Debian Leaders: We Need to Release More Often

daria42 writes "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 Debian Project Leader role, with many pushing for a shorter and more stable release cycle to stop Linux users heading for greener and more updated pastures."

Comments

This is comical..

[Anonymous Coward]

July 2002 .. you've gotta be kidding me.. right ? Another Slasheditor typo ?

I thought Debian was an enthusiasts distro..

Re:This is comical..

[ta bu shi da yu]

Nope. This is correct.

Re:This is comical..

[LordoftheWoods]

Indeed. The whole Debian stable rationalization is actually pretty easy to explain.

I believe the meaning of the word 'stable' is doesn't change often.

Or was it "So placed as to resist forces tending to cause motion."

stable as in stability, right? Isn't stability supposed to be a good thing?

That in mind, I do agree releases a year or so more often would help Debian. But for some people only having to update every few years is a great thing, they don't want upheavals on their servers every 6 months. This is the kind of people Debian stable serves. All of the rest use testing or unstable. They should make the website be more clear that stable is not for desktop users who want recent stuff.
There really isn't anyone working on Debian full time, and it's release pace reflects this. Debian is, well, different.

Re:This is comical..

[LordoftheWoods]

Oh, and I forgot to add.

Unstable - changes often

for any slow people out there. English, anyone?

Re:This is comical..

[digital-madman]

Dude what about RHEL? I mean seriously.. Thats a 5 year support cycle. Sure, I woulndn't pick fedora as a first choice on a server. But Debian is not the only choice with out massive updates all the time.

Re:This is comical..

[Waffle Iron]

However, in engineering, "unstable" often means "buggy", "defective" or "dangerous". That's what comes to mind when people hear that word; they don't refer to their dictionaries to look up less menacing definitions. The term "testing" isn't much better either.

The Debian project should really change their terminology if they don't want to scare people away unnecessarily. Any marketroid would tell them that it would be better to go with something like "Enterprise Edition", "Personal Edition" and "Exxtreme! Edition".

Re:This is comical..

[Anonymous Coward]

"Enterprise Edition", "Personal Edition" and "Exxtreme! Edition".

If you were running for that Debian Project Leader Role, I would vote for you.

We're using SuSE because we can't use pacakges from something called "unstable"

Re:This is comical..

[m50d]

There's one big problem with the Debian system: testing doesn't get security updates. Unstable doesn't either, but they'll get it as soon as the project releases its own updated version. But testing keeps the same packages for quite a while, and is in the right place in terms of modernity/stability for many desktop users. If it got fixes and security updates, it would be a very useable system.

Re:This is comical..

[tuxmaster]

you also got to put in mind the work that it takes to measure "stability". When all in fact someting that looks stable may be completely unstable just becouse you have not tested it yet and it could be a key feature in the program. What would be interesting is looking at the vcs (version control system) that the debian project uses. And the thing about open scouce that makes so many pepole scared of it is that there is no big huge company watching over it with milliions of dollers to pay programers to releace things on time. However Large highly supported Openscouce projects get fast updates simply becouse someone finds a bug sudmits it and the developers fix the bug. You got to remember how many bug fix supports that a OS would get.

Re:This is comical..

[Anonymous Coward]

I thought the three branches are "stale", "rusting" and "broken"...

Re:This is comical..

[tacocat]

I for one am very glad that Debian does not take your advice. The differences between Enterprise, Personal, Exxtreme are arbitrary and subject to interpretation by the user and mardetroids.

My desktop might be your extreme. Who are you to tell me what my installation is?

Re:This is comical..

[schlick]

I believe the meaning of the word 'stable' is doesn't change often.

Hrrm I thought 'stable' was a place you keep horses.

Re:This is comical..

[lakerdonald]

Well, why else would it still ship with a 2.2 stock kernel? Cause it's so old.

Re:This is comical..

[say]

There's one big problem with the Debian system: testing doesn't get security updates.

This is a myth. Testing gets lots of security updates, from both security.debian.org and through the extremely rapid propagation of "normal" upgrades that packages get. Most maintainers seem to propagate security-related bugfixes within hours.

If you use very rare packages with slumbering maintainers, you could probably be in loss of security upgrades, though.

Re:This is comical..

[say]

Any marketroid would tell them that it would be better to go with something like "Enterprise Edition", "Personal Edition" and "Exxtreme! Edition".

Anyone who cares about such things should go use RHEL. Debian is not about marketroid thinking. To those businesses who use more expensive, worse solutions than debian because debian's "modern branch" is called testing: their loss.

Re:This is comical..

[cmdr_tofu]

Well, actually, you are incorrect. Testing does get security updates. I doubt they are made available as quickly as the security fixes to stable, but I guess that depends on the individual package maintainer.

You can see my apt source for debian (TESTING)
security updates.

http://security.debian.org/debian-security/dists/t esting/

Re:This is comical..

[Kick the Donkey]

The differences between Enterprise, Personal, Exxtreme are arbitrary and subject to interpretation by the user and mardetroids.

Please! All the labels are arbitrary, as other commenters have pointed out. Stable, Unstable, and Testing? I've got a pretty good idea what 'Stable' is, but, without looking at debian's site, I can't tell you what the differences are between 'Unstable', and 'Testing'...

With that said, Waffle Iron's suggestions wouldn't work, either (however, I believe he was JOKING. Try turning up the sensitivity on your sarcasm detector. That might help you around here).

Perhaps something more sane like:

• Stable
• Current (formally known as testing)
• Development (formally known as unstable)

To me, those are less open to mis-interpretation, as well as scaring people off with nasty words like 'testing' and 'unstable'.

Re:This is comical..

[pdbogen]

Perhaps, but those are wrong. Testing is what's going to be the new stable when it's done being tested, and isn't current. Unstable is more like current, but is, as the name implies, somewhat often unstable, in every sense of the word. It's the front line of new releases, and sometimes there are casualties. Also, people often forget (or don't know about) Experimental. The newest of the new, the crack commando unit in the war against outdated packages. Anyway, if you really want, you can just call them 'Woody', 'Sarge', and 'Sid'. And, of course, the Experimental branch is secret and nameless.

Re:This is comical..

[Anonymous Coward]

Stupid! I love Debian stable, I don't have any need for more current versions of anything. Very long release cycle is one of the gretest benefits of Debian.

Debian -- Where stable really means stable.

Re:This is comical..

[Anonymous Coward]

it's not comical.. it's right on track... for debian.

if it ain't broke, don't fix it. debian's release schedule is not broken.

debian needs to place an emphasis on the testing branch for desktop users; stable for servers; and unstable for the bold and brave; and experimental for the just plain crazy. and stress that debian is not a n00b distribution or for people who are unwilling or unable to read a txt file.

one of the things that hurts debian is that over the course of the development cycle, _every single version_ of many packages, ends up in the pipeline. instead of uploading every point.point.point release of package xyz, into unstable and then migrating it to testing. seems to me that deb developers just run around in circles for 2-3 years inbetween releases.

someone needs to send debian a c&d against new version uploads so they can finally get sarge frozen; then in 2-6 months you can see a release of sarge as stable.

then instead of continually updating every package with every available new version over the course of the next three years or so... pick and choose major point releases to upload (and maintain) into the distribution. at about 2.5 years out, update packages with the then-current versions and freeze again. at stable + 3 years the next should be ready for release.

Re:This is comical..

[Knuckles]

"Experimental" is different as it is not a complete distribution, but only bits and pieces. Also, stuff from experimental will never be installed by apt-get/aptitude unless you explicitly request it

Re:This is comical..

[Anonymous Coward]

No typo there. Deb is an enthisiast distro, but most enthusiasts use either the testing or unstable branch.

Re:This is comical..

[eraserewind]

I'm not going to dispute your definition, but surely such a system is kind of pointless? Why have official releases at all with such a system? Why not just stick with the ocassional upgrades as they happen.

If you want official releases, better to just make a 6 month release cycle and stick to it. Support decreasing levels of bugfix for the last X releases. Then 1 year old version is "reliable, very few changes" 6 month old version is "semi-regular bugfixes happening", released today version is "acceptable test pass rate. some known bugs, regular fixes expected", and working branch is "compiles. use at your own risk". Released versions don't get new packages, just fixes or minor upgrades. The old software doesn't go away, so it is always "stable", but doesn't get in the way of something more up to date for general users, and bleeding edge people.

Re:This is comical..

[pipegeek]

You can see my apt source for debian (TESTING) security updates.

http://security.debian.org/debian-security/dists/t esting/

Maybe I'm missing something, but it seems that there's nothing there. The package lists for main, contrib, and non-free are all empty files.

Re:This is comical..

[m0rningstar]

I run a hybrid testing/unstable on my desktop. On servers I have to run the same thing in some cases; especially on mail gateways for spam/virus filtering -- the two year old version of spamassassin in Stable is ... inefficient.

However, it's not the instability of the testing/unstable releases that worry me. It's the fact that they are explicity not supported by the Debian security team. So running that on Internet facing servers worries me. A lot.

I don't care about the naming convention. I'm smart enough to go read the docs and figure out what they mean. But when I either have to run two year old versions OR accept a lack of security coverage I become unhappy.

Re:This is comical..

[eventhorizon5]

Some commercial operating systems release new stable versions every 2 years, which is what I consider the optimal release time for Debian (I've run Debian on production servers, and it's important to not have to worry about a system-wide upgrade every few months. I've already started the planning for the upgrade to Sarge on one machine, once it's released - my desktops run Testing).

For example, Sun releases Solaris every 2 years. Debian normally releases every 2 years also, but it's already 2 1/2 years since Woody's release. Even Windows is normally released every 2-3 years (Win2k came out in late '99, XP came out in late '01, but there's been no new desktop OS since then, except for the server OS Win Server 2003)

-eventhorizon

Re:This is comical..

[Ogerman]

Testing gets lots of security updates, from both security.debian.org and through the extremely rapid propagation of "normal" upgrades that packages get.

Yes, testing/unstable gets security updates, but they are not a priority. Priority as in "people are using this in production.. they need immediate fixes." Remember the recent uselib local-root Linux kernel bug? As far as I can tell, it was months before this was patched in testing/unstable kernel images. Or if it was, it sure wasn't communicated properly.

Re:This is comical..

[Short Circuit]

IIRC, the fact that a package is in the "stable" repository doesn't have anything to do with bugfix status. Mostly, it means they avoid adding new bugs by not adding features, and fixing old bugs.

I could be wrong, though. It's been a *long* time since I was up-to-date on Debian/stable philosophy.

Re:This is comical..

[Anonymous Coward]

Perhaps something more sane like:
* Stable
* Current (formally known as testing)
* Development (formally known as unstable)

That's not a naming system.

* Current Release (formally known as Stable)
* Development (formally known as Testing)
* Bleeding Edge (formally known as Unstable)
THAT'S a naming system.*

gewg_

*Hat tip to Paul Hogan.

Re:This is comical..

[eluusive]

• Stable
• Current (formally known as testing)
• Development (formally known as unstable)

Actually this is a common mistake, and shows just how wrong the terminology they use is.

Here are the descriptions of the branches:
testing - Brand new stuff that's quite possibly broken.

unstable - Packages go here after they've been in testing for two weeks without hosing anybodies machines.

stable - Unstable is frozen, (no new packages are accepted) everything in it is then checked for bugs, which are fixed, and then it becomes the stable branch.

So it'd be more like this:

• Stable
• Current (formally known as unstable)
• Development (formally known as testing)

Re:This is comical..

[joto]

So it'd be more like this:
* Stable
* Current (formally known as unstable)
* Development (formally known as testing)

No! The parent was right.

Originally, there were only stable and unstable. Testing was added in between stable and unstable in order to give some extra quality checks so they could release stable more often.

The effect was probably the opposite. Instead of being a testing ground for the next stable release, testing ended up being another intermediary step packages had to go through before they were declared stable. And since any bugs in testing would only be fixed in unstable, the end result was that nothing new would ever end up in stable.

Testing was and is a mistake. For it to work as intended, it should only exist for a month or so before each new stable release, and in this brief period, people should actually care to fix bugs in testing. After the new stable release, testing should go away, and everything should happen in unstable

Instead we have a situation with a testing release that seemingly forever blocks anything from ever reaching stable, because nobody fixes bugs there.

From the user perspective, it's also confusing. Many people are afraid of running unstable, and prefers testing, assuming the packages there have at least a little bit more quality assurance. That is a false belief. Testing has no security team. If there are bugs there, they will be fixed in unstable (usually very fast), and then take a few weeks to propagate down to testing. So security-wise, testing is the worst of both worlds.

The only assurance testing gives you, is slightly fewer completely broken packages. But anyone that actually have tried running unstable, knows that this is very rarely a real problem. Unstable is still "stabler" than most other distros.

Re:This is comical..

[argel]

Testing is what's going to be the new stable when it's done being tested, and isn't current. Unstable is more like current, but is, as the name implies, somewhat often unstable, in every sense of the word. It's the front line of new releases, and sometimes there are casualties

Then how about the following?

Stable
Release Candidate
Development

Re:This is comical..

[LordoftheWoods]

Ha! It will if you pin it up to a normal prio!

I love it..

Package: *
Pin: release a=experimental
Pin-Priority:500

Package: *
Pin: release a=unstable
Pin-Priority:500

Re:This is comical..

[LordoftheWoods]

Although I would not choose those names, I do agree we Debian users need an ambitious leader who is daring enough to propose such small but important changes ;)

Break with tradition

Re:This is comical..

[LordoftheWoods]

Then what is experimental?
Packages "so critical such that they might self-destruct at any-moment?"

Yet I have no serious problems. Yes, there are problems. Problems that normal users would not have a clue what to do with. But I do! And they add spice to my life!! Honest!

Stop making fun of me =/

Re:This is comical..

[LordoftheWoods]

Uh.. maybe there are no known security issues right now?

They're security updates; I'd imagine once theres a new version of the package on all the mirrors the fixes would disappear.

Re:This is comical..

[LordoftheWoods]

I guess.

Though Microsoft has been known to take forever for some fixes.

There was a contraversial study done on this:
http://www.microsoft.com/windowsserversyste m/facts /analyses/vulnerable.mspx

Although Microsoft's description on their site is biased and laughably inaccurate.

MS:
Relative Severity: Windows has the fewest vulnerabilities and the fewest "high severity" vulnerabilities of any platform measured.

STUDY:
"ICAT Classified 67% of Microsoft's vulnerabilities as high security, placing Microsoft DEAD LAST among platform maintainers by this metric."

Some contreversy:
http://www.computerworld.com/securit ytopics/securi ty/story/0,10801,92313,00.html

So, basically:
Microsoft generally fixes fast, but their security still sucks because the bugs allow the system to be completely compromised. By the time the fix is mainstream, its too late.

Re:This is comical..

[LordoftheWoods]

neg, serious ("Release-critical") bugs are not allowed into stable.

Re:This is comical..

[Short Circuit]

Thanks for the clarification.

Re:This is comical..

[eluusive]

What on earth... I just looked at the FAQ and you are indeed correct.

When did this happen? When I first started using Debian unstable was not sid, it was woody, and then it became sarge. And testing was sid.

The page I looked at a five or so years ago was http://www.debian.org/releases/, but that too is fixed now. I know it wasn't this way a year ago either. I made a bet with a friend of mine, who was using testing, about it, and I won when we checked that page out.

Am I just delusional or something?

Re:This is comical..

[Knuckles]

Sure, but then you explicitly request it :)

Re:This is comical..

[chrj]

Here are the descriptions of the branches:
testing - Brand new stuff that's quite possibly broken.

unstable - Packages go here after they've been in testing for two weeks without hosing anybodies machines.

That is not right at all. New packages first go into unstable. Testing is the distribution to become stable sometime. When every goal for the release is met, testing is frozen and all release-critical bugs are fixed. Then it becomes stable when a new version of Debian is released.

In my opinion it is meaningless for a normal user to run Debian Testing, as it is meant for QA and developers to 'test' the packages to go into stable.

Furthermore, when a bug is introduced in testing the user must wait for it to enter unstable first. That process could be quite cumbersome for joe average who thinks he gets a more stable unstable.

Regards,
Christian (running Debian Unstable :))

Kubuntu Hoary Snapshots (KDE 3.4)

[Anonymous Coward]

Here's some fresh Debian:

http://osnews.com/story.php?news_id=10020

Re:Kubuntu Hoary Snapshots (KDE 3.4)

[Storlek]

That's not Debian, that's just Debian-like.

Re:Kubuntu Hoary Snapshots (KDE 3.4)

[Anonymous Coward]

> That's not Debian, that's just Debian-like.

It uses apt, dpkg, and almost every package, including core packages come over verbatim from debian's official repository. Even gnome is the same, with ubuntu-specific configurations. Every last config file is in the same place as Debian. The split is political, not technical.

Knoppix is Debian-like. Ubuntu is Debian with configuration diffs.

well..

[schnits0r]

I would like to be the first to say "duh". Debian is old. Despite it being stable, it's often a good idea to have the newest programs to keep up with the newest technologies.

However, I do find that using a netinstall version of the "testing" release tends to keep up to date with most packages.

Re:well..

[ctj]

I agree I run debain testing as It offers almost the latest software with out more of the breakages you can get with unstable.

That being said it would be good to get a stable release more often. Rather than having to using testing all the time.

Re:well..

[Amiga Lover]

What are the problems with releasing debian more often, that have caused it to become older and older? I think it's 3 years soon since Woody was new.

I've heard it mentioned that some packages are keeping things back, and by the time those packages are ready, there are others being kept back. it's a duke nuke'em kind of situation

Why not aim for a 12-monthly release? Go over by a month or two if absolutely needed, but aim for that. Even if some packages were missed the first time around and left the same as the old ones, then damn... they could have been caught up three times over already (assuming yearly releases over the last 3 years)

Re:well..

[lullabud]

Even if some packages were missed the first time around and left the same as the old ones, then damn... they could have been caught up three times over already (assuming yearly releases over the last 3 years)

I think that defeats the idea of a stable release. The test versions of Debian are released weekly, and from my experience they work perfectly fine. In fact, a few weeks ago I even had a problem installing Debian on a Dell SC420 because the installer didn't have support for SATA, but the following week's release put that support in. I think it's important to realize that the slow release cycle is just for a stable release, which is rock solid, and not for releases in general. Personally, I like the way they do it now.

Re:well..

[dondelelcaro]

The test versions of Debian are released weekly, and from my experience they work perfectly fine.

Testing gets updated daily, not weekly. (Katie runs at least every 24 hours and the mirror pulse happens soon afterwords.)

You're probably refering to d-i which does have snapshots which get updated every now and then, but it itself is updated all the time.

Re:well..

[Billly Gates]

Back in the 20th century Debian was not that old.

My guess is FOSS really took off unexpectingly and Linux became ported to more architectures besides x86 and the Alpha. This caused the folks at Debian to focus on everybody including the atari users.

If a bug was fixed for most platforms but the amiga users (all 15) was still present, then package X would not be updated on any of the other releases. This is whats hurting it.

I hate to say it but the x86, powerpc, and sparc versions should be ahead and have a later version then the others. FreeBSD for example has alpha and powerpc as different tiers of support, although alpha is still pretty stable.

Re:well..

[Anonymous Coward]

> I think that defeats the idea of a stable release.

There's a point where "stable" also becomes "stale". When the flagship in stability lacks so many features that its usefulness is degraded.

Woody has already passed that point for many people, and is only getting worse as newer better software is released. (That's 'worse' in a relative sense, as in further behind newer distros and releases. Of course it's always going to function as well as it did the day it was released, but so will a Commodore 64)

Re:well..

[Anonymous Coward]

I've heard it mentioned that some packages are keeping things back, and by the time those packages are ready, there are others being kept back. it's a duke nuke'em kind of situation

This is 100% NOT what has been going on.

Originally there were claims that the installer was holding up Sarge. After that was solved, Sarge needed security update infrastructure. A year later and we're still waiting on security update infrastructure.

Debian's problem now is the same one that it has been suffering under for the past few years. People take key jobs, refuse to do them, refuse to accept help, refuse to explain what needs to be done, and refuse to get out of the way. What Debian really needs is for a DPL to step up that's willing to kick these folks out of their positions and let new developers take over.....none of the candidates are proposing that, so it's a safe bet that Etch will take every bit as long to release as Sarge has taken.

Re:well..

[Todesmetall]

Back in the 20th century Debian was not that old.

In fact, a new version was released roughly every 12 months, at least in the beginning. Then it took about 18 months from potato to woody, and now three years have passed since the release of woody...

Re:well..

[jav1231]

Agreed. We just tried to install it on a DL380-G3 (not exactly bleeding edge) with a 5312 Array Controller (not exactly new) and couldn't. Oh there's a website here and there offering "Yeah, compile this and copy this to floppy but don't use it and then get 3 cups of flour, 2 cups of water, a shot of Jager, and a pinch of salt..." I went back to the guy and said, "Yeah, you won't be running Debian unless you can get the vendor to provide install information."

Re:well..

[Anonymous Coward]

Debian's problem now is the same one that it has been suffering under for the past few years. People take key jobs, refuse to do them, refuse to accept help, refuse to explain what needs to be done, and refuse to get out of the way. What Debian really needs is for a DPL to step up that's willing to kick these folks out of their positions and let new developers take over

Basically, what Mepis and Ubuntu have done. I've given up on vanilla Debian ever making amd64 an official build and having the repository on their mirror system instead of whoever the hell owns Alioth. Ubuntu has committed to a 6-month release cycle, and if the devs there don't do their job, they'll be fired.

I do have my doubts about Ubuntu's business model though...

Re:well..

[rodgerd]

Which is exactly how Ubuntu has made progress: it's the Debian that had the ruthlessness about cutting largely irrelevant architectures and being prepared to prune the core package tree.

Re:well..

[swillden]

I hate to say it but the x86, powerpc, and sparc versions should be ahead and have a later version then the others.

There is a recent proposal being discussed in the Debian community which would do this. The idea is to establish a set of criteria to determine which architectures should be officially released. All of the other architectures would still be available, but only the official architectures would be part of the release process.

The architectures that currently meet the critera are ia32, ia64, amd64 and PPC. I think what knocked Sparc out was the requirement that 98% of the Debian packages build on the platform, so that's something people interested in Sparc could fix.

Re:well..

[Anonymous Coward]

Interesting, as my list of machines includes the Arm. Good thing they support them all, and not just our pet processors, eh?

Re:well..

[jonadab]

Part of the problem is that the website doesn't make it clear exactly what
"stable" means to the Debian people. It just says, "the latest stable release
is 3.0" -- when most projects say something like that, they mean that's the
latest release that's stable enough for anyone other than one of the project's
own developers to seriously consider using. It would also help if the Debian
front page said *anything* about the testing and unstable releases; at first
glance, it appears that the latest stable release is also the latest release
when, in fact, installing it is a trip back in time to the tune of the better
part of a decade. What the front page *should* say is something along the
lines of this:
Current releases:
Long-term production/stable: 3.0 "woody"
Medium-term current/recent: 4.0 "potato"
Bleeding Edge latest/newest: 5.0 "sid"
(I may have those numbers and nicknames off; I'm not really a Debian guy.)
As it stands, I suspect too many people do what I do: look at the front page,
figure 3.0 is the latest release (after all, it says so), go to cheapbytes or
wherever, nab 3.0, install it, and then try to pick our jaw up off the floor
when it asks us if we want to use a 2.0 kernel.

If it's stable, it doesn't need to be updatedOften

[ABeowulfCluster]

I can see the need for keeping ahead of security bugs, but to change for change's sake is just silly.

More stable releases please

[Neo-Rio-101]

I have no problem playing with aptitude from their latest unstable Sarge (it's great BTW), but it makes it very hard for me to recommend Debian on servers to customers when the latest stable release is eons old. Yes, I know there are ways around this... but let's face it, from a customer point of view it's an small image problem Debian has.

Re:More stable releases please

[JayAEU]

Strangely enough, there are actually people who appreciate long release cycles! I have servers running woody which absolutely need nothing newer and I'm happy about the fact that I don't have to change everything every 18 months.

If the release cycle were to be shortened to said 18 months, it would be nice if Debian were to maintain older releases and not only the previous release, like it it now.

I recommend Debian to my customers as a server platform, exactly because it has the finest package management and the longest release cycles. When stability is the goal, Debian is the right choice!

Re:More stable releases please

[Anonymous Coward]

The people that appreciate long release schedules also appreciate REGULAR release cylces. Sarge is a year or more behind where it should be, and customers quite frankly could care less about the purely internal issues that caused the delays.

Re:More stable releases please

[aug24]

What they need is a new release for people like you (and me on my servers). They could call it Debian 'OldButFuckingSolidAsARock'. Then stable could be, say, approximately a year later subject to pretty damn good stability (after all, no software can be guaranteed perfect yet), unstable and testing to follow.

Three levels just isn't enough to grade sensibly from known-near-perfect to bleeding-edge.

J.

Re:More stable releases please

[Anonymous Coward]

Agreed 100%.

I use Debian Stable because it is VERY STABLE.

Security patches installed automatically with no problems for years.

No other maintenence was done to the system at all.

I'm sick of vendors dropping support for functioning products (WinNT, Win2K, RedHat9, Redhat 8, Redhat 6). Debian was the one vendor that actually stood by its products for the long term. If this is changing, I don't know where to turn.

Re:More stable releases please

[Scaz7]

Although if you look at windows 2000 server it's what 5 years old and still in production. I mean we won't see the next windows server till atleast 2007 which sorta makes debian look reasonably good with it's current release cycles.

I mean the windows world and the linux world are two totally different beast's and i will admit other distro's really have taken two steps forward in the stable branches compared to debian woody but the basis is still the same and IMO debian really does the same job better then a few other current distributions.

In the server world I really only rely on Debian for the mission critical stuff and you know what? So far so good....

Re:More stable releases please

[jlehtira]

While I totally agree with the points about stable, I also need features that the stable relic lacks. I need a stateful firewall and other goodies. And sometimes I even need to compile shiny new software which is sometimes a total bitch on an old distribution.

Sure, debian stable is stable, but look, there are other incredibly stable linuxes out there.

Re:More stable releases please

[Anonymous Coward]

What they need is a new release for people like you (and me on my servers). They could call it Debian 'OldButFuckingSolidAsARock'.

"Old butt fucking"? Get yer lemonparty away from me.

Re:More stable releases please

[MrMickS]

However the current stable windows server version is Windows 2003.

Re:More stable releases please

[xSauronx]

yeah but the thing is, if you get a release 3 years ago and you're happy with it...you dont *have* to update every new release, unless theres security fixes you need to apply. nobody from debian is going to go to your office and make you upgrade.

but if they did a release ever 12, 18, or even 24 months, some people would be more likely to use debian as opposed to another distro that keeps up to date. And since theyre working on debian, I would assume they want more and more people to use it....

otherwise its just a hobby, and they wont care, and they'll just do it however they like.

Re:More stable releases please

[_|()|\|]

windows 2000 server [is] 5 years old and still in production

Whereas Windows is just an O/S, a Linux distribution is an O/S and a pile of applications. This is often considered to be an advantage, but it can be liability, when it comes to long release cycles. For a lot of situations, I'd be comfortable deploying Linux 2.2, glibc 2.1, and even Apache 1.3. I wouldn't necessarily be as comfortable with the Python 1.5, PHP 3, etc., to say nothing of the desktop components. I could build newer versions of the required components, but what do I do when I find that GCC 2.95 can't build it?

Re:More stable releases please

[nmos]

Strangely enough, there are actually people who appreciate long release cycles!

Agreed, but there is long and there is LONG. At some point Stable gets so old that just installing that one newer package you really need (maybe Exim or mgetty) means replacing half your system with non-debian versions due to dependancies. Then when it does come time to update to the next Stable you no longer have apt handling everything as seemlessly as it otherwise would.

Re:More stable releases please

[Bishop]

You are compareing apples to oranges. The Windows release cycle is better then Debian's. The windows release cycle is closer to 3 years, and support is extended for 5 years. XP and 2003 have both superseeded 2000. Support for 2000 will cease in June or July of this year. It would be nice if Debian adopted a more timely release cycle while extending security and major bug support to older releases.

That said the windows release cycle is problematic. New features have a tendancy to get wrapped up with important security fixes. It is near impossible to get just the security patches you need. Finally when MS cuts support users have little choice but to upgrade. With Linux and the other Free systems users have to option of continueing support themselves.

Re:More stable releases please

[drinkypoo]

Personally, I'd like a USEFUL installer CD. Like, one that actually has all the modules on it? It's absolutely fucking ridiculous that I should need a driver CD to support megaraid. It's even more ridiculous that when I use the exact same command line to copy the image to a floppy that is listed in the debian install instructions, the driver floppy doesn't work. (Not sure why I'm supposed to need conv=sync, while back in the day when I made my slackware 2.0 floppies if=foo of=/dev/fd0 obs=18k worked fine, but I tried it for this purpose as well with no luck.)

Why is installing every version of linux but gentoo and fedora such a huge pain in the ass if you have anything other than generic ide or maybe sata now? Or an adaptec scsi card, I guess.

Re:More stable releases please

[rodgerd]

When the release cycle is so long the only way to get acceptably modern versions of perl, Apache, postgresql, python, and the like is to either roll my own or hope backports has what I'm looking for, I stop appreciating it. Too many Debian server installations end up looking like roll-your-own distros.

anecdote

[Anonymous Coward]

Debian was the first Linux distribution I ever downloaded, in the summer of 2003. I was on dial-up at the time (and didn't even have my own line, so I couldn't download 24/7), and I remember being worried that there'd be a new release by the time I was done downloading the first ISO. I mean, open-source software moves fast, right?

Should've relaxed.

Re:anecdote

[Ingolfke]

Gentoo was the first Linux distribution I ever downloaded, in the summer of 2004. I was on a P4 w/ a Gig of RAM at the time (and didn't even have my own beowulf cluster setup for distributing compiling operations, so I couldn't compile at full speed 24/7), and I remember being worried that there'd be a new release by the time I was done compiling the first ISO. I mean, open-source software moves fast, right?

I was right. I finished the compilation, only to have to recompile 70% of the system.

Re:anecdote

[aonaran]

Why were you downloading ISOs over a dialup line, that's what net install is for. download the set of floppies, then apt-get the stuff you actually need.

There's a fine line..

[RubberDuckie]

Between releasing too often, and not often enough. That being said, I run Debian at home, and at work. I can run unstable at home, and get mostly current software. Even testing is stable enough for most of the work related stuff.

In related news...

[Anonymous Coward]

legions of pimply faced linux nerds around the world simultaneously "released" the manual way in support of the mostly irrelevant and ignored Debian community

well, are you done downloading yet?

[Anonymous Coward]

debian

[prurientknave]

I suppose an apt-get answer to yum,portage et-al seems appropriate in exchange for the debian written security patches that would only be included in the stable branch. They should focus on i686 binaries instead. Since such a small minority of debian users are still using 386's

Re:debian

[Anonymous Coward]

what?

No, seriously...

what?

Re:debian

[TardisX]

it's called "stable" because that is exactly what it is. rock solid stable.

It's also nigh-on useless. If the application you want to run isn't also in stable, you're almost certainly screwed.

Unless you want something real simple, like an apache install serving up static web pages, forget it.

I like the concept, but it's taken too far.

Perhaps it should be renamed Debian: Fossilized.

Re:debian

[jdowland]

?

There's no reason to have a seperate i686 branch since there's no benefit over the i386 packages for the majority of programs.

Re:Speaking of which, Suse 9.3 next month

[ctj]

debain testing allow you to update your system as packages become abalable with out having to wait for a full release

There is always BSD

[dbesade]

There is always BSD...

Re:There is always BSD

[Anonymous Coward]

There is always BSD...

Not quite yet. Maybe it will be ready for release in time for etch.

Duh...

[Bruce Perens]

Geez. All I can say is Duh. Yes, you need to release more often. Indeed, you need to release once. Release period. Boy, did I goof by deciding to base the UserLinux release management on Debian. Good idea in theory. And I couldn't get all of the time I wanted to work on the project. But I finally got my act together, but Debian didn't release, and didn't release, and didn't release, and still didn't release. And I will start working on UL again when there is a distro to base it upon.

Bruce

Re:Duh...

[LittleLebowskiUrbanA]

How is it going otherwise? Is the project dead? Have you considered any other distros to base it on?

Re:Duh...

[Anonymous Coward]

See what happens when you don't pay his consulting fees? He makes you look bad.

Re:Duh...

[Soko]

IMVHO, ubuntu is Debian Done Right.

Check it out - I'm certain that they'd like the help of a high profile advocate like Bruce Perens.

Soko

Re:Duh...

[Bruce Perens]

It's ready to go, as soon as Debian makes their release.

Bruce

Re:Duh...

[Anonymous Coward]

Take a look at ArchLinux. It has some nice underlying internals that might be a good start for a new distribution.

Re:Duh...

[fixertechno]

I must agree. User Linux seemed like a dumb idea from the start, especially the software choices that were announced. It's vaporware at best, and needless competition when something that works and for the most part seems to fit the goals of userlinux exists. Hell , its also a more original name. Personally I prefer Arklinux, which is perpetually in an alpha state, but at least kde oriented.

Re:Duh...

[natrius]

UserLinux isn't vaporware. From what I understand, it's Debian with a specific set of supported packages. It's hard for a support company to say "I support Debian" when there's no standard set of software to support. UserLinux fills that void.

Re:Duh...

[jbr439]

I'm sure you've thought of this, but why not base UL's releases off of a snapshot of testing or unstable? I believe that's what Knoppix and Mepis do. It does mean more work for you, as you would have to do some kind of QA on the snapshot, but you would get a release.

Re:Duh...

[ninjakin]

no I would say MEPIS is debain done right.

MEPIS ROCKS

www.mepis.org

Re:Duh...

[grozzie2]

The point of releasing a 'distribution' is to solve shortcomings, and create solutions, not to sit back, wait for others to do the work, then rebadge it with your own name, then try take credit for thier work.

If you really want UL to be 'something' and 'out there', why not just do the required work, and 'get it out'. If you have to wait for a debian update, where is the value add in the UL?

Re:Duh...

[Anonymous Coward]

Until UserLinux has a real support system that consists of more than just Perens' home phone number, it's vaporware.

Re:Duh...

[QuantumG]

Does Mepis actually contribute back to the community?

Re:Duh...

[JerkyBoy]

Bruce is correct, and other developers are saying the same thing. I wanted to use VLC in conjunction with a Tcl/Tk app., and was surprised to suddenly find VLC missing from the "testing" Debian distro. A look at the VideoLan site revealed why:

You should not be using Debian testing unless you perfectly know what you are doing. It is almost impossible to support Debian testing and there are no plans to do it.

Kind of a shock to the system, but the problem seems to lie at least somewhere in /usr/lib...

Re:Duh...

[Matt Perry]

IMVHO, ubuntu is Debian Done Right.

I've been seeing that a lot of people are recommending Ubuntu lately. What does it offer over a regular Debian Sarge install? Is it just some new themes and a new installer or is there something significant that it provides that Debian doesn't?

Re:Duh...

[Bruce Perens]

It's not as if I don't work for Debian. Today I am a volunteer on their corporate board and represent them to various standards organizations. More of my time is devoted to working for the entire Free Software community, and I flew 50K miles last year to represent Free Software, doing things like speaking against software patenting at the EU parliament in Brussels, keynoting a GNOME conference in Norway, lobbying in Washington D.C., teaching law students in Hawaii, and briefing reporters at every LinuxWorld show.

Historicaly, I am the author of Debian's fundamental policy document and did a lot of the early work on their system.

I've paid my dues a few times over.

Bruce

Re:Duh...

[natrius]

Newer packages, a wiki full of documentation, a supportive community and nice package management programs (in the development version). Note that the only one of these that Debian can't have at this point is a stable set of up to date packages. Everything Ubuntu does is shared with Debian.

When comparing Ubuntu with other distributions than Debian, things are a bit different. One of the selling points for Ubuntu for me is that it's developed by a community and has a central package repository. It's been a while since I used a non-Debian distro, so I'm sure much of this has changed, but when I used Red Hat and Mandrake, there was either nothing that compared, or it wasn't visible enough. Assuming that other distros have that now, there's the deb vs. rpm issue depending on which one you prefer. The main issue is that you're never considered a second class citizen in Ubuntu. The other distros have commercial versions with special software and updates you don't have access to. With Ubuntu, everything is free, and they've made a commitment to always remain free.

Re:Duh...

[QuantumG]

I think what he was trying to say was, why would someone choose UserLinux over, say, Ubuntu, if there's no new work being put into it? I don't know anything about UserLinux but I thought I'd do my part to stop this conversation spiraling out of control. We all know you "do stuff" for Free Software, your antagoniser just wanted to know what you do for UserLinux.

Re:Duh...

[LnxAddct]

I understand that Debian folk are not too fond of Red Hat folk, but honestly, Fedora seems like a stable base, or even RHEL4. Grab the source from Red Hat or CentOS (which a few red hat engineers actually help out with). You'll have a modern, secure, easy to use base with an up to date kernel and many added security features like SELinux and services compiled with random memory mappings. I'm not trying to start a distrowar or anything, I personally use both debian and fedora (however I only have one debian server left, used to have a few more). If nothing else, its at least worth tossing the idea around. Best of luck with your project.
Regards,
Steve

Re:Duh...

[Bruce Perens]

The essential reason to choose UserLinux is that it avoids the conflict between Open Source and producing income that all commercial distributions develop while supporting the enterprise. Fully Open-Source distributions won't make a profit over the long term. You have to hold something back like certification or bug reports. And when you do that, the result is something less than Open Source. Rather than give up on fully Open Source distributions, I concluded that we don't have to make money from them. There are enough interested parties to support them as non-profits.

The policy of the UserLinux project is for all development to be carried out within Debian, not within our own repository. Customers can take a much greater role because the Debian organization admits them fairly.

Of course, the long release delay has made something of a fool of me - because so far we've only proven that this non-profit can't get it together to make a release.

There is a lot more in the white paper on the project site.

Bruce

Re:Duh...

[Anonymous Coward]

If you want Stable, you will have to wait. A lot.

I've been using unstable/testing for a long time in my desktop and my server. No problems.

Just grab the latest unstable/testing snapshot and call it whatever you want. How difficult is that?

Re:Duh...

[Anonymous Coward]

And I will start working on UL again when there is a distro to base it upon.

RHEL4 was recently released. Why not base it on one of the RHEL-rebuild projects like CentOS?

(no I am not trolling)

Re:Duh...

[plumpy]

The main thing Ubuntu does is that they seem to focus on providing a stable desktop system. They select packages that they're going to focus on, then package those things REALLY WELL. Like, Hoary (the upcoming Ubuntu release) has GNOME 2.10, (which Debian unstable does not) and it's really nice. They don't try to do everything like Debian does. They pick out which packages they're going to consider "officially supported", and then they make sure those packages are rock solid and make a release every six months.

Of course, since it's basically a fork of debian 'unstable' (with patches flowing both directions), you can also install all those other 'unstable' packages if you change your apt.sources. But the core GNOME desktop system is really slick.

Honestly, I've been using Linux as my primary desktop OS for 8+ years now, so I pretty much know how to make things work, but I'm really impressed by how nice Ubuntu makes things.

Re:Duh...

[Anonymous Coward]

Pfft. 50K miles? I do that in a weekend. You let me know when you've killed for open source. That's dedication.

Please stop being such a leech, Bruch.

Re:Duh...

[chthon]

I have been using Debian since 1999 and I gradually changed from stable to testing to unstable very time after a release.

Since last year, I keep a local mirror of Debian unstable, because I have many systems to update.

The main conclusion I get from using unstable, is that it is not really unstable. The only recurring problem I have had in the past year is that every couple of months the postfix configuration broke upon automatic install, so I had to redo it manually (most of my systems are updated daily through cron).

I am busy investigating selling barebones PC's with a pre-install of Debian, and I must say that the update speed is very large. However, Debian only follows the release cycles of the individual packages.

I think that you need to decide for yourself what features you need, make a snapshot of a local mirror and start testing for feature completeness and to remove bugs.

Updates to this snapshot should be done by manually importing new packages, or by fixing the Debian sources and adding the new package in the snapshot.

This way it should be possible to create Debian based releases in a much shorter timeframe than Debian normally needs.

I think however, if you need to do this, then you need a formal organisation with paid people which can be dedicated to testing and releasing.

The main problem with Debian is that there does not seem to be a dedicated team (paid or voluntary), that has the time and inclination to take this massive snapshot (what, 12000, 14000 packages) and test them all for one release.

Maybe they should only test some subsets which can be released on two or three CD's, release these and provide the rest of the repositories as additional features.

Re:Duh...

[uglyduckling]

I 'grew up' on Debian, but got frustrated 18 months ago when I had to mess around for hours with dependencies - there seemed no point in using Debian stable if I had to use loads of backports and manually install some libraries just to get a working up-to-date web browser and other essentials. I then ended up getting bits from unstable, and my mixed-distribution system regularly broke.

I switched to Mandrake, but really couldn't stand urpmi: it's soooo slooow! Honestly, why does it have to download a multi-megabyte package list? It's a complete dog compared to apt-get.

Then, a couple of months ago, I got a new hard drive for my laptop, and decided to try Ubuntu. As far as I can tell, they do exactly what you describe. There's a list of standard packages needed for a desktop distribution, which are tested and work very well. (The 'main' section). There's then a 'restricted' section, that has packages that cannot be included in the main distribution (e.g. because they don't meet the strict Debian definition of 'free'). Then there's the 'universe' section, that has lots of useful packages that install cleanly, but aren't aggressively tested.

The whole thing works extremely well, has all of the Debian goodness, but with a strict 6 month release cycle.

My philosophy now is: Ubuntu on the workstation, Debian stable on the server.

Re:Duh...

[QuantumG]

That sounds great. The reason I use the present tense is because it sounds almost word for word like the Ubuntu philosophy. Please, do check out Ubuntu. I think you'll like it.

Re:Duh...

[gnalle]

The big seeling point for Ubuntu is that you can have a system that

Has recent packages (Woody doesn't)

Provides security upgrades (Sarge doesn't)

Is somewhat stable (I believe that Warty is stabler than Sid)

Many packages in Sarge are newer than their counterpart in Warty, and similar Sid has newer packages than Hoary. However these differences are small and unimportant.

Ubuntu has focused on a subset of the Debian archive. The packages in this subset are stable and work well. Furthermore Ubuntu has a "universe" archive that contains most of the packages in Sid. Some of the universe packages are uninstallable due to missing files. This can be bad if you are very dependent on a specific program.

Re:Duh...

[nutshell42]

Is somewhat stable (I believe that Warty is stabler than Sid)

Actually the first time I tried Debian, unstable was broken rather often. Only for a few hours in most cases but broken nevertheless. I switched back to Debian when Knoppix came out (point in case: forget Ubuntu; use Knoppix to get a Debian unstable configured and installed in less than an hour =) and I haven't seen a unusably broken package since (In two years? Not sure how long it's been. Did a system reinstall after a hard disk crash). Recently the autofs package's post-install didn't work but the program itself worked just fine.

Re:Duh...

[Donny Smith]

> With Ubuntu, everything is free, and they've made a commitment to always remain free.

The same is with CentOS - http://www.centos.org/ - stable, free, even binary-compatible with a Linux vendor's enterprise edition releases.

>it's developed by a community and has a central package repository.

CentOS is built from "North American enterprise Linux vendor's SRPMs" so in that respect it's less open to experimenting with and it follows the upstream vendor's releases, but for those who like stability, that's a good thing.
Of course, it's RPM based but who cares - it comes with YUM pre-installed so you don't really have to think about it - you update and install similarly to the way you'd do it using apt-get.

If I had time to spend on fooling around with unstable packages and such, I guess I'd try Gentoo... For regular sysadmin folks who don't dislike Red Hat, CentOS is a godsend.

Re:Duh...

[ultrabot]

I've been seeing that a lot of people are recommending Ubuntu lately. What does it offer over a regular Debian Sarge install?

If you have a problem w/ Ubuntu (stable release), it's highly likely that someone can help you - there are thousands of users running the same packages you are. Sarge changes all the time, and Sarge users are less likely to document or help with the problems - why would they go to great lenghts documenting something that might change overnight?

Ubuntu community is also more appealing to newbies, due to lower degrees of elitism.

Re:Duh...

[Ulric]

Usually the "why don't you do it yourself" comment is just annoying, but considering who you're commenting on it's actually a bit amusing.

Re:Duh...

[ultrabot]

UserLinux is that it avoids the conflict between Open Source and producing income that all commercial distributions develop while supporting the enterprise. Fully Open-Source distributions won't make a profit over the long term.

From what I've heard about Ubuntu, they are not planning to become profitable through "evil" routes. There seems to be a genuine wish to help the community and the world at large. Of course that's easier for a millionaire like Shuttleworth, but I tend to believe Ubuntu will develop to be sustainable, if not necessarily highly profitable...

Of course, the long release delay has made something of a fool of me - because so far we've only proven that this non-profit can't get it together to make a release.

Have you considered basing Userlinux on Ubuntu? I.e. create an "independent" service organization based over Ubuntu code? I wouldn't be surprised if Ubuntu welcomed the development, the first phase w/ Ubuntu is essentially about maximizing the number of users.

Re:Duh...

[Anonymous Coward]

I've paid my dues a few times over.

You're not finished paying your dues until you're dead. At least that's the way it works for most folks.

But yes, thanks for all the hard work.

Re:Duh...

[quarkscat]

So that's what happened to UserLinux!

If I could make a recommendation here?
UserLinux should be tied to a linux distribution
that is stable, yet has a 6 month release cycle
that picks up the latest-and-greatest F/OSS
projects.

Slackware is just such a distribution. While
the installation process is a bit archaic in
comparison to some other distributions, it has
never failed to install. It is stable enough
that I am still using Slackware 9.1, but with
the very latest kernel (2.6.11) direct from
"www.kernel.org" (and 2.4.30 when released).

Each release is available on a 3 CD boxed set,
with a "live CD". Slackware's clean design
also makes it the basis for several other linux
distributions. While it is not cross-platform
to the extent that Debian is, Slackware is the
distribution I would hang my hat on if I wanted
to build a release for another architecture.

Just my $00.02 worth.

Re:Duh...

[hawk]

Back when unstable did that, it was only about twice a year.

But when it did, it it left your system dead in the water. And sometimes it was because a single developer had pulled a package upon which many others depended over his own views on license purity.

Eventually, FreeBSD supported all my then-current hardware during one of these, and I switched over.

hawk

Re:Duh...

[billtom]

QG, Bruce knows all about Ubuntu (I hate to play the "do you know who Bruce Perens is?" card, but do you?). In fact, it's a pretty safe bet that he personally knows most of the principle people involved in Ubuntu.

His point about the difference between Ubuntu and UL is that UL operates within Debian, that is, if UL people want to make a change and they can't get it into Debian, then they won't do it.

But Ubuntu takes Debian and then makes their own changes to it. They always try really hard to get those changes back into Debian. But if they can't, then they'll still go ahead if they think it's important.

I'm over-simplifying a bit and I'm not saying which approach is right or wrong, but there is a difference in philosophy and approach between Ubuntu and UL.

Re:Duh...

[jdavidb]

Bruce, I have a lot of respect for you, but your signature has detracted from that. I went to check out Technocrat because of your sig recommending it as a more mature forum ... and found a picture of the Statue of Liberty draped, from moveon.org, with an appeal to get the ad placed on TV. And I thought slashdot was left-biased! Any forum associating itself with moveon.org is not "mature." There is a mature way to make the leftist progressive case, and there is an immature way. moveon.org, and technocrat, are choosing the immature way.

Re:Duh...

[snorklewacker]

Yes, yes ... I'm a junior sysadmin who can make recommendations for new builds. What does UL do for me? My gf is a non-gaming home user. What does UL do for her? It seems to me that UL sounds simply like Debian's commercial arm. If that's so, then yes, I can see how Debian is putting it in the deepfreeze...

Fully Open-Source distributions won't make a profit over the long term. You have to hold something back like certification or bug reports. And when you do that, the result is something less than Open Source.

Certifications seem rather orthogonal to source. In fact, it's the very sort of value-add service that F/OSS advocates from RMS to ESR seem to advocate.

Re:Duh...

[Anonymous Coward]

> I've been seeing that a lot of people are recommending Ubuntu lately. What does it offer over a regular Debian Sarge install?

If you have an amd64, you're supported out of the box with live and install CD's built for it, and amd64 trees at all the official repository sites. No second-tier repository sitting on a lonely server somewhere...

Re:Duh...

[Anonymous Coward]

Bruce, one weakness in the Debian release cycle that has struck me for some time is that a new release waits for every package in the distribution to be brought up to speed.

It's silly for a whole release to be held up so that UnderwaterBasketWeavingPro can be fixed by some (possibly unmotivated) volunteer somewhere.

Seems to me they ought to have a core subset (kernel, libc, bash, X, etc.) that they would release when ready, then bring the other packages on board as the non-essential-package developers get their act together.

In other words, light a fire under their asses by saying not "Pretty please, fix it so we can release", but "We just released, I guess if you want people to actually use your package, you need to update it."

Re:Duh...

[Bruce Perens]

Pfft. 50K miles? I do that in a weekend. You let me know when you've killed for open source. That's dedication.

OK. What's your name and address?

Re:Duh...

[Bruce Perens]

Well, at least you didn't criticize the content.

Obviously we don't agree about politics. The MoveOn ad in question is old, and is there at this late date because I don't want to deal with the slashcode and remove it. It happens I am working on a Ruby on Rails replacement for Slash.

But I would expect to run more MoveOn ads. If I can change a few people's minds in exchange for losing a few other people who just can't bear to look at a MoveOn ad, I'd consider that a good deal.

Bruce

Re:Duh...

[QuantumG]

yeah, actually that was stated in the very first press release from Ubuntu. They indend to make money from support services for Ubuntu and they encourage other companies to do the same.

Debian thoughts

[Anonymous Coward]

As a new Linux user, what I heard from all my friends was, "don't use Debian, use Mepis or Knoppix or Ubuntu." It seems to be the opinion of many that Debian is nice, but it's not worth using a plain version of Debian, because these other distros have built it into something better. At least, that's the impression. So it seems that Debian is losing "mindshare" among new Linux users to a degree.

Re:Debian thoughts

[mabinogi]

Well pretty much the whole point of Debian is to have a distribution that others can take, modify, and re-distribute.
So using Debian derived distributions like Ubuntu or Knoppix is still good for Debian, or at least compatible with its goals.

The fact that it's a pretty good distribution in its own right is more or less just a bonus....

Re:Debian thoughts

[Terrasque]

I personally prefer debian testing/unstable, since I have more control over the system.

But I wouldn't recommend it to a beginner. It's a bit too complex for that, and with distro's like ubuntu why even care?

Ubuntu/mepis/knoppix IS debian, they've just specialized.
Debian is ALL of those, but it's up to you to pick things together to a finished, usable system. And most users have no idea what they want.

Re:Debian thoughts

[smoking2000]

My first Linux distro was Debian Woody when it was still in testing, I learned how to get things running in the Debian way. Which meant alot of tinkering with config files. I'm now quite confortable with that way of using my OS.

I've also setup my mothers laptop with Debian recently, testing flavour, since I refuse to admin Windows in my own network. She can do all the trivial things she wants, browse the net, email her friends and edit some pictures of her kindergartenclass.
One thing she can't do that well is do new things. I always have to setup and configure it for her.

Last weekend I upgraded my Woody install to Unstable and put Ubuntu beside it as dual boot, to see what all the rejoycing was about. And to evuate it as an alternative for my mother to use.

And I must say Ubuntu makes discovering the OS alot easier, since some very handy packages are installed by default, where under plain Debian you whould have to know about them and install them yourself.

So, I agree that for new users plain Debian is not the way to go, unless they are willing to dive deep into the workings of your OS. If you just want to use the OS, instead of tinkering with it, the Debian derived distro's are a fine choice.

Re:Debian thoughts

[Anonymous Coward]

> Well pretty much the whole point of Debian is to have a distribution that others can take, modify, and re-distribute

Also Debian is unrivalled for rationalization.

Re:Debian thoughts

[smchris]

Yes. For me, current Debian releases _are_ Knoppix.

First, there is the legendary Debian installation process that only a linux-from-scratch person would call a wimpy newbie GUI. You end up feeling like a pigeon pecking for an eventual pellet pressing "enter" that many times.

Second, Knoppix hardware recognition is great and knx2hd is an incredibly easy hard drive install in contrast to vanilla Debian, one can choose "old fashioned" Debian, and the base desktop apps you want are already there when you reboot.

Third, there are enough articles out now on the Knoppix remastering process that the entrance bar to that has been greatly lowered. Myself, I've worked up a PostgreSQL server ISO that includes apache, php, perl, ODBC, PgAdmin III, Webmin and desktop icons to the PostgreSQL manual and a debian manual, start/stop database, view log, and the ODBC GUI manager. It's great. I've quit doing bare metal backups.

Paradoxically, these developments may _validate_ Debian caution. All this creative third-party repackaging has a heavy dependency upon the stable foundation Debian provides, right? It's a good thing.

Have to compete with Microsoft

[mr_tap]

Last stable release in 2002 - how can they possibly compete with Microsoft whose last desktop operating system release was in 2001 :)

Re:Have to compete with Microsoft

[Anonymous Coward]

Then again, when was Microsoft's last stable operating system release. :)

Re:Have to compete with Microsoft

[evilviper]

Microsoft only provides the operating system. A Linux distro, OTOH, is expected to provide just about every program that you might ever want to use.

A version of Windows from 2001 isn't a problem, but it would be if it couldn't run more recent programs.

Re:Have to compete with Microsoft

[arduous]

Well, first of all, I'd call Windows XP SP2 their latest release.

Secondly, Windows XP is just the a basic operating system. Debian 3.0 has 8710 packages bundled with it, and all of those packages are now almost 2 years old.

Running a 2002 release of Windows XP doesn't prevent you from installing the lastest version of Mozilla, Firefox or . The version of Mozilla in Debian stable is currently 1.0.0, and Firefox isn't even there!

I've been running debian servers for the last 5 years, but lately I have been seriously looking for an alternative that has a faster release cycle.

Re:Have to compete with Microsoft

[krunk4ever]

That's only if you consider XP, XP SP1, XP SP2 to be one release. If you look at the version #s for these 3 releases, they're actually different.

Re:Have to compete with Microsoft

[xenocide2]

Amusingly, the last Debian stable branch has gnome 1.4 by default. That was the gnome that was torn to shreds by the Sun Usability report as being confusing, pointless and also pretty damned ugly. Testing uses 2.8, which is starting to approach the usability of XP, four years later. Way to give M$ a run for the money there... ;)

Re:Have to compete with Microsoft

[uofitorn]

I understand where you're going with that and all, but, there's nothing stopping you from installing the lastest version of Mozilla, Firefox or . after you install Debian either.

Re:Have to compete with Microsoft

[transiit]

Hi.

I still miss the days of Gnome 1.x before the Sun usability studies made it completely unusable, and I see nothing special about XP as far as usability goes. XP's just got familiarity with such a large crowd.

Remember kids, options are bad, they only anger up the blood.

-transiit

Re:Have to compete with Microsoft

[Terrasque]

And how often do you run mozilla/firefox on a SERVER?

If it's a desktop, you should probably be running sarge/testing anyway.

Re:Have to compete with Microsoft

[tacocat]

I still use Windows 2000 at work.

So what does a stable release of 2002 really mean? Not much. At least it works.

My reason for sticking with Debian was because of their sane release cycles. It is based on 'Release when Ready' not because of some calendar hype. This is one of the core values of Debian. Please don't fuck it up unless you want to look and smell just like any other distro out there.

I guess the real questions should be along the lines of, what currently prevents debian from releasing faster and what can be done to address those specifics? I think the architecture is the only consistent tangible I've heard of.

Re:Have to compete with Microsoft

[natrius]

I've been running debian servers for the last 5 years, but lately I have been seriously looking for an alternative that has a faster release cycle.

Contrary to popular belief, Ubuntu is suitable for server use as well. All the Debian packages are there, just with newer versions. There are two options with Ubuntu currently. You can use 4.10, which is the current version, and you won't have to update for a year, which is when the support (read: security updates and paid support from Canonical) for it ends. Everything in there has been in wide use for six months, so it's pretty solid. The other option is to wait until April 6, which is when version 5.4 will be released. You'll have a full 18 months of security updates, and everything you like about Debian is still there. Check out the FAQ page on server installs.

Re:Have to compete with Microsoft

[droopycom]

"Release When Ready" is a good principle

"But Release When 8000+ packages ready" is next to impossible and rather dumb...

Theres probably something wrong with their paradigm(*), I guess they could also release packages individually or in groups,they kind of do that with testing in fact ...

Stable is supposedly for Critical Entreprise application, but who in this category needs 8000+ packages, including n minesweeper and x IM client?

Maybe the solution is less packages in Stable, just keep the most important component and apps in there. ... Well off course then they will have to decide what make it through and what not, which off course will lead to endless debates, if those debates are not eclipsed by debates on what is free and what is not and if non-free should be kept etc...

--
(*) Off course not i dont really know what that word means...

Re:Have to compete with Microsoft

[eraserewind]

A Linux distro, OTOH, is expected to provide just about every program that you might ever want to use.

And their big mistake is believing that. They should be aiming to produce an integrated system that works well with itself, not a cd full of random software.

Re:Have to compete with Microsoft

[Berzelius]

You may find Sun had made Gnome total unusable. I don't agree with you. The integration and simplicity of the user interface is beautiful - perhaps even on par with Apple.

Re:Have to compete with Microsoft

[transiit]

Apple's interface isn't particulary simplistic or well integrated, by my experience.

Closing windows but leaving the programs running in the background. If the window is not on my screen, and I didn't explicitly say "Go hide somewhere until I fetch you again", it just asking for a bigger mess.

The one-button mouse issue is still an issue. Sure, many immediately replace the abomination, but when the sentiment of "Oh, more than one mouse button is confusing to new users! Ctrl-Option-click is far more intuitive!" still reigns supreme, their interface will continue to be lousy.

They ain't selling easy. They ain't selling "It just works.", they're still selling colored plastic and eye candy.

As a user that never uses drag and drop, had used several mp3 applications before iTunes hit the market and called itself the first, several calendaring apps before iCal, plenty of photo organization/image manipulation stuff before iPhoto, video editing before iMovie, I could give a rat's ass about any of it. None of it is new, they just keep getting away with pretending it is because everyone's hypnotized by their "Genie Effect" window animations.

Don't get me wrong, I did once like Gnome. But when Sun came along and brainwashed Havoc into saying "Everything different is bad.", it lost a lot of appeal to me. Maybe I like having multiple clock applets to choose from, rather than being scared and confused by a selection.

-transiit

Re:Have to compete with Microsoft

[xenocide2]

Did you like the retarded menus with clickable buttons that do NOTHING, and superfluous headings for submenus? I didn't. I didn't enjoy the button styles that were last seen in windows 3.1 either.

I don't think the Sun team was overly effective in their discussion, either. To take the clock example, multiple clock applications and applets is somewhat confusing. Instead, it would have been neat to provide a single clock, which you could put on gnome-panel by dragging and dropping. Part of the problem was that GNOME didn't have a consistant method or vision of doing things with it. Part of the remaining problem is that Sun didn't either. GNOME is neat, but not that neat right now. 2.10 adds so little to the equation. Options can be a good thing, but it only makes sense to figure out a way to make those options accessible in a consistant manner.

Re:Have to compete with Microsoft

[transiit]

I think my biggest complaint from when they switched from the 1.x series to the 2.x series was when I discovered that many of the things that I liked were pulled out with the explanation "Oh, it was too confusing. So we made it easier."

And so I kept banging my head against the blessed defaults, thinking "But I don't like it this way. I used to be able to change it. Why is taking away flexibility considered an improvement?"

Of course, things like gconf don't help either. Making configuration more opaque doesn't make my life easier. Editing name,value pairs in a text editor isn't that bad. A program that uses that sort of configuration format should be nice enough to give you an annotated sample config to start with, and belch out something on stderr if you munge an entry in an unworkable way. The "XMLize-everything!" camp seems to be fretting over a lack of consistency, but it takes away so much. For example,

<gconf>
<entry name="show_hidden_files" mtime="1111271813" type="bool" value="true">
</entry>
</gconf>

As opposed to:

show_hidden_files=1

So what happens when you go off and edit either by hand? Oh wait, we're not supposed to do that. We're supposed to use gconf-editor. Which immediately warns us not to. I understand the comparisons people make with the Windows registry.

I could go on at length as to why dumping all these settings into a big file hierarchy (i.e., ~/.gconf/desktop/gnome/file_views/), with all config files named %gconf.xml, is a bad idea, but that's just implementation boneheadedness.

I want to know why the GNOME team would rather irritate some of their existing users in the name of getting new users, on the advice of the people that have been pushing CDE as the one true UI for years.

As I use GNOME less, it strikes me as a dumping ground for flavor-of-the-month technology.

-transiit

I had to wait for a new stable release

[Anonymous Coward]

to get this FP

i try to release once or twice a day

[Anonymous Coward]

we are talking about pooping, right?

I wonder

[LittleLebowskiUrbanA]

when the new Debian installer comes out, will that kill Libranet? I sure as hell hope not. Libranet is one of the unknown gems of Linux distros. If you can ween yourself of the need for the latest and greatest buggy software releases, Libranet *just works*.
Also, this posting should have been labeled from-the-no-shit dept.

Re:I wonder

[LittleLebowskiUrbanA]

Overrated? Mod was on crack.

Yeah...

[TWX]

It would be really nice if Stable were updated at least yearly. I'm willing to play with Unstable or Testing if it's for my own use only, but if it's for someone else then I may as well either use a heavily-package-based distro like RedHat or SuSE, or Slackware if I'm going to have to build a bunch by hand anyway.

I guess that it'd been awhile since I last installed Debian from scratch, I didn't know that it has been two years.

Re:Yeah...

[hackstraw]

It would be really nice if Stable were updated at least yearly.

Which part of stable is confusing?

You know when I update my server software? When it is so old its not really supported any more, and when I get new hardware, and then it depends.

I like the longer lifespan and dependability of any operating system. If it ain't broke, why change a thing? It was different in the 1.2 and 2.0 days of Linux, when the kernel was not very mature and lacked very important features or had hard coded limitations in it, but those days are gone. The last limitation that has been overcome is large block device support (if you consider 1Tb large). I remember having to use beta kernels, 2.1.125 if I remember correctly because it supported more than 256 file descriptors for a single process, and the production kernels did not without a patch. There was a time when software RAID was not offered in the kernel. Or there were less than 5 or 6 filesystems offered. But now the kernel comes with everything besides a kitchen sink.

Also, the beauty of *NIX systems is the /usr/local (or /opt, I prefer /usr/local) directory to put updated or additional packages that did not come with the installation. For example, on my solaris boxes and even my OSX box, I put things like the latest GNU fileutils and other things like GNU grep, because they are simply more robust featurewise than standard implementations or even older ones. I also tend to roll my own core software on a system depending on its purpose (httpd for a web server, squid for a proxy server, DB server, etc). That way I am free of the distributions mucking with my core services, and often I want additional extensions or whatever that I would just rather do myself. Things like mail and bind, I leave to them to take care of for updates or whatever, I trust them more than myself, maybe that is a bad assumption, but...

Now, these things are different for an end user's desktop machine. But the reason for that is because the GUI world of Linux is more in the growth stage like the kernel was in the mid to late 90s. It is worthy to have updates every six months or less, and it is OK to have the downtime of having your video drivers bomb out on you or monkey around for a while to get the box back up and running. Its not worth the downtime on a server to do this.

Just my 2 cents.

Re:Yeah...

[drinkypoo]

You know when I update my server software? When it is so old its not really supported any more, and when I get new hardware, and then it depends.

Personally, I like to update when there is a security hole, so I don't get rooted. Granted debian is fairly good about security patches...

Re:Yeah...

[hackstraw]

Personally, I like to update when there is a security hole, so I don't get rooted. Granted debian is fairly good about security patches...

Duh. I meant upgrade, not update.

Re:Yeah...

[TWX]

Which part of stable is confusing?

Nothing was confusing. I don't know where you're deciding that from...

I like the longer lifespan and dependability of any operating system. If it ain't broke, why change a thing?

Because people building new servers to have new duties frequently want things that aren't part of stable. They end up having to go to unstable or third party package sources if they aren't willing to compile by hand. Since Debian is designed to update smoothly and in theory is very thoroughly tested before being declared stable, there shouldn't be any major operability or functionality issues at all if someone upgrades from an older stable to a newer one.

Besides, if you're running a critically important server, you should have install media beyond the Internet for your servers. If you want to remain with a frozen version then you have CDs to install from, or your own Stable mirror. Let them integrate new features so that the rest of us can have them.

this just in...

[SuperBanana]

Debian Leaders: We Need to Release More Often

This just in: the Catholic Church says the Earth is round.

In other news, George Broussard admits Duke Nukem Forever "is a little late".

Question- why did it take, oh, 3 years for them to finally come to terms with the fact that their iguana was turning into a dinosaur? It's like they've all been collectively in denial. I took one look at the list of versions in the stable branch when someone suggested I check out Debian. I laughed, and closed the window. Every time I've come across a Debian box, it was "put in by some weird guy who doesn't work here anymore". Debian users preach to me about stability, when I haven't had a linux box do something unexpected in quite some time. Debian's still stuck in the age of obsession with uptimes.

I understand the need for stability, but that means you put more effort into QA, not that you sit on your ass because what you've got works. I mean hell, some distros still ship 2.4; it's an embarrassment that companies like Redhat port BACK improvements made in 2.6 to their own versions of the 2.4 kernel, instead of finding and fixing problems in 2.6.

Re:this just in...

[Anonymous Coward]

Every time I've come across a Debian box, it was "put in by some weird guy who doesn't work here anymore"

Hey, who you calling weird?

Re:this just in...

[dondelelcaro]

Question- why did it take, oh, 3 years for them to finally come to terms with the fact that their iguana was turning into a dinosaur? It's like they've all been collectively in denial.

We've not been happy with the time that it takes to release for AGES now. Potato took too long, woody took longer, and sarge is taking it's own time. The symptoms are known, and much lamented. However, the fix for the underlying problems is far less trivial, and so far no one who is actually capable of doing the work has come forward and done whatever needs doing to fix the actual problem (whatever the hell the actual problem actually is.)

I understand the need for stability, but that means you put more effort into QA, not that you sit on your ass because what you've got works.

Perhaps you've been sleeping through the 300,000 bugs that have been filed on packages in Debian, many of which have been fixed? Or maybe it's just that you don't really understand the amount of work that it takes to actually release a stable distribution without RC bugs on all of the architectures that Debian supports?

Re:this just in...

[Anonymous Coward]

> whatever the hell the actual problem actually is.

The "actual problem" is Debian's ridiclous release criteria -- that is 100% clear to everyone.

Face it, trying to stablize the exact same set of 2000 packages across 11 architectures is valiant but foolhardy. The solution is obvious -- reduce the number of packages and number of archs. The politics are the real problem.

Re:this just in...

[SuperBanana]

We've not been happy with the time that it takes to release for AGES now. Potato took too long, woody took longer, and sarge is taking it's own time. The symptoms are known, and much lamented.

Okay. So, again, why did it take three releases to realize something was wrong? If the symptoms were known, why didn't people just start fixing them? Politics? Funny thing about politics. Even if the politics aren't in your favor, if your intentions are honest, you're stepping up to the plate when no one else is- guess what, it's really hard for others to argue against you without looking petty or controlling.

However, the fix for the underlying problems is far less trivial, and so far no one who is actually capable of doing the work has come forward and done whatever needs doing to fix the actual problem (whatever the hell the actual problem actually is.)

So, basically- you and other Debian people have thrown up your hands and said, "augh, look at this mess, it's huge, complex! We can't possibly fix this mess! Let's wait for someone else to come along and fix our problem."

There's a website for people trying to get their lives back together and pick up the piles of junk lying around the rooms of their house. The common theme is- DON'T try to take it all on at ONCE. Don't sit there and assume you will eventually come up with the most elegant solution to your problems or some genius will drop out of the sky. That's like shooting only for a royal flush in poker.

Divers have a similar mantra, especially ones who do technical diving; nitrogen narcosis exaggerates emotions and a minor problem turns divers into a panic. The mantra- "as long as you're breathing, you're OK". Stop. Relax. Solve one problem at a time (incidentally, the other mantra is not to let problems pile up, because they compound each other; fix things as soon as you notice them...but it's a little late now). Tomorrow, if you see or remember a problem, just solve it. If anything, others might be inspired or encouraged by the activity.

You obviously have a lot of talented people. Get everyone to sit down, make a list of problems. Categorize them. Divide them up and hand them out or post them up on a page. Don't make committees- committees are great at wasting time. When you're behind the eight ball, you don't need a group of people to decide which way is the best direction to move- you've just gotta MOVE. If someone doesn't like what you've done- well, they had three years to do it their way, so tough!

Or maybe it's just that you don't really understand the amount of work that it takes to actually release a stable distribution without RC bugs on all of the architectures that Debian supports?

Why is it that people in a hole always tell others how they couldn't "understand how much work" is it? If it's so hard to make a useful distribution, why did we see a veritable explosion of distributions (some of them based off Debian) in the time Debian hasn't released a single stable version? If they're people who jumped ship, why did they jump ship?

If "too many platforms" is the problem, do what GCC did- stop overcommitting. The GCC team stopped wasting time on a couple architectures nobody was using or helping them maintain but for which they'd have to fret over whether changes would break this or that. Funny thing- nobody's really complained that loudly. If they care enough about that architecture, they either step up to the plate, help recruit people to help.

That advice goes equally for platforms as it does for packages. I remember debian used to be over a half dozen CDs with something like 3600 packages. Focus on core packages; if need be, get people to vote for stuff they want. If something's not ready and nobody could help, fine- it doesn't go in, it doesn't hold up the stable release. If people needed that package, or whoever makes that software gets miffed it was left out, they know they have to help or it won't make the NEXT release either. If nobody notices or cares that package didn't make it- fantastic!

Re:this just in...

[xenocide2]

Why Debian has failed to release:
1. Many developers lack the proper motivation, having used unstable/testing for so long that a new stream called experimental is now used to test software that might actually render your computer partially unusable.

2. The release has gotten so stale that there are about 3 guys left recommending that users stay on stable instead of testing or unstable. Most of the current users have little reason to care if a new stable is made or not.

3. Upsteam projects prefer a mantra of release early, release often, which isn't nessecarily a bad thing, but it gives a very narrow window with which to work with developers to iron out "Release Critical Bugs." GNOME, in particular, has seen several iterations roll through testing.

4. The sheer scale of what Debian provides users in the form of a stable release. Debian packages some ungodly amount of software; its part of what makes Debian damned nifty. And its hard to tell people to stop packaging more when it the costs are a few megabytes on some ftp mirrors and the occasional download. But every package provided is a source of potential stable release bugs. Usually only very common packages wind up being labeled "release critical," such as the GNOME package mentioned above.

Re:this just in...

[dondelelcaro]

Face it, trying to stablize the exact same set of 2000 packages across 11 architectures is valiant but foolhardy. The solution is obvious -- reduce the number of packages and number of archs.

Surprisingly, this actually hasn't been a major blocker for quite some time. If any of the superfluous packages can't get their act together to be in a releaseable state, they are summarily removed from testing.

The actual blocker for the past 6 months or so has been the testing-security support. Before that, it was the fact that we didn't have a working installer.

Re:this just in...

[Anonymous Coward]

so far no one who is actually capable of doing the work has come forward and done whatever needs doing to fix the actual problem (whatever the hell the actual problem actually is.)

Therein lies the problem. The people who know what needs to be done to release refuse to share that information and accept help. Too many key people act like bureaucrats interested in maintaining their positions rather than developers interested in producing a usable distro.

Re:this just in...

[dondelelcaro]

Okay. So, again, why did it take three releases to realize something was wrong?

It didn't.

After potato was released, Anthony Towns implemented testing in an attempt to keep testing in a releaseable state always, so releases could occur more rapidly. That helped, but still didn't really fix the problem.

After woody was released, security support and the installer were serious problems that had stalled the release of woody for quite some time, so more effort was placed into those areas to create a working installer along with a decent security infrastructure. That has helped as well. However, it took quite a while for those to be implemented.

Now that sarge is on the verge of being released, people are analyzing the situation again to try to figure out what else should be done to fix the problem. The Vancouver Prospectus is an attempt to solve what have been identified as the problems for etch.

you and other Debian people have thrown up your hands and said, "augh, look at this mess, it's huge, complex! We can't possibly fix this mess!

No, as you can see above, specific things have been attempted to solve the problem. They haven't succeeded, clearly, but it's not for lack of trying them.

If it's so hard to make a useful distribution, why did we see a veritable explosion of distributions (some of them based off Debian) in the time Debian hasn't released a single stable version?

Distributions based on Debian are rather easy to make, frankly, especially if you're going to standardize on a specific set of packages and only support them. It helps as well if you can throw money at the problem and hire people to work on specific problems. Point in fact, none of the not-for-profit Debian based distributions have every actually released a stable distribution and suported the entire stable distribution for a whole product life cycle. They have different goals for the releases that they make than Debian does, which is quite acceptable for them. [Nothing is stoping anyone from taking a specific version of testing, calling it "stable" and supporting it. The fact that no one has should tell you something.]

Re:this just in...

[Curl E]

Divers have a similar mantra, especially ones who do technical diving; nitrogen narcosis exaggerates emotions and a minor problem turns divers into a panic. The mantra- "as long as you're breathing, you're OK". Stop. Relax. Solve one problem at a time (incidentally, the other mantra is not to let problems pile up, because they compound each other; fix things as soon as you notice them...but it's a little late now). Tomorrow, if you see or remember a problem, just solve it. If anything, others might be inspired or encouraged by the activity.

You obviously have a lot of talented people. Get everyone to sit down, make a list of problems. Categorize them. Divide them up and hand them out or post them up on a page. Don't make committees- committees are great at wasting time. When you're behind the eight ball, you don't need a group of people to decide which way is the best direction to move- you've just gotta MOVE.

This might be a good time to plug Joel on Software's essay Fire and Motion.

Re:this just in...

[illumin8]

dondelelcaro wrote: Or maybe it's just that you don't really understand the amount of work that it takes to actually release a stable distribution without RC bugs on all of the architectures that Debian supports?

Agreed. They definitely are supporting too many architectures at this point.

From TFA: If the proposal ends up being approved by the Debian community, Debian will end up supporting only the i386, PowerPC, IA64 and AMD64 architectures and dropping several architectures such as Sun Microsystem's SPARC and IBM's S/390.

See, this is where they are going wrong. Why on earth would you drop support for Sparc and S/390, two of the most popular architectures in the enterprise market, and keep support for the bastard stepchild that is IA64? It seems like this move to drop architectures is mostly political and probably because Debian receives so much funding from HP. Don't leave two of your best architectures out in the cold.

Re:this just in...

[Anonymous Coward]

Stop supporting the less popular architectures.

Re:If it's stable, it doesn't need to be updatedOf

[Storlek]

This is one of the reasons I'm not using Debian now. It might be stable, have a brilliant program that handles all the installation stuff automagically, and have a great community, but the big problem with it that turned me away is exactly that mindset. The last time I had the inclination to try out something different, I was looking for a non-commercial distro that had recent versions of Gnome and KDE and decent non-annoying package support. Debian had two out of three, but if I got it, it would have been mostly a downgrade for me.

Another really important advantage of releasing more often is that releases attract attention. A new version of something is often enough to get people to try it out, and it could turn out to be very good for Debian. Plus, that's the general mentality anyway -- "release early and often" -- of open-source, and Debian is perhaps the most adherent of the well-known Linux distros to the whole open-source philosophy.

If Debian starts releasing a new version every couple months, I'll be sure to give it a try, and I would imagine many other people feel the same way.

Not a huge deal

[BAILOPAN]

Ultimately, the people who like Debian will continue to use it; likewise Debian's goal should be keeping its customers satisfied rather than trying to sway people away from other distros.

I don't really care that it's not updated because apt is flexible enough to work around that. And if a package is _insanely outdated, usually a newer one is in Testing or Unstable. And as a last resource, it's not like Debian precludes you from compiling it myself.

While more frequent releases would be nice, I like it just the way it is. I feel as if I'm guaranteed that the packages will work together without problems (something I haven't encountered in certain other package management systems). And for the select few programs where the version is unacceptably old (like gaim), I just compile from source code.

Re:Not a huge deal

[Chealer]

"apt is flexible enough to work around that." Really? The possibility of mixing software from several releases is interesting, but you make it sound better than what it is really. Sarge/Sid mixes are really useful, Woody/* mixes bring dependency hell. Of course Gaim is the perfect example of software so outdated that it's broken, but there's more. php4 and 2.4 kernel source have unfixed known holes since monthes. They're just too out of date for anybody to care. If you can't get a secure LAMP under Debian (stable), why use it?

Re:Not a huge deal

[kfg]

"Debian's goal should be keeping its customers satisfied. . ."

Debian doesn't have customers. It has users.

KFG

Re:Not a huge deal

[BAILOPAN]

Yes, sorry, that was a poor word choice on my part :)

Re:Not a huge deal

[kfg]

Just one of my buttons that has been overpushed of late.

KFG

Too much pr0n...

[Anonymous Coward]

...for a second, I thought that read "Lesbian Leaders".

And I, for one....

ahhh, never mind.

Re:Too much pr0n...

[Anonymous Coward]

You're not alone. For just a moment, I thought I saw lesbian cheerleaders.

Not that there's anything wrong with that.

Good news, I think

[Anonymous Coward]

I think this is good news that some of the potential leadership in Debian has reconized this as a problem.

I've been a Debian fan for some time, but I find I am racking my newly built critical servers on RHEL3&4 just because so many of the Debian packages are 'stale'. In a lot of enviroments, running testing is unacceptable and using stable is to far out of date for the intended use of the machine. We are definatly in limbo as far as Debian installs.

I really hope they pull this together, without Debian the landscape changes dramatically for binary stable systems.

But, the biggest problem I can see is that by releasing early and often it creates a larger legacy code base that needs to be maintained but does not have the resources to do so. You cannot effectly update a server farm of hundreds to thousands of machines to a new version within a short legacy cycle, yet it is a huge burden to maintain the legacy code for any lengh of time.

Re:Good news, I think

[runderwo]

RHEL4 I'll give you, but RHEL3? You must be joking. It's at least as stale as Debian stable, if not moreso - plus you get the fun of compiling tons of crap from source because the base distro + 3rd party apt sources come nowhere close to Debian's breadth of packages. I have to admin a RHEL3 box unfortunately, so maybe the situation has improved with RHEL4.

Re:Good news, I think

[Anonymous Coward]

Just compairing PHP4,
Debian stable 4.1.2
RHEL3 4.3.2 IIRC

There are a lot of fuctions not present before 4.3, meaning that you have to go outside of stable to run a lot of current PHP based software.

This is just one example. Sorry if I sound rude, but wtf are you smoking?

BTW, RHEL4 is working very nicely, but I also build and maintain our own 3rd party and custom RPM's for a few hundred of these machines so my view may be different than yours.

Re:Good news, I think

[runderwo]

I don't use PHP, but just about every package I've installed from Red Hat has been a 2002 version of whatever it is, which is comparable to Debian stable. Only the 3rd party sources bother with maintaining up to date software, but they don't have the selection I've come to enjoy. I think the problem is that the RHEL distributions get largely ignored by the 3rd party packagers in favor of the RH desktop distributions (incl. Fedora Core).

Even Slackware....

[bmo]

Is up to date, even considering the head honcho's health problems.

There's no excuse for no Debian stable releases since 2002.

Maybe Bruce should base UserLinux on that.

--
BMO

Re:Even Slackware....

[Mr. Sketch]

True, for the most part. However, even the latest version of Slackware only installs the 2.4 kernel by default.

Re:Even Slackware....

[PedanticSpellingTrol]

Ahhh, Slackware.

Because package management is for sissies.

<3

Re:Even Slackware....

[Hellad]

Right, but the key there is "by default". You can easilly select and install other kernels if you feel like it. The whole idea of Slack is to be using the latest, official stable packages. You can go off on your own and try whatever beta/latest greatest packages you want, but they have always been conservative on kernel issues...

Re:Even Slackware....

[glomph]

Umm, you sound like one of those 'sendmail sucks!' dorks. Slackware has had excellent package management (safe simple upgrades, automatically) since the 8.1 release, some years ago. Keep repeating the groupthink cliches, the Borg loves zombies who ignore facts.

Re:Even Slackware....

[sk8_slacker]

My last Debian installation was Woody, but because of the outdated packages i switched to Slackware. Slackware is a "one man" distribution and manages to be pretty much up to date, so wath's wrong with the Debian guys?!

Slackware rules!

Re:Even Slackware....

[Ulric]

Yes, "upgradepkg *.tgz" is a really, really hard way to upgrade the whole system.

There's nothing wrong with the Slackware package management. It doesn't have dependencies; that is by design. Otherwise it's not too different from anything else, except Gentoo.

Re:Even Slackware....

[sk8_slacker]

I really like the Slackware pkgtools. But there are also other tools like 'swaret' wich ist pretty handy.

Not to mention...

[jrushton]

Not to mention Gentoo.

But I'll wisely keep quiet so not to incur the wrath of Slashdot...

Re:Not to mention...

[Anonymous Coward]

>Not to mention Gentoo. >But I'll wisely keep quiet so not to incur the >wrath of Slashdot... insightful for keeping quiet and not incuring the wrath of slashdot?

Re:If it's stable, it doesn't need to be updatedOf

[Red Alastor]

Try Ubuntu. They have a release cycle of 6 month and the next release due to april is Gnome / KDE. You can even get the preview release now.

The bleeding obvious

[Doyle]

In other news, Microsoft decides that "We Need to Make IE More Secure!".

Good to see the penny dropping twice in one week. ;)

Debian appears....

[zogger]

...just looking at it, to be more of a "base platform" from which people build their own customised distros. This in fact might be an actual model for a future LinuxOS,(OSes in general I mean really) if no standard GNU/LinuxOS ever evolves, just make it incredibly easy to select what sort of computing experience you want, mash a few buttons, answer a few questions about hardware, whatever and etc, and your custom distro gets created, you then download it burn it and install it. People don't really "run" an OS, they want to "run" some applications. They want to just go do stuff with their computer, not really futz with it constantly. Well, I mean the 99% of the other people on the planet. You know, "them" guys.

Anyway, if you look at it that way, it's neither way behind the times or bleeding edge, it's just a big ole pile of apps and kernels that you have access to. Maybe they should just skip the different versions, let Apt sort it out when people go to build their own, make it a remasters dream system instead of trying to be a stock classic distro "OS". Do something different than what MS and Apple and Sun are doing. Make the personalised "your computer" be the primary focus, along with the "easy" part.

Re:Debian appears....

[Bishop]

The problem is that Debian does not want to be a base platform. Their goal is to create an OS for Everyone.

Not sure about more stable releases

[Trogre]

but at least give us a distro that we can use for almost-but-not-quite mission critical applications, like web servers for small businesses, or cyber cafe machines.

There is one very easy way the Debian team could achieve this: merge security patches into Testing at the same time as Stable and Unstable.

Why would this be a good idea? I can't be bothered re-iterating, so here's a paste from a prior post:

Stable? Sadly, not an option due to its complete lack of support for modern hardware or moderm features. It's a marvelous example of what computing should have been in 1997.

Unstable? Far too likely to break at the next apt-get upgrade.

Experimental? Same as Unstable, but worse.

Testing? Probably the best bet, though still not recommended for production use by Debian.org since it doesn't get timely security updates.

Re:Not sure about more stable releases

[Anonymous Coward]

"at least give us a distro that we can use for almost-but-not-quite mission critical applications, like web servers for small businesses, or cyber cafe machines."

Are you saying you can't run Apache with Woody? What does a small business web server need that Debian stable can't provide? I have happy customers for whom I've literally taken P-IIs from storage auctions and turned into LAMP boxes running Mambo CRM systems. Shine up the case, charge only for labor and a service contract and I'm on my merry way. An hour or two more and they have an IMAP mail system with Spamassassin and a Squirrelmail web interface.

And for the CyberCafe machines in most cases people are just expecting a browser so they can check their email and their favorite news site. Even if the version of Konqueror, Mozilla or whatever in Debian stable isn't recent it doesn't break on their aol.com email.

"Stable? Sadly, not an option due to its complete lack of support for modern hardware or moderm features. It's a marvelous example of what computing should have been in 1997."
Modern hardware? For the purposes named above? Like what.... a network card?

As this argument reincarnates again and again I can only hope the Debian team remembers their original rational and stick to their guns, so I can continue to enjoy what for me and my customers is a perfect fit.

Re:Not sure about more stable releases

[Trogre]

The problem begins when you try to run that LAMP system on new hardware - or try to use semi-new features in PHP or MySQL (PHP's gd library comes to mind, but there's a ton of other examples).

Without serious modifications, Woody is not an option for these situations.

Re:Reality check folks

[Anonymous Coward]

Mod parent, self, and grandchildren down, -1, Offtopic!

I most certainly agree!

[MarcoAtWork]

I have used many different distributions (started with one of the very first Slackwares in 2000-2001, stayed with them for a while then moved on to a now defunct German distro, then another defunct one, then Debian, then RH7.x, RH9 and now finally FC3) and the glacial pace of Debian development was what caused me to switch to RH.

Most of the things I needed were in unstable (at the time it was potato I think), and unstable was breaking various odds and ends on a weekly basis and I didn't trust it at all so I ended up jumping ship onto RedHat and have never really looked back.

It kind of sucks, as I did like the idea of using Debian, but when it started getting in the way of me being able to work (aka, things I needed were not available for it) I really had no choice but to stop using it.

Every time I brought the issue up with a debian person I'm always told that 'everybody uses unstable anyways so I should do it as well' which is kind of weird, as when most of your user base is using your 'break at will' branch vs your 'stable' one it should be obvious there is a problem.

Good luck to the debian folks anyways, although I have the feeling it's a bit late to be closing the barn door...

Re:I most certainly agree!

[JonathanX]

You do realize that Slackware has been around since 1993 right?

Stable != Cutting Edge

[akabigbro]

You people have to realize that having a server run forever without a hickup is the key to Stable...Not the latest and greatest. If you want the latest releases, use the Testing or Unstable versions. How many updates does it take Redhat;Suse;Slackware;etc. to be entirely stable?

Re:Stable != Cutting Edge

[Glytch]

And you have to realize that more people run desktop machines than servers.

Why?

[QuantumG]

Why does debian-stable even have to exist? Let Ubuntu and the other distributions based on debian do your stablising.

Re:Why?

[xutopia]

that actually makes sense. Debian sets out to be everything at once atm. But it would make sense to be the best base for specialized distributions by being "unstable", a sort of compendium of things which can/should be on linux. Then other distros can choose to snapshot, bugfix/add/repair and distribute to answer a specific "market" need. If you look at Ubuntu, it's what it did and this is the reason it is so popular. It used a great basis and built upon it.

no shit, einstien!

[RelliK]

Debian developers basically have two options: either reign in the development cycle or rename "Debian Stable" to "Debian Obsolete". I've been a long-time Debian user, but now I too am looking for greener pastures. The question is where to? Gentoo? Fedora? Is there something that compares to apt-get?

Re:no shit, einstien!

[leereyno]

Nothing compares to apt-get, and that is the biggest shame of all.

Lee

Re:no shit, einstien!

[neypo]

Depends. If you are an i686 machine, I highly recommend ArchLinux. Its an i686 optmized distro, its package manager pacman blows everything I've used away.

Re:no shit, einstien!

[Anonymous Coward]

Arch is pretty good... I have been using Arch for about 6 months now, speed is good, can't remember having any stability problems of signifigance and has a good release cycle.

And pacman (arch package manager) is in my opinion just as good (if not better) than aptget..

though the base install does come pretty sparse; I like it like that but others might not.

Re:no shit, einstien!

[Joe Tie.]

I'd say there's a number of programs that compare well to apt-get, such as urpmi with mandrake. The problem is that, at least in my opinion, none of the software repositories are on the same level as debian unstables. It's the only linux distro where I've never found myself having to sit around compiling something or other.

Re:no shit, einstien!

[neypo]

Sorry http://archlinux.org/ - Ctrl + C acted up.

Re:no shit, einstien!

[wyldeone]

Definitely Gentoo. A pain to install, but emerge is even nicer than apt-get. Just 'emerge packagename' and the package is downloaded and installed.

Re:no shit, einstien!

[daliman]

Skip gentoo. I've played with it for about 6 months now, and have come to the conclusion that its unstable is just too damn unstable. Debian unstable isn't too far behind, and is, in the main, pretty damn good. Unfortunately we also use debian for a hundred or so boxes at work; I've been waiting for about 15 months for sarge to become stable now, wanting to upgrade the damn things, but it just isn't happening...

Re:no shit, einstien!

[nns6561]

Try Ubuntu. It's nearly as up to date as Gentoo, but still has all the benefits of Debian. Even better, you can apt-get upgrade to it from a Debian install. I recently changed myself. The upgrade is not entirely staightforward but doable. Better yet, you can always go back to Debian relatively easily.

Re:no shit, einstien!

[LnxAddct]

Fedora has Yum which has a similar feel to apt, but with a simpler syntax, and also apt-get for fedora works very well. www.FedoraFAQ.org answers alot of questions about that stuff if you do decide to check it out. Apparently alot of people have been pleased with Fedora considering that its usage as a server has increased 122% in the past 6 months. It is worth checking out, FC4 is coming out soon and it should be real nice, however I believe FC5 is supposed be damn near magical :)
Regards,
Steve

Re:no shit, einstien!

[andreyw]

"apt-get install packagename" hard to remember ;-)?

Re:no shit, einstien!

[Anonymous Coward]

Sarge is /finally/ almost ready. I've started putting it into pre-production, and haven't had any problems. Go burn an ISO of the pre-release installer to CD, do a network install, and you'll be running quite a nice OS, I think. Nits I have is that they default to 2.4 (come on, an out of date kernel on a brand new release?!), but that's easily rectified. The breadth of package selection can't be beat, and really saves huge amounts of time, unless you run really simple setups. Good sysadmins should all know how to install from source, and they should also try to avoid doing so unless it's really necessary. There are much better (and more fun) ways to spend your time. Anyway, start with the netinst CD image from here:

http://www.debian.org/devel/debian-installer/

That said, if debian *doesn't* fix their release schedule, I think it may be curtains time, at least in commercial environments, and that would be sad. The support from the debian community and the quality, breadth, and management of debian's package selection beats that from the corporatized linux distros hands down. There's really no reason why the same principles that drive the success of free software generally don't equally apply to the production of a quality distribution. It just needs to Get Out the Door. For my part, I am certainly going to find a way to help; clearly help is needed.

Re:no shit, einstien!

[linguae]

If you're willing to switch to a different OS altogether, try FreeBSD. FreeBSD has a Package and Ports system. Packages are pre-compiled binaries that can be fetched and installed, and Ports is a way of installing software through source.

To install Firefox, for example, you can type pkg_add -r firefox, and it would fetch a Firefox binary from the FreeBSD servers and install it from your system. If you prefer to compile Firefox, just cd to /usr/ports/www/firefox and type make install clean. It would automatically fetch the latest Firefox sources and compile them. Ports also resolves dependencies too; if GTK 2.4 or later isn't installed on the system (which Firefox requires), it will also fetch and compile the latest GTK if it isn't installed on the system.

It is also pretty easy to upgrade all of your packages and ports, too.

There are three ways that you can get FreeBSD. Every 5-6 months there is a FreeBSD release (FreeBSD-RELEASE). For example, FreeBSD 5.3 came out last November, and a FreeBSD 5.4 release is slated for April. However, if you want a more upgraded version and track development, there are two directions you can go: FreeBSD-CURRENT and FreeBSD-STABLE. CURRENT is the development branch that adds and tests new features, while STABLE includes the finished features, ready for one of the RELEASES.

You can find out more about FreeBSD here. It has many of the features that you like in Debian, except updated much more often. Only thing to tell you is that FreeBSD isn't Linux; there are some key differences between the two operating systems that you should be aware of.

Re:no shit, einstien!

[toddbu]

Try Mandrake. Always current, great patch support, and urpmi is a dream. We're running it on production servers with virtually no problems. You'll want to shake out new versions before deploying them since the occasional issue crops up, but usually the next patch release fixes everything.

Re:no shit, einstien!

[Anonymous Coward]

Yeah, and apt-get is now available for RPM based distributions just as it's been available for a long time for DPKG based distributions (Debian). Whoopeee do.

Re:no shit, einstien!

[Monkelectric]

Gentoo is for those who want to live on the bleeding edge. Like me (I was also involved in the original formation of Gentoo, so I am defensive:). If stable is important to you, have a look at Gentoo Stable. I know folks who run gentoo on computational clusters and such, so it is certainly possible :)

Re:no shit, einstien!

[guacamole]

Nothing? Not even up2date, yum, and this other tool that mandrake has?

Re:no shit, einstien!

[guacamole]

FreeBSD's ports are fine though I find three problems with FreeBSD's system. 1) The latest versions of packages for which there are ports make files are often not available through pkg_add -r because no one bothered compiling the binary. 2) Compiling all this crap can be slow, specially if your box is say Pentium II-class PC. 3) The ports tree is changing all the time (in other words, it is still sort of unstable..). I prefer RedHat/Fedora/Debian way of maintaining the packages on released distributions. When there is a security problem or some other severe bug in a package, they patch the same version of package that they released originally and then release that as an update. On FreeBSD, I often feel that many packages are being needlessly updated from a sysadmin's point of view.

Re:no shit, einstien!

[evilviper]

Just 'emerge packagename' and the package is downloaded and installed.

I'll call bullshit on that.

Portage has great potential, but it's far from usable yet. Right now, after the first time you've done an 'emerge sync' (put simply, to update the ebuild list), installing any program is likely to result in portage downloading a brand-new kernel, even if you've got 20 different versions installed.

XMMS might be a good example. It (optionally) depends on alsa, and alsa requires the kernel source to compile it's modules. Well guess what? Your installed version of alsa is no longer the latest version available, so it has to be installed again, and your kernel source is no longer the latest either, so that is going to be downloaded and installed to.

This is a basic example. If you want to update one program that depends on gnome/kde libs, good luck, because the latest version of EVERYTHING is going to be downloaded, compiled, and installed. When you come back a day later, it will still be compiling, and filling up your hard drive, unless you are very carful and manually resolve these conflicts.

It's incredibly infuriating. It can be worked-around by manually editing the config files of each ebuild you want to install (and they aren't just simple little text files, either.) but in my opinion, compiling a handful of packages from source in the first place, is infinitely easier.

In case anyone is wondering, I've gone back to slackware after my failed 1+ year experiment with Gentoo.

Re:no shit, einstien!

[ispepalocacoc]

I may be mistaken, but with emerge you can mask any package you like. You can keep any package at any specific version you like, and if a package absolutely needs you to upgrade a package you have masked, then you simply unmask it and away you go. Complete control over what gets installed, and what versions get installed.

Re:no shit, einstien!

[IntergalacticWalrus]

Allow me to return your bullshit call. emerge only updates dependencies if you specify the --update/-u option. And yes, you can update a program without that option. In that case, it will just update the specified program, period. There. No dependencies pulled, that is unless there's something new in the dependencies that needs to be added or updated.

Re:no shit, einstien!

[linguae]

1. Agree. I installed Firefox a few days ago after doing a clean install of FreeBSD. When I decided to install through pkg_add -r, it fetched Firefox 0.9.3! I decided to install Firefox 1.0.1 through ports instead. Somebody needs to upgrade those packages....

2. Agree. My fastest machine is a 475MHz K6-2 with 64MB RAM. Firefox takes about 3-4 hours to compile on this machine (not including dependencies). Even so, I like the ports system because of its effciency, and I personally don't mind waiting a while to compile applications; I have other machines.

3. The reason why the ports tree changes all the time because it reflects changes in the software versions. Plus, if there is any security updates in a package or port, it is very easy to fetch and compile an updated version.

That is one of the core differences between BSD and Linux. BSD is built for installing by source, Linux is built for installing by pre-compiled packages. They're just two different philosophies.

Re:no shit, einstien!

[aanantha]

yum and up2date are crap. On FC3, they rotate through *all* mirrors, even mirrors that are in far off parts of the world. (You can hard code mirrors but you have find them). When a download of an rpm hangs up, you have to kill the process and restart. When you restart, they redownload all the rpms all over again. yum has no GUI. up2date has a GUI but it only lets you update rpms not install anything new. You have to use "Add/Remove Programs" in FC3 which will install the old version off your CD-ROM. Then you can update it. There are a couple attempts at GUIs around yum but they suck. The best is yum extender. The UI hangs when yum hangs. The yum output is in the last tab so if you don't switch to that tab before yum hangs, you won't be able to Ctrl-C yum to abort a download.

I've given up on yum and up2date and switched to apt and synaptic on FC3. Works like a dream. Mirrors can be set up within a CLI for apt. The synaptic GUI is excellent. Fedora semi-officially maintains the apt database but the apt database is always the last to be updated when rpms are updated.

yum and up2date existence is very questionable. They're fundamentally designed around the idea that no new packages will ever get added to the distribution after release. But the Fedora team has a religious attachment to yum so things will continue to suck for new users.

Re:no shit, einstien!

[Look KG486]

Portage. Beats all other package managers.

Re:no shit, einstien!

[Anonymous Coward]

Try portage some time. You'll be pleasantly surprised (unless you use ~x86 (i.e. unstable) packages).

Re:no shit, einstien!

[Anonymous Coward]

i know i'll be flamed for it, but Suse with apt-get is actually quite nice, i was a long time debian user but my desktop and laptop run both on suse 9.2 and i'm thinking to move my servers there too..

Re:no shit, einstien!

[Anonymous Coward]

why don't you try archlinux? a very nice distro and it's package manager, pacman, is a great alternative to apt-get imho..

Re:no shit, einstien!

[evilviper]

with emerge you can mask any package you like

Okay, but going through and manually masking a program at a time is very time consuming. Hence my statement that installing from source is easier than fighting with portage.

Re:no shit, einstien!

[beef3k]

You can use apt with Fedora (although yum is the package manager that's being developed actively to fit in with Fedora).

Same thing:

apt-get install
yum install

Wait a few months for Fedora Core 4, I'm sure you'll like it

Re:no shit, einstien!

[guacamole]

If you don't like the mirrors that up2date chooses for you, feel free to put your favorite mirror into /etc/sysconfig/sources.

Re:no shit, einstien!

[fire-eyes]

Portage in gentoo definately does. Screw the whiners, try it yourself. Don't like it, kick it off.

Re:no shit, einstien!

[scharkalvin]

Sarge is /finally/ almost ready.
Great! When it finally DOES release Sarge will be
only 2 years obsolete compaired with Woody's
3!

Tell you what WOULD be nice. If Debian had an Apt-Emerge option to allow automatic back port builds of packages from 'testing' or 'unstable' from source into the current 'stable'. Granted this wouldn't be perfect, some packages can't be back-ported just by building against older libs, but it would be a start!

Re:no shit, einstien!

[Anonymous Coward]

Nothing compares to apt-get, and that is the biggest shame of all.

Quite right. YaST is much better =P
I was a loyal debian user until I discovered SuSE. Now I can't imagine ever going back. Try it some time!

Re:no shit, einstien!

[Just Some Guy]

I've been playing with Gentoo lately and like what I've been seeing. One huge win over Debian is its concept of "USE flags"; you tell the build system what functionality you'd like to use, and every single package that can support it gets it automatically. For example, setting USE="kerberos" gets you Kerberos (hint: Active Directory) support throughout the entire system, from SSH to Evolution. Contrast with Debian's current setup of having "foo" and "foo-krb5" packages, often at different versions and compiled with different options.

Also, I am not a 1337-optimization fan, but adding USE="3dnow mmx" and CFLAGS="-O2 -march=k6-3" to my setup actually makes a noticeable difference in how fast many apps run on my little laptop at home. I don't so much care that vi now sits idle 4.5% faster, but I do like the fact that I can make SSH connections a lot more quickly than I could with Debian's "compile for 386" policy.

I certainly don't think Gentoo is the Ultimate Linux Distro, and I still use Debian in a lot of places, but it's certainly worth taking a look at.

Re:no shit, einstien!

[techguy911]

No need for Apt-Emerge, just use www.backports.org
Add this line to your sources.list: deb http://www.backports.org/debian stable

Hey.

[JollyFinn]

Hey why not the developers would do with the only incubator they are going to use, get a 9 month release cycle for their favourite (and only child;)

Re:If it's stable, it doesn't need to be updatedOf

[cperciva]

I can see the need for keeping ahead of security bugs...

Speaking of which... *tap* *tap* is this thing turned on? Is anyone from the Debian security team listening? I've got a security issue here... I've e-mailed vendor-sec (3 weeks ago)... I've e-mailed debian-security-private directly (1.5 weeks ago)... are you guys planning on responding some time this month?

(Yes, I'm entirely serious. Slashdot isn't my preferred channel for communicating with other security teams, but the usual mechanisms seems to have failed, and I figure that there must be at least a few Debian people reading this story.)

Greener pastures?

[Anonymous Coward]

" .. to stop Linux users heading for greener and more updated pastures.

And what are those? We migrated recently from RedHat to Debian. Before that we were looking hard for what distro to replace RH with and Debian was the best choice by far. We are running Debian "testing", I admit that. So far, Debian has been excellent, it exceeded all of our expectations. Apt-get/aptitude, it made our life so much easier.

Somewhat shorter cycle would not hurt though. IMO, one release in 2 years is optimal with 6 months of mandatory migration phase support for prior release.

One Word...

[Anonymous Coward]

UBUNTU

Re:One Word...

[Anonymous Coward]

kanotix is the way to go

I never thought of Debian as having releases

[futuresheep]

Debian was the one distro that I never really thought of having official releases. It has versions that are fluid with their packages:

Stable
Testing
Unstable

Each have their own rewards and risks, but the key to me, was that with the netinstall disks, they never went out of date. You never had a CD set full of six month old packages, you had your favorite debian versions latest, usually day old release, a download away.

The new installer is excellent, and with the lack of X based GUI, will still work with a minimal download.

Re:I never thought of Debian as having releases

[killjoe]

The answer is so simple I am surprised nobody has mentioned it.

Trim down the number of "official" packages. Right now there is something like 3000 packages in the debian system. Why not cur that down to a thousand. Take the top 1000 most popular and best maintained backages and call it debian.

The rest of the packages can go into "ports" or "contrib" or something. They would still be there if anybody wanted to install them but they would not hold up release cycles, debian would not guarantee they would work with the rest of the system.

The great thing about debian is that by using stable you are promised that nothing you install will break your system. They can still promise that but just with less packages.

Re:I never thought of Debian as having releases

[Justin205]

It's more than 3000. IIRC, it's more on the order of 8000. But yes, your point is taken nonetheless. But 1000 is probably too few. Maybe aim to take it down to 3000 or so.

I can't find an exact number on the packages in Debian at the moment, but doing a package search for "a" in the stable package set comes up with almost 4000 matches (that is 4000 packages with "a" in them). A description search for "e" brings up 9157 packages. So apparently it's (at least) almost 10000 packages currently.

Re:I never thought of Debian as having releases

[plumpy]

That's essentially what Ubuntu does: takes Debian 'unstable', chooses a set of X packages that are deemed "officially supported", polishes them, and releases every six months.

That course isn't really so possible for a random assemblage like Debian, though. Since everyone who works on Debian is a voulenteer, you can't really tell the guy who's packaging svn-workbench that "your little obscure program isn't a part of the official distribution anymore, please work on something more important". Voulenteers work on what they want to work on. If you tell Mr. svn-workbench guy to stop, he'll probably just stop working on Debian altogether.

Also, most people would argue that the 8000ish packages is one of Debian's strongest points. You don't have to download a random RPM off some website that's of unknown quality and hope it works with your version of FC... In most cases the program you want is right there. I like Ubuntu a lot now, but if it didn't give me the option of installing all the misc. debian packages in addition to its "officially supported" packages, I wouldn't use it.

Re:I never thought of Debian as having releases

[Bishop]

This is ofcourse an excellent idea. It has a big problem though: "who decides what packages are part of the base?"

The way that the Debian project is setup there just isn't anyone with the power to make that decision. The Debian Project Leader has very little power. The decision making process is, by design, decentralised and in the hands of the package maintainers.

Re:I never thought of Debian as having releases

[Chalex]

Good idea. Guess what the Ubuntu guys are doing? :)

They cut down the number of packages to less than a 1000 and call that Ubuntu. All the other packages are still available, but not guaranteed to be up to date.

Re:I never thought of Debian as having releases

[damiam]

That's essentially what Ubuntu does: takes Debian 'unstable', chooses a set of X packages that are deemed "officially supported", polishes them, and releases every six months.

The difference is that Ubuntu is desktop-oriented, and while it's polished, it's not really "stable". Ubuntu is fine for workstation use, but Debian stable does not break. That's why it takes so long to get a release out the door.

Re:I never thought of Debian as having releases

[killjoe]

It's a community effort so why not a vote? Have people submit a list of their favorite packages up to a thousand and pick the winners. I very much doubt you will get a thousand packages BTW so you can add the best maintained and historically the most stable and bug free packages to round it out.

This is why I changed to Gentoo

[Anonymous Coward]

Debian is great, but hey the packages come out too slow!!!

I changed to Gentoo because a lot of the new software took far too long to be released as a debian package. Sure, I could have just downloaded the software, make install, etc blah. But I wanted to manage my packages!

For this very reason I switched to Gentoo.

The only thing annoying about Gentoo is compiling time - which is still quicker than waiting for Debian packages to come out.

Re:This is why I changed to Gentoo

[HoaryCripple]

But I wanted to manage my packages!

You can manage your software just fine in /usr/local.

What's the problem?

[natrius]

People aren't leaving Debian for greener pastures. They're leaving Debian for Debian derivatives. If the last three months on Distrowatch are any indication of how much each distrbution is being used, then Debian is the most important distro out there. Ubuntu is #1, Mepis is #3, and Debian itself is #6. The Debian project has obviously doing something right if some of the most popular distros choose to base themselves on it.

On the other hand, the fact that derivatives are necessary is a sign of Debian's shortcomings. I haven't used Mepis in over a year, but the last time I used it, it was basically Debian installable off of a live CD with easy to use configuration tools. That says that Debian proper is hard to install and lacks user friendly configuration tools. The former problem has been fixed, but I'm not sure the latter has been. Ubuntu is Debian with a shorter release cycle and paid developers to add polish. This shows that users obviously take issue with Debian's long release cycles, and once again, the administration tools. Anyone who is running the development version of Ubuntu right now knows how easy it is to keep things up to date. The newer software also takes advantage of advances on the Linux desktop, such as Project Utopia. I can plug in USB devices, and they just work. It's nice, and Debian proper misses out on things like that because of the age of its packages.

So who uses Debian stable? From the things I hear, it's people who want a long release cycle. Woody users have been getting security updates for however long it's been since the release. People like that. Ubuntu is supported for 18 months after a release, which is likely to be too short for some people. I don't see how Debian loses out from desktop (and some server) users using the derivatives. Ubuntu is the main derivative, and all its work goes back into Debian proper. When etch is getting ready for release, the job is going to be much easier to do, since Ubuntu has already done much of the work ahead. Sarge has been in some sort of a freeze for most of the time Ubuntu has been around, so they haven't been able to reap the benefits of Ubuntu's presence. People getting paid to work on Debian is a good thing, not something to be angry about, which is the sense I get from some posts on Planet Debian.

So if Debian shortens its release cycle, where does that put it in the Linux ecosystem? I doubt they will be able to support security updates for multiple stable releases, which is what they would have to do with a short release cycle to maintain the current length of support. As much as Slashdotters like to poke fun at Debian, it plays a very important role. Does it really need to change?

Debian developers, thanks for making such a great distribution. There are lots of Ubuntu, Mepis, and Debian proper users that appreciate it.

Re:What's the problem?

[cpeterso]

if Debian shortens its release cycle, where does that put it in the Linux ecosystem? I doubt they will be able to support security updates for multiple stable releases, which is what they would have to do with a short release cycle to maintain the current length of support.

If Debian released more often, I think third-party groups (or companies) could more easily support backporting security patches to the previous stable version of Debian than the current situation where third-party groups are "forward-porting" Debian's unstable development branch by including more recent packages and patches (and creating incompatibilities between those Debian-based distros).

Re:What's the problem?

[SuperQ]

Exactly, I have some mission-critical style server that are happily churning away in the corner with woody, I don't want 6 month release cycles, with 18 month drops in support.. it pushes packets, it doesn't surf.

My laptop is Ubuntu, my SO's laptop is Ubuntu.

My home file server is sarge (and maybe MythTV soon)

The router is Bering uClibc

All debian/based.

I liked the solid 2 year release cycle before woody, bo, slink, potato, woody.. all nicely timed... sarge is just getting too fat to get through the damn door. It's about time to put the whole thing on a diet or something.

Re:What's the problem?

[natrius]

If Debian released more often, I think third-party groups (or companies) could more easily support backporting security patches to the previous stable version of Debian than the current situation where third-party groups are "forward-porting" Debian's unstable development branch by including more recent packages and patches (and creating incompatibilities between those Debian-based distros).

That's true. My point is that with Debian's current structure, it can't do both. There is already a Debian with a short release cycle, and that's Ubuntu. If Debian does faster releases, it will only be duplicating effort. I'm not sure what time span people are aiming for in Debian, but let's say it's an annual release. To provide value over Ubuntu's current offering, Debian would have to provide security patches for longer. Ubuntu supports its releases for 18 months, so Debian would have to aim for 2-3 years, which is pretty much it's current length of support. An annual release with 3 years of support means supporting 3 stable releases at a time. I don't think that's practical. As I have said before, Debian's selling point is its long support. From what you've stated, it seems like you'd like Debian and Ubuntu to switch roles, with Debian providing timely releases, and third party groups (Ubuntu) providing security patches and support for older releases. Why switch things around when a working system is already in place?

The eventual argument that gets brought forward in Ubuntu vs Debian discussions is "What happens if Canonical goes under?" Well, what you're left with is Debian with a larger and stronger community than it had before. Many people have moved to Debian based distributions because Ubuntu came along. In addition, Ubuntu's policy regarding community development is much more relaxed than Debian's. After six months, Ubuntu has built up a team of community developers to help maintain packages. With each passing month, more and more names get added to the list. With the addition of a KDE based Ubuntu version, Kubuntu, that list will grow even more. Even without Canonical, Ubuntu will still have its community, and would bring its developers back to Debian proper. Ubuntu could still continue to exist as a branch of Debian.

With so many people using and developing Debian, directly or indirectly, I don't see where all the alarm is coming from.

Re:What's the problem?

[joib]

I liked the solid 2 year release cycle before woody, bo, slink, potato, woody.. all nicely timed.


Actually, if you look at the release dates, you see that each release cycle has taken about 6 months longer than the previous.

I think what hurts debian is not necessarily that the release cycle is longer than most distros, but that it is unpredictable. They have been promising that "this time we'll release faster" for ages, and nothing happens, every release drags and drags. I think they should make a schedule (precise details don't matter), and make sure that they follow it.

Re:What's the problem?

[Havoc114]

http://tinyurl.com/48ok3 For some one who has used Redhat, should they consider using Debian? When I've changed distros that aren't based on a similar ones things can be quite different. I'd imagine that some people, myself included, don't have the time to try out every single distro and find which is best suited for them so they stick to one.

Re:What's the problem?

[Anonymous Coward]

http://linux.slashdot.org/comments.pl?sid=142888&c id=11973242

Some things will be different from Red Hat, but unless you're admistering some sort of server there are only a few things to learn. If you're just using it as a desktop, there's basically nothing to learn. You update and install packages with Synaptic, and all configuration is done with the stock Gnome and KDE tools.

Re:If it's stable, it doesn't need to be updatedOf

[Anonymous Coward]

Uh, nothing has changed in the nearly the past 3 years?

In that time, most of the civilized Linux world has switched to X.org X server, dumped whatever browser they were using and gone to Firefox, switched to the 2.6 kernel, etc. etc. The list goes on.

It's quite simply insulting that the default install comes with virtually every component being out of date. It's a shame because apt-get is so nice. (Yes, I know you can apt-get on other distro's but still)

well then

[b17bmbr]

go fork off. it's open source y'know. seriously though, i'm not too up with debian, but don't they have an unstable tree that keeps things pretty much up to date? and aren't there plenty of packages one can apt-get? and hell, if people want, there's fedora or mandrake for more "current" distros. and nothing stops someone from compiling a new kernel if they should choose. debian tends to be the more linux guru distro anyways, unlike many more user friendly ones. big deal. choice is great. distros like lindows, lycoris, and especailly knoppix use deb as their base. so, big deal if they want to be more methodical.

stable = insecure

[Anonymous Coward]

chkrootkit doesnt find packet sniffers!

snort doesnt pick up port scanning or anything for that matter

gaim cant use any other protocol other then jabber or some shit!

i wouldnt use it for a server my self its too old and insecure..

long live sarge!!!

Re:stable = insecure

[Anonymous Coward]

Using sarge sounds like a good idea except there is no guarantee of timely security updates with sarge since they have to pass the 2 weeks "testing" phase in unstable before actually entering "testing" i.e. sarge.

So if you are that security conscious but you are using sarge you haven't been paying very close attention I guess!

Really you should just scrap the whole Debian mess for BSD or a more modern distribution of Linux.

Re:stable = insecure

[Anonymous Coward]

you have a point there my friend. i will be moving to bsd in time.

the thing with debian is its in limbo. stable is ment to be secure but it fells very insecure. unstable is ment to be insecure but its feels secure because it has the latest software.but because it has the newest software doesnt automaticly mean its secure.

my point from long live sarge is that i hope for sarge to become the new stable.

Re:stable = insecure

[Anonymous Coward]

I agree that the age creates some security issues, but they are better solved by backports of particular packages. Chkrootkit and snort in this case. Backports.org has them both.

What is not OK about this solution is that the admin has to figure it out and accomplish it himself. It seems to me that Debian would be improved by recognizing a few different classes of packages. Security packages would have their own upgrade cycle, weighted toward pushing out essential features and data updates (chkrootkit, etc.). Core system packages would stay very similar to the current release cycle, which is weighted toward stability in every sense of the word, and not to features (gcc, libc, etc.). Application packages would, preferably, actually branch into a couple of options. Maybe the original stable version with backported security patches but another option for the latest release of the application backported/compiled for the stable core Debian distro. There's really not much of a difference here: if Backports.org was an official Debian project and part of Debian stable, then it would nearly be what I'm describing.

Re:stable = insecure

[Anonymous Coward]

Then install those things from their respective sources if you feel the need. Most programs which are run on Deb stable need to serve HTTP/FTP/Whatever, and be stable and secure. Not necessarily up to date (backported bugfixes excepted of course).

Re:stable = insecure

[Anonymous Coward]

The other thing that pisses me off about stable is they backport the patches to old versions. Not only does this take longer to code and test but it makes your server look to attackers like it's running vulnerable software. So when some kiddy scans your net and see all these antique versions of sendmail it gets his attention and he starts to focus on your servers even though they aren't vulnerable. I don't want to bring any more attention to my net than needed. Having every script kiddy banging his head on an old version of bind or sendmail wondering why his sploit isn't working really is something I'd rather avoid.

Re:stable = insecure

[Narchie Troll]

Um, why are you using gaim on a server?

Re:Bush wishes he was Irish

[Anonymous Coward]

The point was that he sounds incompetent and acts like a drunkard, you fool. Just like you sound like a dick, doesn't mean you are one. Got it?

#1 Reason Why I Don't Choose Debian For My Bus.

[SlashChick]

We have over 100 Linux servers, but we chose CentOS as our default OS. We could have chosen Debian instead. In fact, the control panel we use for our customers (DirectAdmin) runs on Debian. But here's the #1 reason I didn't choose Debian:

[hypothetical scenario]
Customer: "What operating system version do you use?"
Us: "Debian unstable."
Customer: "...unstable??"

The close-behind #2 reason is the installer, but I understand that's getting fixed. IMHO, Debian should strive to release a new stable version every 6 months, with 12 months being the maximum time between new stable releases. As it is, I cannot justify using Debian for business purposes when their offering that coincides with what we need is labeled "unstable".

Re:#1 Reason Why I Don't Choose Debian For My Bus.

[Lunix Torvalds]

Out of curiousity, did you try running Solaris x86, and if not why?

Re:#1 Reason Why I Don't Choose Debian For My Bus.

[uofitorn]

Why not just tell your custumer that you use "Debian"? If he knows enough to ask which release, I think he'd be less likely to cry foul when you tell him Unstable....

Re:#1 Reason Why I Don't Choose Debian For My Bus.

[pasamio]

My university runs and supports Debian Sarge.

Re:#1 Reason Why I Don't Choose Debian For My Bus.

[Terrasque]

Tell them you're running Debian S.I.D - that should fix it ;-)

Re:If it's stable, it doesn't need to be updatedOf

[Storlek]

As I said here, it might act like Debian, but Debian it's not.

A notable problem with using "spinoff" distributions is package compatibility. Can I install any .deb package on Ubuntu without possibly causing binary version problems? Similarly, can I build a package on Ubuntu, give it to a Debian user, and be sure that it'll work properly on their system?

This is a problem with rpm-based distributions; I don't know if apt handles it in a smarter way than rpm, but I've been burned by it and I'm hesitant to try and see. While on the surface everything may seem to function properly, you never know when doing something seemingly innocent like installing or upgrading a package can open up a huge can of worms. I know; I tried installing some packages from my Mandrake 8.2 CDs on a Red Hat system. The first couple worked without any problems, but I tried installing another package that happened to mess with some other file that was already on the system, and it broke several other seemingly unrelated programs.

Question?

[deathguppie]

whats the only thing that takes longer than a full Gentoo compile....

Re:Question?

[Anonymous Coward]

waiting for GWBUSH to get kicked out of the white house maybe?

which will never happen

Re:Question?

[Terrasque]

Longhorn?

Re:Question?

[WillerZ]

Meh. Troll away all you like, but emerge -e world takes under 12 hours on my desktop box. Many things take longer than that.

AMD64 architecture helps...

Just so everyone knows...

[Anonymous Coward]

Debian _UNSTABLE_ is shipping Xfree 86 4.3. There have been, quite literally, _thousands_ of bugs fixed since then.

Stable does not always equal good.

Good explanation.

[ta bu shi da yu]

Someone mod the parent up?

Re:If it's stable, it doesn't need to be updatedOf

[Pandora's Vox]

Debian and Ubuntu are currently similar enough that i have yet to hear of this happening, though i'm sure it's possible. note that the ubu dev model is something like this: snapshots of debian unstable every 6 months, with fixes applied and fed back into "vanilla" debian. as such i think that we're going to continue to see them being very similar.

-Leigh

food for thought

[iggymanz]

fedora: the blowfish sushi of distros, exciting, dangerous and for daredevils. It may kill your machine

redhat: the cafe food in the basement of the megacorp, great food but at airport restaurant prices.

novell/suse: the suits come in the front and pay to sit down and get served the same great food most of which is given away at the soup line in the back.

white hat: sneaks the food away from redhat and does the soup line thing. Some seasoning missing.

mandrake: tastes like redhat with somewhat better seasoning and operated kind of like the suse restaurant

gentoo: gourmet ingredients for you to build your own 9 course dinner, hopefully you don't starve in the meantime

debian: stale, week-past-expiration date bread that won't hurt you, and some rather tasteless but nurishing year-old jerky to put on it.

Re:food for thought

[jcuervo]

debian: stale, week-past-expiration date bread that won't hurt you, and some rather tasteless but nurishing year-old jerky to put on it.

Debian/unstable: like the "secret menu" at In 'n Out. It ain't on the menu, but it's good. :-)

Re:If it's stable, it doesn't need to be updatedOf

[wdd1040]

Why not post it to somewhere anonymous like SANS, etc?

They can contact the teams in other methods I'm sure, and if not, they can publish it and force a fix.

debian

[VAXGeek]

it's important to look at debian as a concept as a whole. when you say "debian stable", you're talking a particular snapshot of all these programs 3 or 4 years ago that have been analyzed and proven stable. if you are looking for what linux provided as a whole 3 years ago, you are probably in the right place. why is it so bad to have a clearly defined role for this "stable" distrobution? it's called "stable" because that is exactly what it is. rock solid stable. if you want fancy jazz, no one stops you from using testing or unstable. despite the scary connotations, testing has proven to be stable as well.

whoops!

[MarcoAtWork]

I thought I started using Slackware in 1991-1992, but looking back at my old emails it was around October 1994, my bad.

Separate Testing and Frozen

[nns6561]

This should help reduce the long time between releases. Every 6-12 months convert testing into frozen. Frozen should be considered a release, and supported as such, with security updates for ~18-24 months. Periodically, when few bugs exist in frozen convert frozen into stable. Maintain security updates for the last two stable releases. The idea is that for mission critical servers, stable would be used, but frozen would be sufficiently stable for workstations and less reliable servers. Just pushing out security updates for testing isn't enough because the continual upgrade process is too much for lightly administered machine. This also insures that testing is never frozen for long periods of time, which is good for desktops that want to be running the latest software. I like the concept of supporting fewer architectures, but this needs to be done properly. They should only allow bugs in the rare architectures block packages in stable. Most of these architectures are only used for servers or other rarely updated machines. Plus, since stable is not the release, these architectures won't block releases.

the fuss about Debian's "cycles"

[l3v1]

This is a "once more" new iteration of the same old idea of Debian updating their stable branch not often enough. And as always, I have to respectfully but totally disagree.

For one, people should really understand and see, that not all Linux distributions are just there to suit the newbie (l)users' desktop needs. This is just the attitude people gather while being full-blown Windows users and then fiddling around with some Linux, thinking it's cool and if he can't find his way around, then at least that';s another reason to bash.

Debian's stable branch is just _the_ perfect distro for servers. You can argue with this statement, but I will _not_ listen to home users' hysterical crap about the newest kde/gnome being necessary. There are places where that simply doesn't matter.

Where I spend my working hours very few people use Linux distros on their desktops, really very few, but almost all our servers are Linux based. The two of them where I hve root access are Debians. One is a current stable Woody, being web&mail&db&cvs&related server which I installed last year because the previous machine had a major blowup. The other is a Debian Potato (!) which is the previous [i.e. before Woody] stable branch, which is our dns server, up and working for ... well, since about the Potato release.

No desktop environments, no x, just good stable and reliable code which I trust and - most importantly - _very_ _easy_ to maintain.

At home I use Debian SID for about 4 years now. Updated about weekly, _very_ stable and usable. It has all the desktop fun I need. Most important: it hasn't been reinstalled since the first install just always copied over to the changed machine (about once in a year, I always hand-build my machines ever since I became acquainted with the screw driver), updated the necessary stuff and keep it always apt-get dist-pgrade-ed.

For me, and for many others out there, Debian - and now the quite many Debian-based distros, hey, there are even Debian SID-based distros now (!) - represent _the_ _GNU/Linux_ _distro_. For the others, there are plenty of others you can use and that is exactly why Lnux distro forking is a Good Thing, try not to forget that.

Re:the fuss about Debian's "cycles"

[Juergen Kreileder]

Debian's stable branch is just _the_ perfect distro for servers.

Believe it or not, there are people who need something more up-to-date than woody on servers.

Re:the fuss about Debian's "cycles"

[vonsneerderhooten]

What couldn't you install that wasn't included on the woody release??
Please, I'd like to know.

Re:the fuss about Debian's "cycles"

[Anonymous Coward]

Ja wohl. I can't imagine anyone writing a NEW application that targets Woody's 5 year old applications.

Re:the fuss about Debian's "cycles"

[l3v1]

And that's why there are Sarge and SID, these can very well serve as your so new application's testing ground.

Re:the fuss about Debian's "cycles"

[l3v1]

Believe it or not

Believe it or not, I also live on this planet.

Well, agreeable, newest Oracle or DB2 [just the quickest two examples that came] could cause severe trauma for every inexperienced who wanted to install them on Woodyr4. And well, Samba is also not the latest. That's true. But than again, who told to always need to use Debian [Woody] ? For general purpose server services [dns,gw,route,@,80,svn/cvs,my/psql,.......] it is just fine. (BTW, no serious trouble should occur with the above two on SID).

Re:the fuss about Debian's "cycles"

[Juergen Kreileder]

Anything that needs decent threads support for instance, woody's glibc doesn't support NPTL.

Also, it defeats the purpose of a distribution if you have to build and maintain apache2, mysql-4.1, cyrus-2.1, exim-4, etc yourself.
And no, running testing or backports doesn't solve that. There are no timely security updates for testing and backports.

Re:the fuss about Debian's "cycles"

[JBv]

I agree to some point.

In servers, I couldn't care less about X or fun and games. I do care about hardware support and features.

I really wish the chasm between stable and testing was not that wide. It is stupid that debian stable net-installs with no problems in a pentium 200 Mhz with a 10baseT nic but fails to install in 1 year old computer because of hardware compatibility with the nic and/or scsi controler.

Re:the fuss about Debian's "cycles"

[loopkin]

This comment is totally insane.

I happen to run several servers on woody, and i can tell you: it's almost impossible to run it without using backports or handmade packages.
The reasons:
- hardware support (2.4.18 linux kernel is very very very old in that aspect)
- lacking (at all, or partially) big recent software/features (amavis/clamav/spamassassin/samba/...) that HAVE to be there
- blocking bugs that will never be corrected for various reasons (ex. faulty FTP client in woody's PHP, that is so old that it's unmaintaned by the PHP team).
- old packages that make running recent advanced software on top of woody impossible (think apache2+subversion and so on)

So, no, woody sucks for servers too, and a lot. And i can provide you with a lot of instances in which i had big problem with too old woody's packages on a _server_ (nss-ldap, unrar, ...)

Re:the fuss about Debian's "cycles"

[joshua]

The last straw for me with Debian stable, even for server use, was when I installed Debian stable on a server only to find that the network card wasn't recognised. And I'd have to compile a kernel on another machine, burn it to a CD, and copy it onto the server before I could see the network.

The network card was from that minority manufacturer, Intel.

Re:the fuss about Debian's "cycles"

[l3v1]

This comment is totally insane.

I love you too :]

it's almost impossible to run it without using backports or handmade packages.

The problem is 1). you assume nobody else makes backports of the needed stuff or doesn't compile whatever is needed from sources if _necessary_ 2). I also didn't explicitely state that the default packages/install (which for some of us means a plain base install then the rest by apt&hand) can be tailored to your specific needs, which I felt evident.

- hardware support (2.4.18 linux kernel is very very very old in that aspect)

Where did I say you are not allowed to run newer kernels ? I never run debian kernels, currently 2.4.27 and 2.6.10 are the defaults here.

lacking ... that HAVE to be there

Well, for you, and everyone has to specfify his _own_ needs, or is that not that self-evident as it should be. Woody and some hand-installed&maintained applications completely fill our needs.

ex. faulty FTP client in woody's PHP

That would be a problem only when I'd use it :P

apache2+subversion and so on)

There are some of us who constantly get bashed for using 1.3 series apache, but we do. We also happen to use cvs and not svn (yet). But these are _not_ valid excuses, I know, still they are good to prove my point: if it's too old for _you_ use something else. Until then, Debian stable is fine.

Thing is, for every problem there is a solution. One is to make your way with the existing setup. Another is to change it for something else. I'm not against any of these, but if I _can_ choose, then I stick with Debian.

Re:the fuss about Debian's "cycles"

[Anonymous Coward]

How about php 4.3? My workplace requires testing on debian stable and php 4.3 to interface with oracle9, and I had to build fresh packages from source to get it running.

To say that nothing has been developed since 2002 (stable's release year) that is needed on servers is silly.

Re:the fuss about Debian's "cycles"

[loopkin]

Ok, what basically i'm saying:

1- woody + backports or handmade packages is NOT woody. why ? because you have to _maintain_ it by hand (or trust the backporters to do so efficiently).
in other words, woddy+backports or handmade packages is NOT anymore a stable distro. and, i agree, stability is something important for a server, and especially in a company.
2- it's becoming almost impossible to be able to use an unmodified woody on a server, because of the various reasons i've written down. and i'd say it's the case if you've been trying to set up a server with woody on a modern machine in the past 18 months (personnal experience: you want GE, SATA, or so, and, for software, try to set up a mail server in a SME without antivirus & so on).

now 1+2 = Debian sucks for servers.
since it sucks also for desktops, this means that Debian has no future if they can't get themeselves out of that big problem.

now of course, if you use woody only as a DNS server on a 3 years old machine, it's ok. in any other situation, you might run quickly into trouble.

if it's too old for _you_ use something else. Until then, Debian stable is fine.
thing is that it's not that simple:
- when woody came out, for instance, you could set up a mail server without antivirus and spam filtering (it was really another era...). since then, you had to add these features. but you can't do it while keeping an unmodified Debian stable. that means that not only Debian is getting obsolete on new installs, it's also getting obsolete on old installs. and it means too that it can be ok today, and not tomorrow.
- if Debian isn't able to provide a solution in stable for all/most of these problems, that leaves no choice but not using Debian on servers. developping a distro that nobody can use is pretty much pointless isn't it ?

"use something else" is not an answer to the question "is Debian stable useable on servers ?", and the answer to that question is, i'd say, 75% no, 25% yes. Now if it's good for _you_, that's great, but it's very far from being the case for everybody (i've installed or administrate about 20 Debian servers).

Re:the fuss about Debian's "cycles"

[cmdr_tofu]

This is mostly true, *but* I have found that even on servers Debian stable falls a little short. For instance if you want to configure postfix/amavisd/spamassassin, you will find that debian's postfix and amavis are woefully out of date. It's not even an issue of stability, they are just using an old release. If I have to get the production release of _someSoftwarePackage_ by compiling it myself or from a backport source, I lose all the advantages of Debian's patch management. I was very excited about testing becoming stable, but I am not sure what the timeline is for that. Since I am not a Debian volunteer, I cannot really complain. They are doing a good job, but things are not yet perfect. Other distributions may be catching up to debian with alternatives to apt (I have not really checked them out), but I don't think they will have the same quality assurance of Debian. Fedora RC5 might be great for playing nethack, but would you really want to run your business on it?

Re:the fuss about Debian's "cycles"

[Ih8sG8s]

I'll agree the idea that a known stable base, coupled with skilled "roll your own" know-how makes for excellent servers. I will strongly disagree that Debian is the only way to go, however.

I have never seen any distribution that I could not make do exactly as I wanted. There are some that are better placed for the purists like you and I, agreed. Debian is not the only answer, sorry. Slackware for instance is easily as stable, and has a release cycle which includes timely jumps to newer technology.

I'm not here to tout Slackware over Debian, only to say that for the purists, Debian is not the only choice, and any purist could tell you so.

Re:the fuss about Debian's "cycles"

[nmos]

I agree (and said so in another thread). So, does anyone know of any server oriented debian based distros?

Re:the fuss about Debian's "cycles"

[stinky wizzleteats]

(l)users

I think you just hit upon the real reason Debian is losing users and/or failing to attract new ones.

Re:the fuss about Debian's "cycles"

[Anonymous Coward]

OK, I'm done writing my new application. How long should I sit around and wait for Sarge before deploying?

Re:the fuss about Debian's "cycles"

[damiam]

Woefully out of date? Is there a missing feature you really need? Or do you just have an aversion to low version numbers? As you say, Debian has excellent QA, and it doesn't come from rushing new releases.

Re:the fuss about Debian's "cycles"

[Anonymous Coward]

If the stable release is useless on the desktop, why don't they remove all the desktop oriented stuff? That would allow them to reliably release stable on whatever regular schedule would best meet server users needs. They could call it debian server or something.

Re:the fuss about Debian's "cycles"

[cmdr_tofu]

For instance if you want to use amavisd.new (which has a lot more features than amavis like per user
SA settings), you cannot use the postfix supplied with Debian. An MTA is a pretty important thing,
and unless you can have a backport you have to fight against the package management system to get yours working. There are backported versions of the packages available (thanks Norbert!), but there is no guarrantee that patches will be available in a timely fashion. It is for these reasons, that I am now looking very seriously at Gentoo.

If you ignore all the jokes above..

[Anonymous Coward]

*insert obligatory masturbation joke here*

One year term?

[toddbu]

I'm curious to know how any organization can move forward when its leader can be replaced every year. A one year term gives you just enough time to get a good night's sleep between campaigns.

Re:One year term?

[l3v1]

Well, fresh blood can mean speedup. Fresh people wanting to prove they're right can lead to overall improvement. And in one year they can't have enough time to settle in and slow that much down. And maybe one year is just enough for people being able to keep up the pace they have at the beginning.

All this is just hypothetic, but seems enough to justify.

What!?

[LongShip]

What in UserLinux requires that Debian release a new stable release? Why don't you select the best packages from Stable + Testing? Then, you get some cutting edge stuff where it's important and you get stable where it's important. That's what many other Debian-based distributions do.

Claiming that UserLinux needs Debian to release a new stable release so that UL can use it sounds a bit like a big cop out.

There is no UserLinux, there is only Debian.

Re:If it's stable, it doesn't need to be updatedOf

[benjcurry]

Check out Arch Linux. It's a bit young, but up-to-date, fast, elegant and great package management.

American Leaders: We Need to Quit Being Asshats

[jfm3]

jfm3 writes "The lack of diplomacy, ecological consciousness, and ability to think critically since September 2001 is fuelling the campaigns of many candidates for the country's American President role, with many pushing for a politer and more intelligent foreign policy to stop Americans heading for greener and more peaceful pastures."

Perhaps this does not appear on topic, a crime against slashdot of the highest caliber. Of course, both statements are equally absurd because of their obviousness (at least to me).

Devil's Advocacy: Why more often?

[aendeuryu]

Here's my thing. I like debian's stability. In a way, it guarantees a certain degree of consistency amongst debian-based platforms. There's an interoperability on my computer right now that I really like. It's a creation of a sort of de-facto standard. It's pretty calming, to be honest.

Meanwhile, you've got Fedora and Mandrake and Suse constantly pushing towards more advanced releases, but with all those changes, that means there's a lot of programs that are going to need to be updated more frequently. Right now you've got Ubuntu, Knoppix and Mepis based off a reliable distribution. It's a solid foundation. Change the foundation, and all of a sudden you'll throw those three distributions off-kilter.

Wouldn't it be easier for the applications and distributions to achieve their own respective nirvanas of perfection if the goal posts aren't moving?

[/devil's advocacy]

Alright, to be honest, I don't necessarily agree with all of the above, but I REALLY don't agree with change for the sake of change. What seems to be happening here is that a lot of people are saying that debian is stagnating. Yeah. Okay. The formula for mixing hydrogen and oxygen hasn't been changing all that much, and yet it's still pretty useful as is. Want to start messing with that? (Yes, I know there's a big difference between water and debian. I just want to see some of the proponents for change start mapping out the gray area)

So debian's getting stale. So what? It injected new life into my laptop, and I'll tell you what, that ISN'T stale to me.

How about making the Testing branch default?

[carlmenezes]

Well, how about adding in a "server" branch? Those in the know can use this - it will be what used to be called the "Stable" branch. However, for 90%+ of home Linux users, they would like the latest releases, so how about setting the default to "Testing"? It would make sure that most default installs have quite up-to-date packages with minimal bugs, while expanding the testing base greatly.

I have never had a problem with the Testing branch - it's been stable enough for me and been running for well over a year with supervised updates and I love it.

As a desktop user, what I wouldn't like is seeing KDE 2.2 come up after installing Debian after I have just seen all the news on the net about the 3.4 release. KDE 3.3 would be fine and I'd be willing to wait some time for 3.4 to get to me, but 2.2 would be un-acceptable. I would have been much happier if the install simply defaulted to the testing branch and installed even KDE 3.2.

Re:How about making the Testing branch default?

[tacocat]

Adding another branch to manage would only make things slower. The last thing Debian needs is more branches to the process.

Before you recommend this, why don't you prove that the current system does not work and then identify why it doesn't actually work. This is an arbitrary statement not based on fact.

I do agree with the suggestion to drop many of the hardware branches in Debian or at least put them all onto their own release cycles and then lockstep only the best four (arbitrary number) and let the rest die off by attrition. There is evidence that some delays come from coordinating architecture. There is also evidence that this is not the only problem but coordinating so many packages is hard.

To put it into perspective. At work I still use Windows 2000 and at my previous job I used Windows NT 4 up to the year 2000 when I switched.

Re:How about making the Testing branch default?

[carlmenezes]

I think you mis-understood me. I suggested only changing the default install so it uses the testing branch.

All I am suggesting is add more users to the default testing branch by making THAT the default branch for an install. Those who want to use the Stable branch can always do so.

The Testing, Stable and Unstable branches already exist. Nobody is adding any new branches. All we're doing is increasing the number of users testing new software by an order of magnitude. Won't that help in catching bugs faster, faster fixes and ultimately, a better release schedule?

Please try to understand a person's point before you criticize.

Project Management 101

[buddha42]

Debian has three options

1. Reduce the scope of the project
   • fewer architectures
   • fewer packages
2. Add development resources
   • hard to do with volunteers
   • inject some money (bounties?)
3. Streamline operations
   • reduce some of their bureaucracy and excessive policies

Personally I think they would be best served by doing a little of each.

Re:Project Management 101

[natrius]

Reduce the scope of the project

I disagree. One of the greatest things about Debian is the scope of the project. I can install almost anything and not have to hunt around the internet for a package. It's all in one place. I think the currently proposed approach on not releasing the lesser used architectures at the same time at the others is the correct approach. Abandoning them completely would be foolish, but having a whole release held back by problems with software that's not even heavily used is a problem.

Add development resources

This has been done. Ubuntu. People are paid to work full time, and their work goes straight into Debian. This also takes care of the issue Slashdotters have with the long release cycles, since people can download a new version of Ubuntu with the latest version of Gnome, KDE, etc. every six months. The problem it doesn't solve is that of people who want to run Debian stable, but can't use the ridiculously old packages for commonly used web programming languages. The release cycle needs to be shortened, but not by too much.

reduce some of their bureaucracy and excessive policies

You call the policies excessive, but it's thanks to their efforts that is possible to run a computer based on completely Free software (and Free documentation, which is probably the issue that prompted this point). Sure, their policies often err on the side of idealism rather than pragmatism, but I think it's beneficial for the entire community that they do this.

Re:Project Management 101

[IamTheRealMike]

I disagree. One of the greatest things about Debian is the scope of the project. I can install almost anything and not have to hunt around the internet for a package. It's all in one place.

Which says everything about the futility of trying to package every program you could possibly want. Debian hasn't released for years because it's far, far too big. If Debian was in fact an OS and not a constantly shifting, usually broken snapshot of a subset of the free software world, it would release far more often.

To be honest, the best way that Debian could ensure rapid, predictable releases would be to jettison their package repository entirely. Crazy idea, yes?

Well, think about it. What is Debian the "OS" famous for? It's pretty, easy to use installer? No. Its slick, integrated desktop? No. Its power as a server - not even that, really, if you compare Windows 2003 as an OS to Debian as an OS (ie, excluding the random programs that come with it) you'll probably find that W2K3 makes things that require Guru level knowledge in Debian much easier.

In fact the only thing Debian is famous for is having lots of packages. To get decent versions, you need to use unstable which means that often things won't actually be installable, and occasionally an apt-get dist-upgrade will break your system, but hey it's better than no packages at all.

The solution seems obvious - abandon the unproductive and unscalable centralised repository system and release Debian as a core OS that provides the essentials, with the rest of the software provided by third party binary packages.

Re:Project Management 101

[natrius]

The solution seems obvious - abandon the unproductive and unscalable centralised repository system and release Debian as a core OS that provides the essentials, with the rest of the software provided by third party binary packages.

If this happened, there would be no Ubuntu. There would be no Mepis. There would be no Knoppix. All the Debian derviatives exist solely because of the monolithic Debian package repository. Sure, Ubuntu has its own repositories but they all come from Debian unstable.

Furthermore, to me, it doesn't seem like that is what the Debian project is about. Debian provides all the free software that runs on Linux and is worthwhile to package. They also ensure that it is possible to run a computer on free software. Without the central repository, that goal couldn't be accomplished. To do what you suggest would take far less Debian developers than currently exist. These developers would move on to the develop for the third party repositories, but those would be far less useful than Debian's repository and cause massive duplication of effort. Plus, you state that Debian's main strength is its packages, but you want to take those away? Sure, there would be quick releases, but the released product it self would be much less useful than Debian is now.

Debian has problems, but the massive overhauls people are proposing are completely unnecessary. Right now, there is a Debian for everyone. Woody is too old. The release cycle of Debian proper needs to speed up somewhat, but not radically. That gap has already been filled by other Debian based distributions.

Re:Project Management 101

[IamTheRealMike]

If this happened, there would be no Ubuntu. There would be no Mepis. There would be no Knoppix. All the Debian derviatives exist solely because of the monolithic Debian package repository. Sure, Ubuntu has its own repositories but they all come from Debian unstable.

That's not true. The thing that makes Ubuntu great as a desktop and Debian lame is all the work that has gone into the core OS part, not the packages. The Ubuntu universe is riddled with uninstallable software, and main ships with out of date software too (eg, Inkscape).

Furthermore, to me, it doesn't seem like that is what the Debian project is about. Debian provides all the free software that runs on Linux and is worthwhile to package.

It quite clearly does not, at least not by my definition of "worthwhile". What about all the commercial games out there? I guess they're not worthwhile packaging because they aren't free enough. What about up to date Wine packages? I guess it's not worthwhile packaging that, so upstream has to do it instead. Even if you take a massive leap of faith and claim that Debian is able to package everything anybody might ever want, it still has big problems with freshness even in unstable.

They also ensure that it is possible to run a computer on free software. Without the central repository, that goal couldn't be accomplished.

I don't see that logic. Having software pre-filtered means you don't "accidentally" get un-free software, however not everybody shares the same definition of free. See the ridiculous discussion over the Xorg "nv" driver on debian-legal for a good example of that. You can still run a computer using only free software if you want, you just have to not install non-free software. Assuming the core OS is free, that's not hard at all. That has the advantage that for the majority of software you can make up your own mind about what is free or non-free, instead of having Debian choose for you.

These developers would move on to the develop for the third party repositories, but those would be far less useful than Debian's repository and cause massive duplication of effort.

I don't see that logic either. Third party packages (not repositories) would be provided by upstream projects like on Windows or MacOS. Those packages would work for anybody, using a technology like autopackage. Therefore their utility is hugely increased, and duplication is decreased, because the software need only be packaged once for all Linux users.

Plus, you state that Debian's main strength is its packages, but you want to take those away? Sure, there would be quick releases, but the released product it self would be much less useful than Debian is now.

No, Debian would just have to focus on writing a great OS instead of packaging as much stuff as possible. So that means things like, a graphical installer, slick integrated desktop, nice config tools etc etc.

Debian has problems, but the massive overhauls people are proposing are completely unnecessary. Right now, there is a Debian for everyone.

That clearly isn't true otherwise the candidates would not be claiming that Debian users are "leaving for greener pastures" and Ubuntu would not have appeared out of nowhere with a vibrant community almost overnight (guess where they mostly came from ...)

Re:Project Management 101

[drinkypoo]

One of the greatest things about Debian is the scope of the project. I can install almost anything and not have to hunt around the internet for a package.

Yeah, you can install almost anything as long as it's really old. Kind of like the range of colors available for the model T.

Re:Project Management 101

[natrius]

That's not true. The thing that makes Ubuntu great as a desktop and Debian lame is all the work that has gone into the core OS part, not the packages. The Ubuntu universe is riddled with uninstallable software, and main ships with out of date software too (eg, Inkscape).

Ubuntu doesn't just take the "core OS" from Debian. It takes everything from Debian. Without Debian's packages, they'd have to do all the work themselves. I've only come across one broken package on Ubuntu, so I don't know where you're getting "riddled with uninstallable software" from. Ubuntu doesn't ship with out of date software. They take a snapshot of the latest software when their Upstream Version Freeze comes along. For Hoary, that was Janurary 3rd. The version of Inkscape that will be in Hoary is 0.40. This is how almost all distributions that actually put out releases work. There's no way to have a stable release unless you freeze the packages and have a bug fixing period.

What about all the commercial games out there? I guess they're not worthwhile packaging because they aren't free enough.

What I said was that they provide all the free software that's worthwhile to package. Why would Debian developers package commercial software? No free distribution does that. Companies can provide their own packages like Skype does.

What about up to date Wine packages?

From what I heard, the Debian maintainer for Wine hasn't been doing much lately, so one of the Wine developers is working on getting up to date packages for Ubuntu (which will of course trickle down to Debian).

Third party packages (not repositories) would be provided by upstream projects like on Windows or MacOS. Those packages would work for anybody, using a technology like autopackage.

Believe it or not, program installation and update management is better in Linux than it is in OS X or Windows, because you click a button and your entire system is up to date. I don't want to have to hunt around getting the latest versions of each piece of software on my computer. Ubuntu and Debian do that for me. Autopackage does play an important role though. Not every project can build packages for every distribution, or find someone within those distributions to make packages for them. For the distros that a developer can't support directly, they can provide an autopackage that will work for everyone else. However, that's not the way I'd want to install software by default, because I'd lose the ability to update. About a week or two ago, I changed all the "warty"s in my apt sources to "hoary"s and did a dist-upgrade. My whole system was updated. Windows and OS X users can only update the operating system and the programs that come bundled with it. It'd be nice if more developers of small projects would provide autopackages, since they often only post a package for one distro and a generic tarball.

No, Debian would just have to focus on writing a great OS instead of packaging as much stuff as possible. So that means things like, a graphical installer, slick integrated desktop, nice config tools etc etc.

That's not what Debian users what Debian to focus on. Debian users use Debian for the package repository. That is the selling point for Debian. If you take that away, it's just another distribution. You also take away Ubuntu, Mepis, Knoppix, etc. because all of those are based on the fact that Debian has a huge repository of software for you to install. Faster release cycles, graphical installers and configuration tools are obviously much less important to users than the package repository. Whe else would people using Debian and Ubuntu? From what I understand, Fedora has configuration tools for tons of things. People still choose Debian based distros. Wh

Newest software

[HoaryCripple]

Build tools are up to date on debian. Build your own bleeding edge sofware and place in /usr/local. Easy. And I don't want to hear about "But I don't waaannna build my own! There should be a package!" Tough.

Re:If it's stable, it doesn't need to be updatedOf

[natrius]

A notable problem with using "spinoff" distributions is package compatibility. Can I install any .deb package on Ubuntu without possibly causing binary version problems?

AFAIK, packages within Debian itself aren't even compatible with each other. If you're running unstable and you want to give a package to someone running testing, you're out of luck. Why is it a surprise that Ubuntu packages wouldn't be completely compatible? From my experience with Ubuntu, it seems like most Debian unstable packages are forward compatible to Ubuntu, but I doubt the reverse is true. This makes sense. Ubuntu has more up to date packages than even unstable at some points, since Ubuntu applies it's own patches, and the Debian maintainers may not apply them immediately. If they add the Ubuntu repository at a low priority and try installing your package, it'll probably work, but some of their libraries will be updated to Ubuntu versions. That's a bad thing, because it might break future updates within unstable for them. Maintaining package compatibility and achieving Ubuntu's goals at the same time would be impossible to do.

By the way, Ubuntu isn't a "spinoff" distribution. It stays with Debian unstable, then freezes the set of packages and stabilizes them. For the next release, they start over.

Re:Speaking of which, Suse 9.3 next month

[Anonymous Coward]

Where can I download this Debain you speak of? Is it any relation of Debian's?

Except...

[sp0rk173]

FreeBSD maintains the same kind of stability WITH a more current release schedule. 5-stable (unlike 5-release) will give you a very stable system. 5-release will give you a pretty rock solid system, though unbreakability is not guaranteed. Use 6-current and you better expect breakage, though it's not guaranteed. The last -stable FreeBSD milestone? Nov. 6 2004.

Before there's a shitload of replies about 5 sucking - yes it did suck when it was strictly a new technology release. Now bugs have been patched and more things have come out from under the giant lock. Speed has increased, as has stability, and it has earned the -stable tag. The point of this post is just to say stable != extremely out of date. stability is just well-tested, well-written code.

Re:Except...

[Anonymous Coward]

Err no, speed of FreeBSD 5 has not really increased. It is still an old dog here... or did you have some links you can show us?

Re:Except...

[Anonymous Coward]

> The point of this post is just to say stable != extremely out of
> date. stability is just well-tested, well-written code.

One thing to note is Debian's stable is meant to be not just rock-solid, but also "unchanging" stable. Both meanings of the word apply.

Meaning if you install a debian stable, it absolutely positively will not change, except for security bug fixes. It'll be the same system now, tomorrow, in six weeks, and in six months. You won't get a feature change on a debian stable system that messes with your server that may very well RELY on those features acting as they do.

Unfortunately having it stretched out to "unchanging for 3 years" is far too long. I'd like to see 18 months absolute maximum.

Re:Except...

[Anonymous Coward]

That's getting close to Apples to Oranges though. FreeBSD is an operating system, Linux is a kernel. That said, I think a distro like debian could benefit from doing things similar to FreeBSD. Make a stable solid base system that doesn't change much, but update the "perpherial" software more. You don't need changes to 'ls' everymonth, but many people want to keep up with versions of PHP and such. I think many Debian people would be happy with a base system that only changed once a year provided that they supported that system for at least 2-3 years.

Re:Except...

[Anonymous Coward]

How about a reply pointing out that these two statements are incompatible

yes it did suck

and

stability is just well-tested, well-written code

the point is, Debian won't release a 'stable' that sucks, and is not actually stable, and needs to 'earn' a stable tag. You won't see stable needing a big upgrade to actually be stable and not suck. It will be stable, and have occasional security patches.

Re:Except...

[sp0rk173]

FreeBSD's -stable tag means this as well. Very rarely are features backported to the -stable branch from -release or -current. Only security fixes.

Re:Except...

[sp0rk173]

You misread. FreeBSD 5.x sucked in the -release stage. After they ironed out the kinks and squashed the bugs, 5.x became -stable and ceased sucking.

too late...

[Anonymous Coward]

I was a long time Debian user but I jumped ship about 9 months ago for Fedora.

You could release a new stable every 3 months starting next week, it's too late, I've had it with Debian's crap.

Re:If it's stable, it doesn't need to be updatedOf

[Anonymous Coward]

File a grave or serious bug against the package in question.

I release every night...

[Anonymous Coward]

nt

If we agree it's an image problem...

[godless dave]

Maybe all they need to do is call "unstable" something other than "unstable". Call Woody something like "stable server" and call the most stable "unstable" release "stable desktop". Better yet, let some company make a spinoff using those terminologies, and keep the Debian people working on Debian instead of terminology.

summary

[evulgenius]

Debian stable is too old. It doesn't work on latest x86 and PPC hardware. Testing is fine for desktop, but for people who need stable and secure system for servers it's not an option. Since there is no security support for testing and there still are some bugs.
So we really need stable releases more often. Doing it by dropping some architectures makes sense to me, if you can't buy the hardware anyway. Also developers can still work on their favourite architecture and release if they keep up to the speed those 4 most popular architectures are releasing. It just means that i386 won't be waiting if there are some bugs on m68k.

And yes, I run debian testing

Look who's using slashcode!

[Thud457]

I told ya it was the tool of the debbel!

FreeBSD release often and very stable!

[Anonymous Coward]

FreeBSD releases constantly and it's very very stable and secure...and they are even responsible for their own kernel development! All debian has to do is take other people's software and package it! FreeBSD developers have to develop the kernel, test it, package and then port software from Linux or other *nix to BSD as need and make packages etc.

So really Debian has NO EXCUSE for the lame release schedule.

Security

[Anonymous Coward]

I used to use redhat exclusively. Until the 3rd time I had been rootkitted.

Then I switched to debian after finally listening to some advice.

Have never been rootkitted since (was about 2-3 years ago). Will never ever go back to redhat.

On debian

[Anonymous Coward]

Umm debian is well in need of some dusting off. I LOVE their pkg management system. However might I sugest they use a modle like freebsd? Mabie they need to admit the modle used by Gentoo has merit.

This just in ...

[Mind Socket]

... you need oxygen to live.

Is that the best you can do?

[Anonymous Coward]

Customer1: "What operating system version do you use?"
Us: "Linux 2.6."
Customer1: "Cool."

Customer2: "What operating system version do you use?"
Us: "Linux 2.6"
Customer2: "No, I mean, what distribution?"
Us: "Debian Linux."
Customer2: "Cool."

Customer3: "... What version of Debian?"
Us: "Debian Sid."
Customer3: "Cool."

Customer4: "... You mean, Debian UNSTABLE, right?"
Us: "Yep."
Customer4: "Cool, I run that on my home machine! Bugger about that Sarge release schedule, eh?"

mods:parent is clearly redundant

[themusicgod1]

All I can say is Duh.
pay attention, mods

Re:mods:parent is clearly redundant

[Bruce Perens]

Phooey on you. My post was submitted before that other.

Re:mods:parent is clearly redundant

[themusicgod1]

by schnits0r (633893) ... on 23:52 17th March, 2005
by Bruce Perens (3872) ... on 0:00 18th March, 2005
you missed by 8 minutes.

Maybe a bit too often...

[ic3p1ck]

My only complaint is that the testing version of Debian is updated a bit too often. I dislike having to get 10-20MB of packages every week to keep up just in case there are some security updates included (Debian security notifications are only done for the stable release).

I would prefer something in between stable and testing, updated reasonably often with new packages (and features) and also have security releases in between as required.

Re:Maybe a bit too often...

[tacocat]

Get a clue. What would you put in between stable and testing...

The more layers you add to the process, the slower the process will become and the worse the problem will be.

Try being part of the solution and not part of the problem.

Re:Maybe a bit too often...

[ic3p1ck]

That was unwarranted.

If you'd been reading the rest of the discussion you would have noticed that people suggest that the stable distribution is more appropriate for server type environments where stability is paramount. I'm merely suggesting a 'stable' distribution more appropriate for general use (such as mine).

Your post contributes nothing to the 'solution', so I suggest you become a part of the solution instead of being a jackass.

Re:Maybe a bit too often...

[Short Circuit]

You think you've got a problem? My packages were last updated in December, via approximately 15 CDs. I haven't had to reinstall on this machine for about five years, thanks to periodic sets of CDs containing all the Debian/testing packages.

I just got my first home internet connection in three or four years a few weeks ago, and apt wants to go out and grab 950MB of packages. Time for another CD set...

Re:Maybe a bit too often...

[tacocat]

The last thing you can do as a proposed solution is to introduce another layer of complexity to the process of package release management to Debian. What I am hoping to contribute to this whole discussion is some experience with the Debian packages and the other distribution package/release processes.

In order for Debian to release anything it has to pass through three levels of testing an confirmation: Unstable, Testing, Stable. But technically that's wrong since they've also introduced Experimental to the process.

The problem with all the proposals to bring in additional layers and variations to the Debian release architecture is that they add more conditions that have to be met in order to reach the final Stable version. And the bitch that everyone has against Debian is that it takes forever to reach a Stable release.

At each level (Unstable, Testing, Stable) you end up hinging pachages upon dependencies of other packages. With 16,000+ packages out there, running into dependencies is relatively common. When you end up with a dependency that hangs a segment of the release tree, it can take quite some time for that final dependency or combination of dependencies to clear a release branch. For example, perl 5.8 took a very long time to reach Testing. If you were to add another layer of release cycle to the process, you will only serve to delay reaching the Stable architecture.

Using this example, the effective outcome of adding another architectural layer to the process is this: It takes even longer for a package to reach Stable, making it even less likely that someone will use it, even for a server.

The net outcome is that your efforts to better Debian has made it only worse.

And regarding the arguement that there should should be a Workstation and a Server platform in parallel to each other, who is going to be the arbitrator of what constitutes packages for a Server and packages for a Workstation?

Creating a delineation between the two necessitates two classes of users, Administrators and Users. Linux doesn't have to work like that. If you believe that is should work like that, then I would prefer to see you choose a different distribution that supports this abitration of packages belonging to Workstations, Servers, Enterprise, and whatever.

One of the finest qualities that Debian has, in my opinion, is that there is no discrimination between the Server and Client user installations and they have what I believe is the finest release architecture designed and in practice to date, Open Source or Private.

When the discussion starts up on a place like Slashdot that someone needs to make Debian better by adding release variations to the process, I say to them that they need to better examine the performance of Debian and better understand what works and what does not work in Debian and use that information to cite your evidence supporting your proposals to fix Debian by adding variations to support.

FreeBSD

[Anonymous Coward]

I agree 100% with sp0rk173. Well said! Outdated and archaic doesn't mean is stable. As he already said, FreeBSD is released in a more often basis and is rock-stable :)

Woody has been updated along the way

[Anonymous Coward]

Up to and including r4. Most people who want to use Debian as a desktop use Testing anyway.

Release Names

[felixdzerzhinsky]

I think the actual names of each release are part of the problem. Someone else notes that its pretty hard to sell to management " want to run your business on Debian Unstable". How about: stable = server testing = testing unstable = desktop

Two words...

[Anonymous Coward]

...Ubun-tu

Re:These distros are not based on debian.

[Anonymous Coward]

Wow, you really have absolutely no clue what you're talking about. Go read the website.

A Knoppix easy run and HD install CD would work

[Anonymous Coward]

They can keep the DOS config install screens for those know what they want for each computer (ie sysadmins) but for normal users there should be a Knoppix CD with an install to HD option. Like Kanotix.

The users would be able to boot and say: wow this looks and works great, I want to have this on my harddisk next to the documents that are still on there and just click 'make it so' Then they would be number one of easy to use linux again. Others have shown them the way now put it all together.

pinning

[LordMyren]

debian's package management system includes the ability to pin. that is, to attach various repositories/package trees of varying distributions with varying priorities. all my systems start stable and quickly recieve a good number of testing grade packages. because of dependancies, this means my system is usually ~50/50 stable/testing. i then usually add some non-system-metal stuff from unstable like KDE, gnome, & staroffice.

i also have a long list of external package repositories from apt-get.org. some of my systems also track ubuntu packages as well. i run ubuntu's Xorg package set on my laptop (better acceleration, maybe one day working Xorg Suspend-To-Ram on my ancient ATI mobility ). it works perfectly transparently, including xcompmgr & all.

the nice thing about debian is it lets you mix and match very easily while resolving all dependencies very nicely & very cleanly. also, you can set up your own repository very easily to take a sample collection of packages from kingdom-come and mirror it so it looks like a somewhat cohesive single repository. with apt-build coming along nicely, you can even cleanly and efficiently maintain your own patched versions of packages as they evolve, making it easier to recompile all your programs for Heimdal kerberos instead of MIT, for classic example.

who gives a rat about stable? just pin what you need. debian distro is really about empowering the user to whatever ends with the most direct simplicity. distros like ubuntu are there for those who just want a single clean complete desktop distro.

Myren

Re:The last STABLE os form MS

[TheRealNecator]

I even can't remember the last stable OS from Microsoft.
Was there ever one?

Insightful?

[grolschie]

Well, first of all, I'd call Windows XP SP2 their latest release.

Well then, if you are going to count minor versions of the same release, then the latest update to Debian 3.0 was released on January 1st, 2005! Call it "Service Pack 4" if you like, we call it Debian GNU/Linux 3.0r4.

Running a 2002 release of Windows XP doesn't prevent you from installing the lastest version of Mozilla, Firefox or . The version of Mozilla in Debian stable is currently 1.0.0, and Firefox isn't even there!

Get real! Do you install Mozilla, Firefox, etc from WindowsUpdate.com? I think not! Nothing stops you installing the above mentioned software on either Debian "stable" or Windows XP. As with Windows, Debian "stable" users can download it from mozilla.org or other 3rd party sites.

Better terminology needed

[cupraman]

Debian has always made a problem for itself by using 'stable' as a version description. It's fine if you know that 'stable' means 'not likely to change much', but to most users the word implies that all other versions are 'unstable' which make them think that it's likely to crash a lot. I think a more relevant description would be 'static'.

All servers I install are Debian and initially I used stable but now I use testing and have not had a single problem.

For servers, Debian's great. For desktop, it's still great except that you use Knoppix or Ubuntu instead which take care of providing the latest and greatest package versions. Underneath they're still good old rock-solid Debian!

More Releases

[dancingmad]

As a hard up typical Slashdot reader, I too feel I need more releases.

heheheh

[halcyon1234]

"We need to release more often"

{giggle}

Really?

[Compact Dick]

Watson: No shit, Sherlock.

Re:If it's stable, it doesn't need to be updatedOf

[_Hellfire_]

Can I install any .deb package on Ubuntu without possibly causing binary version problems? Similarly, can I build a package on Ubuntu, give it to a Debian user, and be sure that it'll work properly on their system?

Actually...Yes. Yes you can

In fact the system I'm writing this on is Ubuntu Warty and I have the Debian Sarge repositories loaded in my sources list. I've got quite a few Debian packages loaded on my system with no breakage whatsoever. I've heard people refer to this type of setup as "Debuntian".

I wouldn't do anything stupid like apt-get upgrade (I comment out the Debian stuff for that) but for installing specific packages you're pretty safe.

yeah, last week

[Anonymous Coward]

Mozilla 1.0 is the bomb.

Re:If it's stable, it doesn't need to be updatedOf

[micheas]

Having just migrated from Ubuntu to sid I was very supprised at how well this went.

Ubuntu has upped the version of each package from the debian version by "ubuntu" this means that the next package compiled from a more recent version will replace it. It seems to follow debian policy as well.

This is a far cry from Knoppix, which I personally wish would remove the install to hard disk option. Knoppix is so far from debian in versioning that upgrading to debian is a horrific pain (there was a base package that had it's version promoted from 0.2.x to 2.x so when 0.2.78 comes out it doesn't get installed. I can't remember the package, but it was a core, or almost core package.

That said, I wouldn't develop from Ubuntu, but from sid, as Ubuntu is a fork that may or may not get it's changes into sid, and has stated that it will sync back with sid. Also sid isn't as nasty as one might think, because there is the experimental branch, that depends on sid, that gets most major changes and first (KDE4, GCC4, Gnome3, etc.)

My uninformed opinion on the matter.

Re:If it's stable, it doesn't need to be updatedOf

[Meumeu]

I can see the need for keeping ahead of security bugs, but to change for change's sake is just silly.

Mmm, drivers maybe?.

Debian's role

[salesgeek]

The best way to run Debian is not to run Debian. It is to pick one of the better derivative distros like Ubuntu, Mephis, Libranet, etc and go. Debian even in it's current imperfect form has generated some amazing distros - some of which make the case for Linux on the desktop and some of which make fantastic servers. Debian itself is difficult to install compared to some of it's derivatives and a little more difficult to manage.

Debian is fueling an incredible ammount of invention and innovation, and I for one am happy to have benefited from the project.

Misleading Nomenclature

[ajs318]

Debian is a victim of its own success.

It's an absolutely massive project. There are about ten thousand packages, all including metadata for full automatic dependency checking and resolution. Each of these packages is available for each of a dozen architectures, and there is consistency across all platforms. Debian is Debian; whether it's running on an Intel, a PPC, a Sparc, an ARM or whatever. The user need not know what lies beneath the skin of the machine; the procedure for doing something should be absolutely the same whatever is inside.

For a project of that sheer size to work, it's pretty much got to be ruled over with an iron fist -- if not literally, then those involved have to act as though it were so.

Woody is out-of-date for desktops; I don't think there is any question of that. KDE 2.2? Hello? And it's not exactly up to the minute for servers, either: it's still pushing Apache 1.3, for crying out loud!

The real problem stems from the fact that before a package can be accepted into the Stable release, it has to be shown to be bug-free on each of twelve architectures. So if it segfaults on a steam-powered toaster, it can't be deemed fit to run on an 80386.

But that's just the ideal for the Stable distribution. There are two other Debian distributions, Testing and Unstable. Whenever someone creates a brand-new .deb package, it goes into Unstable. The rules are, if you run packages from Unstable, and they break, you don't bitch: you fix them, or you keep your trap shut, but you don't bitch. Once a package has been in Unstable for awhile, it can go to Testing. When the project leaders are satisfied that the current state of the Testing distribution satisfies all the criteria and is fit to call Stable, then a new Stable distribution is born.

Testing is actually the Debian distribution you probably really want to be running if you have an 80386-type machine. Yes, security updates get ported into Stable in good time; but Testing probably has newer versions of packages anyway which are likely to have the security patch in by default. It's safe to run on servers iff you read the news and you know how to apply a patch and compile a package from source. {And if you don't, then what the hell are you doing running a server?} But Unstable is actually quite reasonable. I've found it to be no worse than Fedora or Mandrake: any problems I've had with packages not installing or not co-operating turned out to be due to mis-specified dependencies, requiring cunning use of manual override and package searches. So no worse than any RPM distro there :) It's not the packages themselves that are unstable; rather, the versions are unstable, simply because the maintainers keep putting in new versions as soon as the .debs are put together. I wouldn't run it on a server; but on my laptop, which is behind a firewall, it works very well, and I'm also using it on my work desktop {an AMD64}. All that being said, I am tempted to try Kubuntu -- it's just like Ubuntu but with a KDE desktop {sorry, but despite my best efforts, I really can't get to grips with GNOME}.

It's also worth remembering that every Debian-derivative -- Ubuntu, Linspire and so forth -- started out as a copy of the Unstable tree.

Re:Misleading Nomenclature

[damiam]

Yes, security updates get ported into Stable in good time; but Testing probably has newer versions of packages anyway which are likely to have the security patch in by default.

That's misleading. There's a dedicated security team for Debian stable, and they backport fixes quite quickly, sometimes even before the fixed version is released upstream. No such team exists for testing, and an updated package will take, at minimum, two weeks to get into testing after hitting unstable (and sometimes much longer).

Re:Misleading Nomenclature

[Anonymous Coward]

I guess following should have been the name for debian versions:

unstable latest
testing current
release release

I can bet that debian testing is far far more stable that stable version of number of linux distros. Better if we don't bring into picture CTRL-ALT_DEL operating systems here.

Re:The last STABLE os form MS

[a_n_d_e_r_s]

Remember back in the mid-80ies - there was a OS from Ms that was called Xenix - based on a obscure OS from Bell Labs called UNIX - Xenix was selected as the main multi-user OS for IBMs latest AT based ranged of computers at that time.

Well it was some time ago.

That wouldn't be really true...

[Anonymous Coward]

I mean, the difference between "stable" "testing" and "unstable" isn't just in haow updated the software is. "Testing" and "unstable" don't get official and prompt security updates as "stable does, and also testing can be pretty "unstable" if you happen to try an update on (or a fresh install of a snapshot from) the wrong day... I know, I **** up my system this way! ;-)

FreeBSD all the way!

[Anonymous Coward]

It's a rock-stable and up-to-date *true Unix*!

So two Linux advocates walk into a bar....

[dspisak]

one says to the other "Man, I hate having to wait so long for my distro to update!". The other then asks the man "What distro is that?". "Gentoo" he replies. The other advocate chuckles "At least your distro wasn't last released 3 years ago".

George Bush: We need more oil

[Anonymous Coward]

Since we're in the business of reporting the obvious....

packages

[diegocgteleline.es]

FreeBSD maintains the same kind of stability WITH a more current release schedule.

FreeBSD doesn't have packages for most of things and for a few platforms. Compare that with releasing 12000 packages (14 CDs, IIRC?) for 10-12 architectures. Is not that FreeBSD sucks, they work great, but is not fair to compare two things that are not really the same. And BTW, the 4.X -> 5.3 step has not been exactly "fun".

(and don't come saying "this is the proof that ports > packages. Time has showed everybody that packages are valuable, I don't want to start recompiling libc or X.org because of a critical security bug when I have a spike load, ok?)

Re:packages

[Zemplar]

Fine, don't use ports on FreeBSD if you don't want to. A simple "pkg_add -r [package name]" works great on FreeBSD, and arguably better than apt-get.

Re:packages

[Anonymous Coward]

Yeah, I love the way pkg_add, etc., does dependency
checking without errors, and gives you options to
install dependencies. I was installing from the
most recent "STABLE" and I got errors installing
packages that conflicted and it didn't give me
any option of fixing them, and then I had to go
back and fix it manually. The packaging system
is PRIMITIVE compared to apt-get, but ports ARE
awesome.

Re:packages

[Necrotica]

FreeBSD doesn't have packages for most of things and for a few platforms. Compare that with releasing 12000 packages (14 CDs, IIRC?) for 10-12 architectures.

Wrong. I just cvsup'd my trusty FreeBSD 5.3 server sitting right beside me and it has 12,611 ports available.

Re:packages

[mindstrm]

Although that's true, it ignores the reality of the situation.

For a huge majority of tasks that the business world wants to use these systems for, BSD does the job nicely. Yes, Debian takes forever to release a new -stable because of the huge number of tasks and platforms. Nobody is calling them stupid or lazy.

THe fact remains: In the current state of affairs, organisations want to use current software, with new features and all that. Debian stable is years behind the curve now, and that's a LONG time in this business.

Yes, it's stable, and that's appreciated, but stable software that's outdated doesn't help.

Re:packages

[Billly Gates]

agreed.

I do not bother with the packages on FreeBSD anymore. They are always broken. I can never figure out what is wrong. Ports too if you cvsup alot for security fixes. I admit I do start with ports and then use some packages. Wrong compiler settings with dependancies can cause alot of strange core dumps and bugs.

Its too easy to brake something but ports are better. Ports are not perfect either if you cvsup. You can read the CHANGES file all you want but I could not get java to work for over a month. The port required an old version of a patch file that was no longer being served.

I switched back to Windows because I needed java.

Re:packages

[Anonymous Coward]

I'm the same person who made the original comment.
I had the same problem. Java was uninstalleable.
I tried everything in there and believe it or not
the only way I was able to get it to work, was to
actually change a linux package and install it on
freebsd. Yeah and the core dumps are bad. Plus
why by default do they put errors unto the main
terminal screen from other processes?

I think Debian is much better for package
management, though I think the stable branch should
be MUCH more up-to-date. But otherwise debian
is a really good integrated system, i.e. when you
install a package, it is updated on all of your
installed window managers, etc. Really nice
integration.

Debian: inactive distro

[Donny Smith]

Alright, I'm just kidding here based on distroWatch.com definition ;-)
" (i.e haven't released a new version in over 2 years and their web sites don't give indication of work in progress)."
http://distrowatch.com/stats.php
Onc e upon a time Debian was the most respected and popular Linux distro. I still love it but its popularity has dwindled under the onslaught of new distributions. Who would have thought...

By the way, the popularity stats are very informative:
http://distrowatch.com/stats.php?sec tion=popularit y
Fedora is going down - despite Red Hat's reanimation procedures, Debian too, while Ubuntu and CentOS are going up...

When was this year?

[Kaseijin]

By default Portage does not update dependencies unless necessary (as specified by the dependent ebuild), emerge is run with -D/--deep, or (IIRC) the old version has been removed from the Portage tree and is no longer supported. Packages can be forced with -O/--nodeps and pinned with /etc/portage/package.mask; as I'm sure you discovered, changes made to /usr/portage/ are obliterated during the next sync. With the exception of /etc/portage/package.mask, which was added early last year, it's been this way since early days. I'm trying to figure out a time when this stuff wasn't documented, wouldn't have been answered in five minutes on the forums or mailing lists, and yet was recent enough that you could expect your experience to be relevant, but I can't come up with anything.

I'm no philosopher

[Kaseijin]

BSD is built for installing by source, Linux is built for installing by pre-compiled packages.

What does that make Mac OS X and Gentoo?

Re:I'm no philosopher

[linguae]

Well, they are exceptions to the general rule. Mac OS X has a BSD core, but it isn't generally used like a BSD. The software that Mac users use is generally closed-source, proprietary software, which is why Apple didn't include a ports system in OS X (because its traditional userbase has no use for it); you'll have to download Fink or Darwin Ports separately. Plus, even though I'm very fond of Mac OS X, Mac OS X is used very differently than FreeBSD (or any other BSD) is used, and I prefer the BSD method.

Gentoo adopted a variation of the FreeBSD Ports system called Portage, which I hear is very nice. I also like how Gentoo (like Debian) doesn't just dump packages on you that you don't want; you get your basic userland and you emerge/apt-get your way into getting the environment that you want. In fact, if I had to switch back to Linux, I'd use Gentoo or Debian unstable. They're like BSD, except with the Linux kernel and the GNU userland.

Re:If it's stable, it doesn't need to be updatedOf

[cperciva]

This issue affects other operating systems, not just Debian. It will be disclosed according to the schedule agreed upon with the other vendors; I'm not going to disclose it early just to spite Debian.

It would be nice, however, if the Debian security team were aware of this issue before the disclosure date rather than after.

Release more often....

[Anonymous Coward]

I think geeks release quite often...

Especially when watching Debian Does Dallas...

Debian release process difficult

[sagneta]

The current debian release process worked at one time. PResently however a release is much larger as it encompases about every concievable piece of software you might wish to install. That's a tall order. I think perhaps the Gentoo system, which I now use exclusivly, might have the best idea. Distribute a base system. Allow the user to include/exclude and upgrade whatever they want as they move along. Their call. I have emerged items (download and compile) and then reverted them just as quickly if I found an issue. I think that is the only way given the huge amount of software available and the possible library interactions. I think Debian biting off a lot. I respect them greatly for trying however.

There's a better chance

[thenefariousone]

of apple releasing a two button mouse!

Re:If it's stable, it doesn't need to be updatedOf

[grendelkhan]

I have plenty of sarge and sid debs installed on ubuntu hoary - and I even use mariliat when I apt-get upgrade and to grab certain non free pacakges I want. No problems whatsoever.

Maybe testing is to up-to-date?

[Anonymous Coward]

I for one, would be happy with sarge releasing with a 2.4.x kernel only, apache 1.3.x and KDE 3.1.x if it happened today. Maybe they try to release too up-to-date software?

Re:If it's stable, it doesn't need to be updatedOf

[Zemplar]

I'd be interested in trying this. Did you just change your respositories in Ubuntu to reflect Debian sid? Was it really that easy or did you have other problems?

Missing the point

[Anonymous Coward]

Honestly, who uses stable as a desktop system? I use it on a server and it is PERFECT for that role. It has virtually no package updates because it is rock solid already and it's easy to get security updates.
I have used debian unstable on my desktop but i switched to gentoo because unstable was just that - too unstable.
IMO debian unstable is updated regularly enough, its just sometimes some of the packages are flaky. They need to work on a way of putting the packages going into unstable under a little more scrutiny.

Release more often, ohh sir.

[davro]

Debian, Rules Servers. Ubuntu, Rules Desktop.

Something similar to gentoo?

[cheezemonkhai]

No i'm not talking about the spoilers and alloy wheels.

Seriously Gentoo has x86(stable) & ~x86(testing) and there equivalents for each platform, and different packages are considered stable or not on a per platform basis.

SUrely something like this for debian, with prehaps core architectures being released together (eg x86, ppc & Alpha).

Also how about Stable, Release and Testing/unstable as better names.

Testing & unstable all sound like they are broken, when infact testing usually isn't.

And now, here come the apologists

[magnwa]

OHH! But Debian stable is "STABLE." It's rock solid. Why would you need anything else? I mean, come on, you want stable, right?

I used to be into Debian, until I realized that stable also meant featureless.. and that they would continually be behind on the Stable release. I'm sorry, but I can't use old versions of software , even on the server side, when newer versions are more optimized for what I'm doing.

Re:And now, here come the apologists

[PigleT]

Considered testing? Especially, with what appears to be security-updates to testing as well these days?

I know, in general terms it makes a mockery of the design goal in having stable,testing,unstable the way they are: but in practice, something that I can use to apply updates weekly or security ones daily, with a continual smooth one-or-two-packages-a-fortnight rate for life, is really nice.

Re:And now, here come the apologists

[magnwa]

I've considered testing, but the problem with testing is that it still doesn't get nearly the support or help that stable gets. Something goes wrong in testing and I'm told "Well, it IS testing."

Nah.

I'd rather just use FreeBSD or something that is supported instead of relatively ignored..

Has anybody got metrics?

[synotia]

When considering optimizing software, the first thing you should do is get metrics, use a profiler, find out exactly what's slow, and what needs to be fast.

It seems to me that before *optimising* the Debian release strategy, either by removing architectures and/or shortening the release cycle, one should consider the metrics:
* How many servers/workstations use stable, testing and unstable (and for what reasons).
* How many users of each architecture are there? (and for each of stable/testing/unstable)

With this information, the Debian crew can make informed decisions on where their focus should be, what needs to be fixed, and who for. We might find, for example, that Debian stable (as it stands) is used quite heavily, and to keep good support for it (call it "really-stable" or something). We might also find that the majority of Sid users are on x86/PPC/AMD64 and to scrap support for other architectures in Sid.

Who can say until we have the metrics.

Re:Has anybody got metrics?

[ccwaterz]

This might answer some of your questions:

http://popcon.debian.org/

But how is that different from Ubuntu?

[khasim]

The essential reason to choose UserLinux is that it avoids the conflict between Open Source and producing income that all commercial distributions develop while supporting the enterprise.

And how is that different from Ubuntu?

The policy of the UserLinux project is for all development to be carried out within Debian, not within our own repository. Customers can take a much greater role because the Debian organization admits them fairly.

Again, how is that different from Ubuntu?

Of course, the long release delay has made something of a fool of me - because so far we've only proven that this non-profit can't get it together to make a release.

You will always have that problem.

The solution is to base your releases off of a SUB-SET of Debian. As Ubuntu does with Gnome. As UserLinux was supposed to be based upon Gnome.

Rather than delaying while some package you won't even be including waits for its maintainer to fix it, you should be shipping your release.

Remember, one of the key aspects of UserLinux was that it would NOT ship with two dozen text editors, one dozen image viewers, etc.

Look at the packages, not whether Sarge is released or not.

Re:If it's stable, it doesn't need to be updatedOf

[micheas]

I had a small problem, but if I remember correctly it was something like evolution and evolution-server were in a funky state in sid, I think I solved the problem by uninstalling evolution and then installing gnome which reinstalled evolution.

I might not have had to do that if I had used apt-get dist-upgrade instead of apt-get upgrade.

But I'm not sure, I do remember that the whole proces was about two hours, including download time (dsl at 1.5k download in the United States).

What "support" means.

[khasim]

It's hard for a support company to say "I support Debian" when there's no standard set of software to support.

Actually, it is very easy. The same as a support company can say "we support Red Hat".

What "support" means will vary with different vendors.

The local shop might "support" Red Hat because one of their techs uses it and likes it and has a book on it.

Oracle might "support" Red Hat because they have a formal contract with Red Hat and technicians on both sides who work to duplicate and resolve issues.

The local shop would be able to recommend NIC's and sound cards and such and be able to configure them and help you get your CD-burner working. But they wouldn't be writing code to provide you with new features in that app.

Oracle would be able to find the error in the code that caused your database to hog memory under certain circumstances and get the patch to Red Hat to be included in the next patch. But they wouldn't be helping you rip CD's.

All UserLinux was supposed to do was to pick the "best" (by Bruce's criteria) app in each category and focus development/support on those in order to provide a Free platform for users and businesses to port their software to in a business-friendly fashion.

The problem is that Bruce lost focus and is, instead, waiting on Debian to move Sarge to "stable". Debian supports almost every hardware platform out there and hundreds of apps. The problems slowing down Debian should NOT be affecting UserLinux.

15781 packages in sid-main-binary-i386

[Bishop]

grep -c "^Package: " /var/lib/apt/lists/ftp.debian.org_debian_dists_sid _main_binary-i386_Packages

We're listening ..

[stevey]

Although unless you could post a subject, or the mail account you mailed from it'd be hard to tell.

There are literally hundreds of messages going to the security@debian.org alias - and vendor sec also gets a lot of spam. This is one reason why sometimes I've lost things.

Of course that's likely not to be what's happened to yours, maybe it just got queued up behind all the other things that we're working on.

Does that help?

Feel free to ping me with another copy if you like.. Actually forget I said that, I've just found your mail and I've personally not responded because of the lack of details - we already publish our private keys on our webpage so asking for them again is extra work when we've got lots to do.

Vendor-sec / Debian can do lots of things your particular case you might think of a more appropriate person to pass it onto - obviously I don't wanna give details here.. Grr.

Re:We're listening ..

[cperciva]

I've just found your mail and I've personally not responded because of the lack of details - we already publish our private keys on our webpage so asking for them again is extra work when we've got lots to do.

In the case of my initial vendor-sec posting, well, I don't know who exactly is on that list, so I couldn't search around for everybody's PGP keys -- thus my comment of "if you want to know the details, send me your PGP key". (Also, it's much easier for you to paste your pgp key into an email than it is for me to find your pgp key -- especially when you consider that I'm trying to promptly inform over a dozen different security teams.)

But now that I know you're awake, and now that a sibling post has linked a FAQ which lists your key, I've emailed you all the details.

PPC alternatives?

[grikdog]

Have you noticed how people have started saying Apple has great hardware? I was going to put Debian on my iBook this weekend, but this thread has coldcocked that idea. A shame too because the PPC iso's downloaded, checksummed and burned perfectly. But I need a real desktop, I need OpenOffice.org and I need a Linux bootable from (and installable to) my Firewire drives (not my IDE internal drive). Maybe my path to Nirvana is to stop applying OS X updates and wait for the corporate open source money oozers (COSMO) to fund a rational distro. The last time I looked (admittedly it's bit awhile), even YDL was recognizing my trackpad, but wishing it would go away.

Re:PPC alternatives?

[poofyhairguy82]

PPC alternatives?

You want Ubuntu.

Re:PPC alternatives?

[grikdog]

Ubuntu is a Debian "wisdom distro" -- i.e., an attempt to pre-select features for an idealized (or dumbed down) user. Ironically, I do appreciate the attempt made on my behalf, but my first and foremost desideratum is the ability to install to, then boot from, external media , e.g., a partition on a Firewire drive like SmartDisk's CrossFire. Since Ubuntu can boot from CD -- heck, since Mac OS X can boot from a SmartDisk, for that matter! -- I know what I'm asking is feasible. But it's just not an option for home users. I suspect the real reason is, nobody at Debian has the time to make it happen, but there will be a population explosion in Linux User Kingdom just as soon as installing to and booting from iPod (to name one pointlessly expensive external hd) becomes possible.

debian

[XO]

From someone who's been running debian on his box for about 2 years now, and every day hating it more and more....

If you don't have your apt-get set to automatically update you to everything in 'unstable' and 'experimental', your system is automatically unusably out of date.

At least for a personal box.

Debian stable should be used for servers.
Debian shouldn't be used for anything else.

Debian Woody is too old for servers

[Bishop]

We don't need the newest KDE or Gnome, but newer versions of PHP, Perl, Postgresql, and Samba would be nice. A new 2.4 kernel would be nice too.

It would be less work for us to move to a new stable Debian every 12 months then to deploy servers that are a hodge podge of Debian Woody and backports, or Debian/testing.

Re:If it's stable, it doesn't need to be updatedOf

[snorklewacker]

> I can see the need for keeping ahead of security bugs, but to change for change's sake is just silly.

Of course it is. But when it comes time that you have to upgrade a particular major system, and you find that dozens of packages from the rest of your system have to be dragged kicking and screaming with it, it's nice to know that the transition won't be so abrupt and brutal that it breaks god knows what because the packages from today don't integrate with what's left from 3, going on 4 years ago. Source distributions have less problems with this, but they're not immune either. Debian does an incredible job at keeping the migration path smooth, but the longer they wait, the worse it's going to be.

Re:If it's stable, it doesn't need to be updatedOf

[Anonymous Coward]

It's been three weeks since your contact with vendor-sec and nothing's happened. You fulfilled your ethical obligation to the vendors, now it's time to fulfill it to the users: it's full-disclosure time.

I read down, and the debian guy says there was a lack of details, but they didn't even fire off a $MOREDETAIL boilerplate response. Maybe the community will be good enough to fill in the details then...

I am actually happy

[Anonymous Coward]

I am responsible for a medium sized network (12 companies, 50% of them in remote locations, about 300 workstations), and we use debian woody for everything network related. I am actually very happy with the release cycle because it lets me keep the TCO down. How? The time i have to prepare the next generation production environment is much longer, therefore i can prepare it without having to hire more people. The same people that manage the production environment are building the next generation using sarge so they are getting hands-on training simultaneously. Anyway, if there is a latest-greatest software that we *need* and can't wait till the next production environment cycle we can always build it from source.

The point is, i believe the release cycle is excelent for my servers. On my desktop i use either sarge (work desktop) or ubuntu (work laptop and home desktop).

P.S. forgive my english, it is not my 1st language.

Debian should fork itself

[sadiklis]

Go with Server and Client releases. Server would be updated less frequently, trying to go in sync with major server software releases (Apache, PostgreSQL, etc.). Client version could be released more often... and it should probably be sub-forked further into Gnome/D and KDE/D projects releasing themselves in sync with Gnome and KDE releases.

Or not...

[Anonymous Coward]

On what basis do you base your statement?

For an absolute newbie migrating from Windows, Mepis is a very good choice. Knoppix is also a good choice but Mepis is a bit better for the newbie category.

The problem with Mepis is as soon as you change a source in your sources list, as soon as you decide not to use the Mepis "user utilities" or other Mepis specific scripts or administrative tools, then the first question is, why did you do that, and statement to follow is, its not Mepis anymore, you are on your own. After that, try getting help in debian-user channel on irc if they find out that you are running Mepis and not debian proper.

Why deviate from the Mepis specific admin utilities? Why should one have to answer that question? I'll do it anyway. Mepis loads kde's version of APT-Synaptic. Some months ago, I ran into a bug that was traced to kde's apt-synaptic (can't remember what they call it right now. I tried googling for the error but came up empty handed. Someone helped me out and we traced the problem to specific lines in the kde app which must have been fixed by now because it was reporting specific line numbers in the error. Suffice it to say that no package updates would work because of the error. Installing synaptic and related apt applications fixed the problem. Some other changes were necessary in the sources list to keep my system working. One of the warnings in Mepis is if you change the sources list without using the Mepis specific utilities, the changes will be overwritten with the defaults the next time you start the Mepis utilities. So as soon as you start to deviate, you basically are on your own. As has been made clear in the message boards and in irc. Mepis irc channels are a little more subtle, you basically get question after question as to why you deviated from Mepis proper, and how can you expect to get help after doing that. Not to knock this answer, because it is coming from other users who love Mepis and are very helpful to newbies, but aren't experts beyond Mepis so can't help beyond Mepis. The help received is absolutely appreciated, and I'll continue to recommend Mepis to Linux newbies migrating from Windows, but because of the problems encountered when you change something to more closely track Debian on which Mepis is based, I can't recommend it for business users or for those that know they will make some changes to more closely follow debian at some later date and don't want to lose all support that Linux normally provides.

If you get the 3rd degree while asking for help or can't get help from the Mepis channels, and have to hide the fact that you are running Mepis while trying to get help in debian-user, what good is the distro, except to a newbie who wouldn't make any changes to a distro based on a different distro where tens of thousands of packages are available? Are Red Hat users denied help if they use Red Carpet instead? Are Suse users denied help if they use apt or synaptic instead of yast for updates? Or some other app for something administrative instead of yast?

Mepis may rock for newbies. For everyone else, it's a big Linux world out there.

Re:If it's stable, it doesn't need to be updatedOf

[metamatic]

Or try MEPIS, my current favorite distribution.

It's basically Debian Unstable, plus a kernel pre-configured so everything just works (e.g. Centrino wireless), plus better hardware detection, plus Mozilla with all the usual plugins.

All on a bootable KNOPPIX-like CD, so you can see if it'll work with your hardware before going ahead with the install.

Debian has no competition in a server environment.

[Hohlraum]

None. Nada. Zilch. I was hardcore Redhat since pre RH3 (before that I was a Slackware zealot). Once they released the new installer for Debian I decided to give it a run. HOLY CRAP, what an amazing OS. Rock solid. All the packages "just work" with one another 99% of the time. Pretty amazing OS, if you haven't tried Sarge with the new installer you are really missing out.

As far as the desktop, after trying FC4t1 and noticing practically zero actual technological advancement I decided to try out Ubuntu. Nice desktop OS. Once 5 is out with xorg as the default xserver I'm switching my workstations at work over to it from Fedora.

Run stable/testing simultaneously: use "pinning"

[Anonymous Coward]

Testing is actually the Debian distribution you probably really want to be running if you have an 80386-type machine.

For servers, I would disagree. Instead, you want to use pinning to run a mostly stable system, with the specific server packages of interest pulled in from testing. For example, say you want to set up a webserver running Apache 2.x instead of Apache 1.3. Then you simply pull in the apache2 package from testing. All dependencies will be met from testing as well, so a small portion of your system will end up being from testing. But most of your system will be rock solid from stable. You can still run apt-get dist-upgrade periodically without too much concern. For more info, just search google on "apt pinning".

Re:If it's stable, it doesn't need to be updatedOf

[PxT]

> I've e-mailed debian-security-private directly

From the FAQ - the proper place to send it is to security@debian.org

It doesn't matter

[zogger]

You are agreeing with me in a way. I DID say make it easy for this *everyone* guy to have a distro. Look at distrowatch, every-freekin-day there's some new "distros" being "developed" based on Debian, because people want their "own" version of computing reality, but then they all need "community support" and "paypal" contributions. It's the same dang wheel is being reinvented constantly, and there's no absolute need for that if it was taken to the next logical step. So I am saying just admit that this is apparently what a lot of people want, and rethink how it's done, just make it wicked mucho easier simpler to "roll your own" for anyone who wants that. Joe Blow wants mostly every game in existence and runs the newest heavy video stuff, swell, mash a few choice buttons, he downloads or gets shipped 8exactly that*. suzy schmoo wants a ton of biz apps and it has to be uber secure and have all sorts of encryption and whatnot, poof, she got it. ma and Pa six pack want an educational deal for the kids and some light gaming and casual web surfing, poof, they enter the fields, make their choices and it gets spit out, exactly what they want, no more, no less. That's something no one else is doing, not any of the big guys making money at it that I mentioned. It could be the way to make this Linux OS thing really take off on the consumer desktop. And if debian don't want to do it, there's maybe some other folks who could see the potential here. I bet quite a few (million) people would be willing to pop ten or 20 bucks for an OS they KNEW 100% in advance would be exactly what they wanted, would work flawlessly right out of the box, and would be guaranteed to work on the hardware they got, without having a ton of kruft in the install they got absolutely no need or desire for. Just a biz and developer model of doing this I haven't seen really discussed yet, like I said "something different".

Moderation comment

[mr_tap]

My parent post got moderated insightful, funny and flamebait - it probably deserved the second two more than the first :)

Response

[Anonymous Coward]

Debian Users: No Shit!

Re:If it's stable, it doesn't need to be updatedOf

[giantsfan89]

Packages within Debian are compatable, but not always drop-in replacements.

Some packages are split into smaller packages with the next version. You could have package X 2.6, which includes a lot of stuff, and then in version 3.0, X gets split into X-doc 3.0, libX 3.0, and X 3.0. Obviously, you can't use X 3.0 from testing and put it into stable, because it relies on libX 3.0, which is not in stable. You'd need to copy over libX 3.0 and X 3.0.

If package Y 4.5 depends on package Z 4.5, again you'd need the same thing. You'd have to copy all the dependent packages down. And hope they work. Or look for a backport.

I recently installed Debian Stable on a server with an Intel e1000 NIC in it. The kernel in the stable install didn't include support for my kernel. So I had to grab about 10 packages from testing, including a later kernel and all of it's dependent packages, before I could get networking support.

Making the point explicitly

[Kaseijin]

Neither BSD nor Linux are "built for" one kind of package or the other; individual distributions are built around specific package managers. Yes, Portage was inspired by FreeBSD ports; if that makes Gentoo Linux a BSD, its kernel makes it Minix.

Mac OS X has a BSD core, but it isn't generally used like a BSD.

A BSD, like anything else, is used however people use it. FreeBSD 5 isn't exactly 1BSD itself, and the developers consider Mac OS X part of the family.