The lack of a new stable release of Debian GNU/Linux since July 2002 is fuelling the campaigns of many candidates for the project’s leadership position, with many pushing for a shorter and more regular release cycle to arrest user departures.
Debian leaders: Faster release cycle required
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
So they’ve decided that they speeding up releases might be important. Now we only need to wait 2 years for the voting, mailing list arguments, and delays before they begin to actually implement it. Maybe a couple years after that we’ll get a quick release. Woot!
Well, it’s not much problem for me to run a Testing or SID branch of Debian, however, if I could get a stable release form Ubuntu with same stuff that sits in unstable Debian, then I have no chooice. 🙂
I’m still running Debian on my hope server, but all my desktop and servers at work are taken by Ubuntu. Oh, and a Slackware laptop.
I guess fastering the releases cycle wouldn’t harm Debian much.
Uhm.. brought into sharper focus with the Kubuntu article below this one – this site rocks, love the contrasts
I used the sid branch as my dekstop machine for about a year and I never had a problem.
Even with sid some upgrades are a bit slow, though.
Every day I launch apt-get update apt-get dist-upgrade.
But today I saw that GEdit has been update from 2.8.2 to 2.8.3.
Hey gnome 2.8.3 has been released about one month ago!
And I’m not speaking of a move from 2.8 to 2.10,“ simply” from *.*.2 to *.*.3!
FreeBSD updated it to ports on 12 Feb 2005!
(and hey packages for -stable are built every 1/2 weeks if I remember right)
Does it mean that FreeBSD Team is not doing the right tests for that package ?
What’s the real reason for such delay for a small dot.dot.fix ?
p.s.
GEdit is just an example.
The explanation is easy. At present more open source software gets written/updated in the given unit of time then Debian developers are capable to pack. May be you have noticed the following DD mantra: “Stable release of Debian will be shipped when it is ready”. In present situation it means NEVER. This turns the whole Debian idea into mission impossible and no wonder finally at last some of the participants start to scratch their heads.
This was a VERY frustrating situation half year ago, but now it’s OK. Ubuntu is a very good distribution. In fact you get full power of Debian as it should have been and some Afro-skin as bonus. Why not ?
debian stable should always be stable – there is a niche which they serve.
a solution is to have a new stream or branch which serves the less conservative desktop users.
Why ubuntu is better than debian ? Why ubuntu don’t contribute
with packages to debian ? Why the two projects don’t join ?
The communities need 2 distro so much like ? Debian is a world
where ubuntu survives.Think that! In debian you are in control.
Having the right leader is important, and the leader should not be selected based on seniority. The natural leader probably will not volunteer for the job, but anyway, I agree that a shorter SDLC and release would make Debian a more visible choice.
Why ubuntu is better than debian ? Why ubuntu don’t contribute
with packages to debian ? Why the two projects don’t join ?
The communities need 2 distro so much like ? Debian is a world
where ubuntu survives.Think that! In debian you are in control.
Ubuntu does a freeze of the Debian packages every 6 months and creates a stable release from it. Ubuntu focuses on the desktop (specifically Gnome), Debian doesn’t. Ubuntu does contribute bugfixes and improvements upstream to Debian, but it doesn’t make sense for them to contribute packages as they have different stable trees due to different release cycles. Ubuntu is good for people who like Debian and Gnome as it provides more recent stable packages, but I doubt it’s definition of stable is as stable as Debian’s. It does make sense to have a desktop distro with frequent releases, and to have a distro with longer release cycles for those who don’t care about having recent packages, so Ubuntu and Debian should remain seperate as they fulfil different requirements. Though I do agree a release cycle approaching 3 years is a bit of a joke and should perhaps be shortened a bit.
if I could get a stable release form Ubuntu with same stuff that sits in unstable Debian, then I have no chooice.
Please notice that Ubuntu can currently support only a small amount of all the packages that are available in Ubuntu’s package repositories. (Hey, have you ever wondered how the whole Ubuntu release can fit into just one CD?) All the packages in the universe repository are officially UNSUPPORTED — Ubuntu has taken them from Debian’s unstable branch and bug reports against these packages are sent from Ubuntu developers straight to Debian developers. Ubuntu developers only take responsibility for the stability of the packages in Ubuntu’s main repository (which contains Xorg and the latest Gnome, and not much else).
As an on-topic side note, it’s good to see finally some sanity in Debian’s release policy. IMO, they should support less architectures and less packages for stable releases, which would make the release cycle considerably faster.
Note that the entire set of Ubuntu supported packages do *not* fit onto one CD. There are over 2000 packages in the supported set, covering the desktop (what’s on the CD), all the popular FOSS server applications, and heaps of other stuff. The main repository covers a lot more ground than “Xorg and the latest GNOME”. 🙂
The main repository covers a lot more ground than “Xorg and the latest GNOME”.
OK, I was a bit unfair. Still, the Ubuntu main repository covers a lot less than most new Ubuntu users seem to realize. Ubuntu definitely wouldn’t be as popular as it currently is if there wasn’t all this unsupported software available in the universe repo. But as long as Ubuntu keeps contributing packages back to Debian, I consider Ubuntu as a Gnome-oriented cutting-edge development branch of Debian.
For me Gnome 1.4 on Debian-stable is still too unstable.
Maybe if debian’s installer was a little easier to use you’d get more users, as would Gentoo. If Debian’s president is going to be focused on attracting more users, why not make it easier on them adopt it?
Everyone who uses knoppix or mepis or ubuntu with a harddrive (not referring to those who use them as live-cds) install is a potential debian user.
“my first focus will be on our release schedule. I have a proven history of releasing software on time, on schedule.”
Are you kidding me? scheduling is the problem? Anyone can release software on a schedule, how about WORKING SOFTWARE?
I use Knoppix at home and love it–after I customize the crap out of it. Basically I love debian, but wouldn’t mind if it was easier to install.
So when there’s no release time no one will be ready for a release because if my package/responsability is not ready doesn’t matter so I can take 1 or 2 weeks to make ready and with no package versions specified for release I can be putting new packages and freeze “when its ready”.
Don’t get me wrong I like debian but woodys age hits me bad.
Debian NEEDS scheduled releases even if it’s 18 – 24 or 36 months and every voluntier should work with that on mind and don’t loose the way.
And if free ( libre ) software moves Fast debian should be faster than it is now. who really uses sid or sarge without security as a producction server ?
Think : Do you believe that sarge with gnome 2.8 will be attractive as a desktop when there are other distros with gnome 2.10 ? all the work to put gnome 2.6 and 2.8 on sarge where is now that gnome 2.10 is released and gnome people want gnome 2.10 and sarge is freezed ?
Have you ever really used the (current) installer? It’s totally sane and makes sense. No it’s not graphical, but it’s clear. CFdisk to create your partitions is easy (lots easier than fdisk for example). Loading your modules with modconf (during install) is quite easy too.
You don’t have to be a genius to understand Debian’s installer and if you actually read the stuff that’s on your screen you will see that installation is a breeze.
Havind said that, the new installer (which is on it’s way) makes it even more easy to install it.
Actually yes I have used the new installer–that’s why I mentioned it would be nicer for NEW users.
No, you don’t have to be a genious to install it. CFdisk IMHO is the same as Fdisk in terms of complexity–if not more since you need to know how to setup a swap file first, then a linux partition.
Have you tried windows XP’s installer? It’s not super graphical off the bootcd until 1/2 of it is installed.
Yes the new installer is LEAPS AND BOUNDS better than the old, especially considering the new one’s hardware detection.
But if the goal is to attract new users (since current ones are obviously capable of installing debian,) the installer ought to make it VERY EASY to run. If Debian’s installer was the same as the other distros I mentioned (knoppix takes a little work, but whatever,) MANY MORE PEOPLE WOULD USE IT.
As far as Release dates go, forget development: When it’s tested as stable, then it is stable. Put together a TEST schedule and TEST PLAN. That should dictate the release schedule for their Stable branch. That makes much more sense than simply releasing on a set time–if it ain’t tested how do you know it’s stable?
The new debian installer doesn’t use cfdisk. It uses partman based on libparted, including a few autopartition schemes. Are you sure you tried it? There is a new (third) release candidate planned for the 23th of march. http://www.debian.org/devel/debian-installer/
I guess I was using the old “new” installer from a while ago. When I get home I’ll give it another try with a new ISO.
anyways–The very things that make Fedora and Mandrake attractive to new users should be at least examined for debian IF the goal is to get new users, or to attract more “distro hoppers.”
Debians package management rules, and synaptic is great. Is there a debian DVD ISO out there?
I for one would be happy with sarge releasing with 2.4.x kernel only, apache 1.3.x and KDE 3.1.x if it happened today. Maybe they try to release to up-to-date software?
Wait for RC3, there will be a DVD iso too.
…my Sarge systems haven’t had any updates in days. :/
Change your mirror.
Duh!
is Duke Nukem 4ever ready by then too?
I am using Ubuntu since I suffered filesystem corruption in Debian’s partition (I used it for 2 years, more or less happily). Since I “had” to switch to Ubuntu, I ran into quite a bit of bugs, in a distribution which calls itself stable (ok, so it was v1, maybe Hoary will be better).
I don’t think Ubuntu can act as a Debian replacement for all of us, I think of it more as “Debian-based Mandrake”. Some packages will never receive all the attention they got in Debian (Common Lisp anyone?). Plus, mixing Warty+Hoary was troublesome, to say the least, while Debian testing+unstable was almost trouble free. So, in Ubuntu you can’t always get the latest and greatest, even if you run their development branch (which is a frozen snapshot of Sid, except for Gnome related things and a few others which receive active development).
Debian chokes itself in bureaucracy. For me the best release policy I have dealt with is the one of Gentoo. “Stable”, for the most part, updated continuously, and “unstable”. Plus checkpoints 4 times a year, which receive extra bug squashing. Gentoo’s only downside, really, is the compiling from source part.
Anyway I have the deepest respect for the work done in Debian, and I hope they will take the bold steps needed to reorganize the way Debian is made.
I really don’t understand what all the hype surrounding Ubuntu is all about. It’s nothing more than the debian installer and some modified packages. Why would you reinvent the wheel when Debian has been doing the work for the last decade. What happens when the debian installer starts getting old? Will Ubuntu create their own installer? If it’s a pretty installer you want, just use Progeny’s Componentized Linux which uses Anaconda as the installer, is compatible with Debian’s Sarge branch, and includes a great choice of software. Okay, Ubuntu has managed to get X.org working on Debian, but there are repositories out there now that will allow you to install X.org on native Debian. I haven’t seen anything in Ubuntu that Debian can’t do or is worth doing. And it isn’t the easiest Debian system to install. Give Progeny a try. Personally, I use the chroot method to install Debian. But that’s me.
I like what Matt Garrett had to say about of releasing more often and getting sarge to go GA. I have been planning to use debian for my sparc’s and I would hate to see it go.
Linux for these other cpu platforms really helps alot of people out. Systems are expensive and hand me down hardware can go a long way with linux. I hate to see all these systems become land fill.
I guess I should help out more and I have to find out how. We could use a dev leader that gets people involved to get releases happening faster.
I just gave Debian a shot for the first time about a month ago. (For context, my first ever linux install was Slackware ’96.) The plan was to put it on a TiBook. . . and up until about two days ago, I was still trying to hassle with Woody (which is most definitely not cut out for PPC computers, as far as I can tell.) It took me _that_long_ to figure out that Sarge is more than stable enough for most desktop installs.
Please, Debian, come up with a new word. In the rest of the OSS world, “Testing” means, “You’ll be filling out lots of bug reports and hopefully doing some debugging yourself.” While I realize that Sarge isn’t as rock-solid stable as Woody, I don’t get the impression that it’s bad enough to warrant the label “testing.”
> I really don’t understand what all the hype surrounding
> Ubuntu is all about. It’s nothing more than the debian
> installer and some modified packages.
Actually, it’s a bit more. It has a 6 month release cycle. Every version, SID is frozen and stabilized. For the next 18 months, this version recieves security updates. Before the next release cycle starts, all customization and additions are fed back into SID (making it more stable). The development version branch is then based off of SID and all changes are done directly on it until the release.
Essentially, Ubuntu *is* Debian and it helps make Debian stabilize quicker, so there’s no reason for diehard Debianers to frown on it.
BTW, it’s a bad idea to mix Warty and Hoary. Hoary is essentially “Warty Experimental”. Mixing SID and Experimental is at least as troublesome. If you want some of Hoary’s features in Warty, it’s better to use one of the backports.
I can’t afford a new computer. I don’t know how old this one is, but the front panel is held together with sticky-tape, and the CD drive opens only on good days. If I was to run any other distro, it would have to be at least 2 versions out of date, otherwise it would crawl or just not go. Yet I can run the current version of Debian and have access to patches and security fixes. Debian makes me happy.
I like what Matt Garrett had to say about of releasing more often and getting sarge to go GA. I have been planning to use debian for my sparc’s and I would hate to see it go.
I haven’t read the discussion around it but i’d also find it really sad when the SPARC, ALPHA, MIPS ports -among others- would go. Its not as if these people would work on the 4 remaining ports then? Its not as if all these ports are exactly the same now or support the same software? I mean, thats not the case with NetBSD either.
Nah, frankly, i don’t think thats the actual problem. The problem is (IMO) rather that there has to be a hard line. A tight schedule, such as the OpenBSD and the GNOME teams are following. All the dates are set, so people can have rightful expectations and work ‘to’ these. Now, i’m up for a bit of flexibility too and don’t say “go use that (e.g. OpenBSD) instead”; the thing is that sometimes, a bit of force and authorianism pushes to force something to a certain way, isn’t all that negative, instead of keeping argueing and improving. Sometimes, we have to do instead of thinking about doing. Perhaps minimizing project’s goals (such as less architecures, 2 editions (server and workstation/desktop) may help.
Just a mind-boggle anyway 😉
I don’t see a problem in Debian’s release cycles. I’m using testing and it works very nice. Is this “problem” just wanting the newest software because it’s the newest?
There’s a point in the argument that many users try out Woody to start with, the naming of Sarge isn’t that genious after all this time. But that’s why it’ll become the ‘stable’ Debian in the coming months (or years?). Ah it’ll come when it’s ready. No need to have a tight schedule, let the people relax and get the ‘work’ done while having fun. Not to hasty, it’s easy to forget to play in this fast world…
The problem is that testing DOES NOT HAVE SECURITY UPDATES and that’s a big problem
The problem is that testing DOES NOT HAVE SECURITY UPDATES and that’s a big problem.
It does, but until frozen they’re not official DSAs. You still see the security fixes in the changelog. There still *are* fixes and they’re getting there pretty fast it seems, faster than in the past (except for kernel-image IMO).