the stability of Debian for the server.
Ubuntu has become the Fedora to Debian’s RHEL.
As for the packaging problem, Debian has always been at a loss for non-repository packages. Packages are usually made for RedHat (now Fedora) and maybe a couple other RPM distros. Even Gaim has never done .deb packages and they are a pretty big project. The fact is that Debian users get their software from repositories.
Well, hopefully autopackage will make all this go away anyway.
I too use Ubuntu (actually Kubuntu) becuase of those handful of custom/backported packages. I didn’t mean any offense, I only wanted to clear up any misunderstandings.
IIRC there is an X.org package in experimental, which will eventually make it into sid–once the logjam around testing clears that is. The updated APT which Ubuntu uses is, I believe, the one from Progeny, which is also in experimental. And as for updated Gnome and KDE, they will move into sid once the logjam around testing clears.
This is really being taken out of context. The original discussion was regarding UserLinux being nothing more than another distribution based on Debian <a href=”http://www.advogato.org/person/mjg59/“>link. The criticism was that UserLinux wasn’t contributing anything back while claiming that it was Debian’s fault that they had failed to release.
Ian Certainly has a different problem in that Ubunto is offering many of the advantages of Progeny and he is right that Ubunto becoming popular is a major problem for cooperation between Debian distributions. But the context here was a Debian response to claims by user linux.
>I had to report 4 times more bugs in 3 months running Ubuntu than 2 years with Debian testing
>Big surprise in case you were running Hoary…
I wasn’t, that’s why I mention it. You know what the worse kind of bug is? The bug that is caused by a fix of another bug (that didn’t affect me in the first place). I am thinking of certain kernel patches (computer can’t poweroff, no DMA for DVD by default – vanilla worked great)
To my knowledge there’s a bug somewhere in Hoary final what crashes “Hardware Monitor” – an applet, compiled from source. The same applet works fine in Warty and Arch/Gnome 2.10. It’s definitely Hoary’s Gnome fault – but they won’t touch it, because hard. mon. “is not supported”
This is Mandrake I tell you, NOT Debian.
I think the problem with Debian is that it is too ambitious. How else can we explain that Ubuntu has catered to an entire group of people (ie: the “too popular” claim) in such a short time? Its been more than TWO YEARS since Debian actually *released* anything.
Ubuntu really does have a different aim than Debian and I don’t see why Ian is saying what he is saying (unless its because he runs a company selling a competing distribution also based on Debian — but I doubt that could be true). Is he worried that Ubuntu users won’t return to Debian? I’d guess that most Ubuntu users won’t — but they probably didn’t use Debian full-time prior to Ubuntu anyhow. Again: Debian does *not* hit the same points that Ubuntu does, period.
So if Ubuntu is popular it is really a sign for Debian to wonder what they are doing wrong. Perhaps if they abopted a policy similar to Ubuntu: instead of releasing 10 million packages as part of the distro, re-engineer the repos so that we have “base” [ie: GNU/Linux/(Gnome|KDE)] with the standard stable/testing/restricted and then additional repos that hold applications (again stable/testing/restricted). With proper meta-packages in “base” packages in “apps” can be marked with versioned dependencies so things “just work”. It would mean a much shorter release cycle and it would make a lot of the Debian based distros redundant.
Ubuntu is showing the way to Debian’s salvation. The “here’s everything in the world” mentality is really not very practical anymore and moreso, offers little benefit to anyone. There are plenty of packages that cater to very narrow groups and those packages should be seen for what they are and packaged together in their own repositories. The way it is setup now, each interested “group” pollutes every other interested group with baggage. That’s not helpful.
Brandon’s tenure will be critical — if he takes appropriate steps now, he can turn that ship around. Otherwise, Debian will either self-implode or will eventually lose enough mind-share that it becomes a really small (ie. base-size) distro anyhow.
Actually I agree with what you say. Debian’s huge bureaucracy is choking it and this is a problem to be finally acknowledged by its developers.
> Is he worried that Ubuntu users won’t return to Debian? I’d guess that most Ubuntu users won’t — but they probably didn’t use Debian full-time prior to Ubuntu anyhow. Again: Debian does *not* hit the same points that Ubuntu does, period.
Yet you can read enthusiastic comments such as “Debian’s dead!” by people who probably have never and would never use Debian. “But I use Ubuntu, which is sort of Debian, the difference is that I can handle Ubuntu”. Debian != Ubuntu
I wonder why ubuntu has gotten so popular? I use Mepis a drivative of Debian. It’s very hardware aware and works great, been around longer too.
I never have tried Ubuntu, not sure why. Something about it I can’t put my finger on, maybe that I heard it was back by a mysterious corportion.
Anyway Debian may suffer from this and ten again maybe not.
I don’t see why. So we have to wait a little longer for Sarge, so what? It’s up to date and I definitely have a need for a stable Sarge so I don’t have to upgrade for a few years.
I guess we don’t share the same definition of “up to date”…
Personally I hate having to be a sysadmin on a daily basis when apps change their configurations. Sarge will be on my servers and my desktop for quit a while.
For server, I wholefully agree with you. For desktop usage, I guess we have different needs. Still, judging by the current delay, we might be dead before Sarge is ever released…
“Update Debian more quickly then!! That is one of the reasons people are going for Ubuntu. Ubuntu is all the happiness of Debian but quicker.”
Debian updates every day. People are using Ubuntu either because they are lazy or because of the hype. For a true Debian user like me Ubuntu is a straitjacket.
“#1) ubuntu is not alone as a debian derivitive. ”
It is the only and first fork, though
Maybe Debian should have seen it coming when all the other distro’s were focusing on a more speedy evolution towards the server/desktop environments. While Debian stayed with older versions and the rest of the distros expanded to more promissing versions of the apps, people lost interest in Debian.
Debian has it’s own goals, namely servers and such, thus it shouldn’t event try to compete on a desktop. If something does everything, than it doeas nothing ‘well’. Slackware went KDE only – good choice considering that GNOME was rather poor on it anyway. Ubuntu is GNOME only – and it should be (i am aware of Kubuntu) – better do one thing well, than more poorly. The diversification is the key – there are so many distros out there, yet they can not aim for the same users – e.g. desktop.
I wish the Debian team very best – they must continue their excellent work!
But they should consider constraining time releases to about 18-24 month, no longer. AFAIK they already have decided to drop development on “unpopular” architecture – good decision.
…They could take a snapshot of Sid, patch the bugs and make their own distro. NOT modify packages so that they are incompatible with Debian, as even they admit.
Ubuntu is most certainly not a fork. At least not by any definition of that term I’ve ever heard. It’s a point release branched off of sid.
From Webopedia:
Definition of “fork”
“To split source code into different development directions. Forking leads to the development of different versions of a program. Forking often occurs when the development of a piece of open source code has reached an impasse. The project is forked so that the code can be developed independently in different ways with different results. ”
Which is exactly what Ubuntu is doing
Rubbish. All changes are made available to the upstream, and future releases will continue to be made from freezing sid rather than being managed outside of the common infrastructure. Debian would be foolish to not take the Cannonical packages for X.org, Gnome 2.10 and KDE 3.4 and use them as the base for experimental. All that would be required is changing a few file names and they’d be fully conformant with Deb packaging policies.
Yes of course Debian Woody is stale for the desktop.
My reaction? I am using Kanotix 2005-2, which 100/% Debian compatible, mostly Sid, KDE 3.4 from experimental.
With Kanotix I have ever used pinning, giving testing a higher priority. Never broke a thing and nothing I can’t have from the huge Debian repository. Every novice can tailor Kanotix to their needs, while Ubuntu/Kubuntu feel completely “locked”, on purpose.
“…They could take a snapshot of Sid, patch the bugs and make their own distro. NOT modify packages so that they are incompatible with Debian, as even they admit.”
But Sid isn’t “Debian” — Woody is. That’s why all these derivative distros exist in the first place. If you are running Sid, you aren’t “compatible” with Debian either. Getting Sid released as stable is an important step that must happen soon. Yet that won’t cure Debian. How long till the next stable release after that? And what of the intervening period?
Like others, I feel that it is high time to rethink what Debian is. Perhaps Debian ought outgrow the monolithic mind-set and instead learn how to become an umbrella for a distributed model. Progeny has the right idea — delineate the entire distro into components depended on a small core. It makes everyone’s life so much more sane. Surely Ubuntu and the other Debian distros would be happy to adopt a stable, up-to-date core? They would then be free to focus on what’s interesting to them, working with Gnome, co-ordinating desktop standards with other distros, etc. Likewise for other interested parties working on other components. If something like that had already existed, there probably never would have been Mepisis’ and Ubuntu’s and Knoppix’s and the like. Instead, there would have been Debian/Gnome and Debian/KDE (or maybe Debian/Desktop) groups and others like Debian/Server, Debian/Netserver.
Then again, maybe not. Even those few groups I mention can have widely differing (and probably conflicting) needs for a core component. One-size-doesn’t-fit-all so why is Debian still stuck in a postion where it strives for that? The absence of a stable, current Debian has lead downstream distros to take on too much. For example, whereas a Debian/Gnome is more likely to stay true to Gnome as it would have a smaller mandate than a full distro allowing it to work more closely with the Gnome community. Instead we see distros like Ubuntu (which is really great, btw) changing basic functionality of Gnome 2.10 even though 2.10 is itself newly released. Becuase they have a “full distro” mentality, they take on development issues that should be left to the upstream provider. Because they are used to long upstream release cycles from Debian, they carry this idea over to all packages. In an effort to improve cohesion they diverge from the base. Which isn’t entirely bad except that they are not the only divergent point. All of the downstream distros are diverging even from each other. Everyone is forking in a way. I don’t mind if a distro makes changes to integrate components: but that’s where it should end, at the integration phase. Debian’s monolithic release model has broken down in the face of a quantity of riches is leading to a fracture of the entire distro market. It is fine to scold distros for breaking with the base but we must remember that Debian’s organizational weaknesses are the root of this issue.
Debian is an extremely important distro. I very much hope it finds a way to both achieve its goals and satisfy users en masse (all users — not just those already using it).
“But Sid isn’t “Debian” — Woody is.”
First time that I hear that.
“That’s why all these derivative distros exist in the first place.”
Most derivatives are either Sid or, less often, Sarge.
“If you are running Sid, you aren’t “compatible” with Debian either.”
Assuming that Sid isn’t Debian. Indeed, downgrading to Woody is pretty difficult
But who wants that?
Other than that I find yours a good post.
Ubuntu could be considered a fork in some definitions, but that isn’t a bad thing. According to wikipedia, even the Linux kernel forks and we benefit from it.
Who cares if Ubuntu forks Debian? The license says it can, and Ubuntu gives back patches and bug fixes.
Am I a fork of my father? Is that a bad thing?
I think the only way for Debian to go is to adopt a faster release schedule for most of its software (perhaps not the crucial parts of the system and the kernel). This way, Ubuntu will work nicely with Debian. In fact, I think it could be an extension of the Debian Desktop project.
As for Ian, I believe he only believes that Ubuntu is bad for Debian at the moment.
And what is exactly wrong with that? Honestly. A duplication of efforts isn’t always good… but I believe it’s for a good cause, though. Like I said, I prefer the flexibility of Debian (although a custom install of Ubuntu can do the job, somewhat) but it sure needs a swift kick in the ass. As both projects are open-source, there is always hope that Debian will adopt some of the good things from Debian.
…from Ubuntu, of course.
“Am I a fork of my father? Is that a bad thing?”
The bad thing about forks, in linux, is incompatibility. There is already plenty of it.
The example again is the rpm based distros: SUSE, together with Debian, is my favourite. But while till now a deb was always a deb, when you are looking for a rpm you must know: “for which distro”? If I am lucky the occasional Fedora or Mandrake rpm might work on SUSE, but I can’t add an entire Fedora repository to my SUSE APT.
Exactly the same thing might happen to Debian and Ubuntu: in future when looking for a deb we’ll have to know which sort of deb.
Now considering that Debian has so many derivatives which are fully compatible, I can’t foresee much good for them.
The popularity of Ubuntu is the clearest sign that Debian’s organisation is completely failing. If it wasn’t for Ubuntu, I would’ve switched to a non-debian distribution some months ago. Woody is so outdated it’s laughable. For a usable system you need to compile half of the applications by hand or use third party repositories with all the risks of breakage and security flaws. You might as well use a more modern distro right away. With Ubuntu, at least I’m getting something usable that’s debian-based.
Also, Ubuntu gets attention because it actually brings in the innovative new stuff for desktop use. A debian release is usually “ah, nice, the same things other distro’s brought us 2 years ago”. For something that takes this much time, the results really should be more impressive. For example, the new installer is nice, but really not much better than those of other modern distro’s. It’s all decent, but not exceptional.
As said in other posts, a time-based release structure is really neccesary. If Debian doesn’t solve their organisational problems, they will become irrelevant as a stand-alone distro very soon (if they are not already. I say that as someone who likes Debian a lot and has used it almost exclusively since Hamm. If things continue like they are now, Debian might become a glorified “Linux Standard Base”, a foundation for building other distro’s. Unfortunately, this will mean death for Debian because it will lose large part of the active community around it. Users and developers will focus their energy directly on improving the ‘end user’ distro instead of having to deal with the huge bureaucracy that the Debian project has become. That way, the project itsself will seriously run the risk of becoming irrelevant in the long term.
As for debian actvity: for the public, development doesn’t count. It’s like saying “We never released our application, but we have a lot of commits and you can check out our cvs!” The developers might remain active in their own circle but in the eye of the end-user a lack of releases means the project is dead. They are building a cathedral that is never going to be finished. The developers might churn out a highly outdated release every 3 or 4 years but nobody will care anymore. Testing or Unstable are nice to toy around with but they don’t have production status with regular security updates etc. and they break when you upgrade them (don’t tell me they don’t – i’ve run both testing and unstable for years – they do break when upgrading sometimes) You might as well use Fedora or Gentoo. Debian really needs to get their act together and _release_ at least once a year or I forsee a very uncertain future for them.
To be honest, i’m not very confident they will be able to pull it off. I remember posting a similar rant on slashdot _years_ ago when Potato was released (thats almost 5 years ago!). They had the same problems back then, and it seems it only got worse.
Is the only reason I’m using Debian today.
I’m sorry Ian, I just can’t agree with your assessment. Build a better more up to date distribution.
You are falling into the same FUDmachine trap that you used to speak against.
Ubuntu just solved Debian’s “problem”. I’ve been following for 3-4 years now and the consenses seems to still be the same. Debian really hasn’t kept up… the Sid jokes are true…it’s a 2.5 year old “beta”… Sarge is pushing a year as “Alpha” but it’s supposed to be “the future”.
Somebody has to draw a line in the sand and say it’s “done”.. Debian has consistantly refused to do this.. they still call stable & supported 3 years old! That leaves out 90% of the really interesting stuff that you read about. Ubuntu would seem to fill the void nicely. If Debian want to keep their 18 month cycles, they should get used to other people taking up the slack… most of the debian “fanboys” will tell you as much… they don’t want to “babysit” either.
My opinion is that Debian should welcome Ubuntu… they’re holding to the same values after all, and let ubuntu “carry the masses”… after all, debain really doesn’t want them anyway.
I have to side in favor of debian in favor of ubuntu for several key reasons, most notably:
1). debian actually works on my system. Ubuntu causes incurable insta-freezes with x on bootup (bad module selection I guess). I’m not going to get into details, since I don’t want to dig through various fanboy responses telling me how great ubuntu is and why I’m a moron for ubuntu not working correctly (also, yes I have tried starting up with -s to see what’s going on — not worth fixing imo).
2). ubuntu setup creates a single super user account. What the hell is this?! This is an incredible step backwards! This is the worst thing I’ve seen a distro do in ages
3). Ok, I just really hate #2. I really wish they’d rethink this!
that was, um, the point I was making.
I agree totally
maybe the debian project wants to replace the Linux kernel with FreeBSD’s kernel. when they released developmental copy of Debian GNU/kfreebsd, instead of putting that needed time into sarge.
maybe the debian project wants to replace the Linux kernel with FreeBSD’s kernel. when they released developmental copy of Debian GNU/kfreebsd, instead of putting that needed time into sarge.
Maybe a not so bad idea.It would speed up the current work-flow without the need to drop some arches and last but not least maintain the current stability,that is if you trust the FreeBSD core team,(i do).Not original,but why reinventing a different wheel everytime when ther’re proven solutiuons at hand,MacOsX.
“none of the other debian derivitives have harmed debian in the least. ubuntu will never supplant debian if it builds on a new sarge every six months. ubuntu will alwas benefit debian by making sarge more stable.”
I presuem you mean Sid, not Sarge as Ubuntu derives from Sid. No point in using any code from Sarge as it is already ancient (even yet not released!).
I tend to agree with many others, Ubuntu rise is due to Ubuntu being what Debian should be.
“I have to side in favor of debian in favor of ubuntu for several key reasons, most notably:
1). debian actually works on my system.”
Fair enough.
“Ubuntu causes incurable insta-freezes with x on bootup (bad module selection I guess). I’m not going to get into details, since I don’t want to dig through various fanboy responses telling me how great ubuntu is and why I’m a moron for ubuntu not working correctly (also, yes I have tried starting up with -s to see what’s going on — not worth fixing imo).”
a) Browsed the Ubuntu Wiki?
b) Asked about in the excellent adn FRIENDLY forums?
c) Tried submitting a bug report? I did and voila, X.org works on my Thinkpad X30 these days.
“2). ubuntu setup creates a single super user account. What the hell is this?! This is an incredible step backwards! This is the worst thing I’ve seen a distro do in ages”
Hmmm…superuser…such as root. Yeah, a really bad idea.
Seriously, IF you want to have a root account you can pick the more advanced installation option when booting from the install cd, but I guess you never bothered to try that.
“3). Ok, I just really hate #2. I really wish they’d rethink this!”
Several reasons became two and both are moot.
ubuntu setup creates a single super user account.
No, it doesn’t. In fact, it doesn’t even enable the super user account. Ubuntu uses the sudo machanism to grant certain users root priviledges. That’s totally not the same as creating a single super user account.
One thing that needs to be remembered is that Debian is not your ordinary distro that fights for popularity among desktop distros. Rather, Debian is developed and maintained by a large international community of volunteers who each have their own interests. Some like to work supporting less popular architectures, some like to experiment on making Debian work on BSD or Hurd kernels. And, of course, many developers try to make Debian GNU/Linux function just like your ordinary distros.
That Debian is such a scattered group of different projects of various interests is both its strength and its weakness. No millionaire can ever buy Debian or tell it what to do. Debian does things its own way, the Debian way. Debian is slow to move but Debian won’t die when the money runs out. And most Debian developers couldn’t care less about popularity contests.
But I’d still insist that Debian should now make some quick and tough decisions to force a new stable release out as fast as possible. If only to ensure that Debian will stay that last shread of true software freedom in this hard, calculating and commercialist rat race world we’re currently living in.
Debian itself is a combination of many projects and many, many other projects are based on Debian. I welcome all these projects. It doesn’t matter if you use Ubuntu or Mepis or Knoppix or Damn Small Linux or Kanotix or Libranet or whatever. As long as you’re using a Debian based distro, you’re also a Debian user. There’s one swirl to rule them all… And that’s Debian.
“Debian itself is a combination of many projects and many, many other projects are based on Debian. I welcome all these projects. It doesn’t matter if you use Ubuntu or Mepis or Knoppix or Damn Small Linux or Kanotix or Libranet or whatever. As long as you’re using a Debian based distro, you’re also a Debian user. There’s one swirl to rule them all… And that’s Debian.”
Well said! I for one consider Ubuntu to be my preferred _flavour of_ Debian.
debian for me! track testing, stay away when big changes are happening! more stable than most and mostly up to date!whats wrong with that
plus seems my older hardware is not supported as well with the new 2.6kernels….
ok heck, i will admit i still love old woody, no gnome or kde, just a light window manager, let gmc draw the icons and do my file manager work…… and i am good to go and the fastest booting 300mhz with 64mb on earth
i think
I’ve read here that ‘LinEx is going to base on Ubuntu’. I want to clarify here that that’s simply not true. Ubuntu doesn’t fit.
Debian + (Suse-Suckfactor) = Ubuntu
IMHO
alter as you see fit.
-nX
… I’d run Winblows, MacOS X, or Be. If I wanted a developer platform that offers lots of server capabilities, maybe sprinkled with *basic* desktop components, then I’d go linux. As it is, I fall into the latter category with my work laptop, and I run Debian experimental. It’s great. But really, to go with Linus’ original call to arms for Linux, I don’t expect Linux to “just work” for me – I expect to have to make it work. After all, if it just worked to install X.org, what would I say I did over the weekend at the water cooler the next day? It just worked? Boring. I prefer to work for my satisfaction when it comes to Linux. I don’t deny the convenience of apt-get blah, but then I don’t depend on it to have my stable, secure, up-to-date system. That in mind, I consider Debian a non-distribution – the best kind. Its more of a starting point for my various Linux installs. Get the core in place, usually consisting of a kernel to boot with, a filesystem, and a compiler suite, and build from there. If a Debian repository has what I need, then great, I save a little time. I don’t get all pissy because I can’t use a repository though.
Security for any system is not something you just install, you work for it. OpenBSD certainly ships a system more secure out of the box, but thats mostly because they take the guess work out of configuring the system for sysadmins that don’t know what they’re doing. Yes, you have to deal with software security issues (buffer overflows, etc.), but do you really want to depend on someone else to handle the security aspects of your software if you’re the one responsible for it? I’d rather get my hands dirty, or not be responsible for it in the first place.
I suspect that a lot of the developers working on Debian are just as eccentric and alternatively motivated as any developer out there can be. Maybe Ian cares that fewer people use Debian as an actualy distro, but if some geek’s got the fortitude and committment to be a Debian developer now, as-is, that geek probably won’t jump ship on the Debian project – whatever they decide to do with it.
Just to point out the obvious.
Debian stable is “Woody”
Ubuntu is Debian testing “Sarge”
Kanotix is Debian unstable “Sid”
all work just fine on any system I have installed them on.
I actually have had more problems with Debian “Woody” and Ubuntu “Sarge” than Kanotix “Sid”
I am yet to have any issues with Kanotix and frequently use it to salvage or repair other systems, Linux and Windows.
Just my 2 cents