Ian Murdock: Debian "Missing a Big Opportunity"

Natester writes "While Debian struggles to get its next release (Etch) out the door, the project's founder, Ian Murdock, has spoken out about politics, the lack of firm leadership, and Ubuntu's meteoric rise in prominence. Murdock believes that Debian is "process run amok" — nobody feels empowered to make decisions, leading to the sluggish rate of progress."

Firm Leadership

Sometimes you need firm leadership to make decisions rather than stagnate by trying to please everyone all the time and doing nothing.

Re:Firm Leadership

I thought Open Source was about each of us having it our way instead. Compilation without representation is tyranny!

Dazed & Confused

The immaturity posed by this comment being modded insightful is just sad.

While the sheer number of packages in the Debian repository is awesome, you are confusing _choices_ with a lack of focus. Debian's NOT pleasing everyone. They can't.

There will be many out there probably like you who are reassured with a self-contained environment that a Ubuntu provides. They have x number of apps configured a specific way that works okay in many situations but is really poor if more or something different is required.

In my business, I need to have log reports formatted a specific way. Well, there just so happens the log analysis package I use is in ubuntu's "universe." e.g. should work, but it's not an official distro package. Good news, it's quite well supported in debian's main package repo.

This is why ubuntu is kind of like AOL way back in the day or Microsoft server apps for good system administrators. Once you figure it out, you realize the limitations and move on.

When you are ready, Debian's there. Still Free.

Re:Firm Leadership

It still has a place as a meta-distribution. Gentoo has been fairly successful on that model.

Re:Firm Leadership

Open source is about choice, but Debian is about providing a distro that does what most of their users are supposed to want. It's still a tyranny - the tyrany of democracy.

Re:Firm Leadership

If we all wanted it our own way, we wouldn't need Open Source. We'd just write all of our software ourselves. Really, everybody wants to tweak a couple of things slightly, and leave the vast majority of everything entirely unchanged. The point of Open Source is that we can make just a couple of slight modifications, and we don't need to start from scratch. But this, then, relies on there being software that is close to solving our problems.

It's the bikeshed problem: everybody agrees that we want a bikeshed, and that it needs to be painted to keep from rotting, and nobody has a particular color it has to be, but nobody feels empowered to go out and buy paint, in case somebody turns out to be deeply offended by the color choice. Someone needs to take the initiative and pick something, and if anyone turns out to care, they can repaint it later.

Re:Firm Leadership

And then when someone finally does go ahead and pick a colour it's the ugliest brown you've ever seen. Yeesh, what were they thinking?

Re:Firm Leadership

I'm skeptical as to whether a firm leader would be able to keep all those developers together working on Debian. It may work for Ubuntu, but Ubuntu has much fewer developers. And they get paid.

If I were a coder I would be much more likely to volunteer my time to Debian than Ubuntu. I'd rather donate to a fully democratic system than a benevolent dictatorship. And if I'm already coding for a project and they decide that they're going to "empower" someone to ultimately say what goes and what doesn't, I'd be more likely to quit contributing code.

Re:Firm Leadership

Dam I wanted to mod in this thread but I can't mod you fairly: "fair point but I really disagree" (underrated I suppose?)

Anyway, I feel that a benevolent dictatorship is actually the prize winner in the dev cycle. Once you get into the "please everyone, get a majority vote" mode of operation you run into endless debate as one side tries to convince the other side of the merits of their idea(s). Now I want a unanimous decision on a jury, but for a distro I want a clear path and direction. The dictatorship forces that path to exist. While I may want the path a different color, so long as it's going the same general direction I am then I'm OK with it.

The direction I'm interested in is a mainstream linux that I can deploy on joe sixpack's computer. I want a linux that is as friendly as OSX, and as compatible with hardware as Windows. I want a distro for the masses, and thus while you are entitled to fork it and tweak it, I think the main tree should be ruled by an authoritarian, rather than a committee.

I also think that the "open market" will decide this for us. Suppliers (donors of code and money) will "sell" to their ideals and buyers will install to their needs.
-nB

Re:Firm Leadership

The direction I'm interested in is a mainstream linux that I can deploy on joe sixpack's computer.

That niche is well supplied by Ubuntu. If you think the space needs competition to keep things fresh, it has it: Linspire and Mandriva.

Debian *doesn't* need to target that niche at all. The existing policy of slow, steady progress and periodic rock-solid server releases produces a distro that's an excellent basis for projects ranging from Ubuntu to Embedded Debian to build from. There are occasionally cases where Debian is a bit slower than it should be (multiarch for example), but that doesn't mean that what Debian is should be thrown away to make an Ubuntu replica.

Re:Firm Leadership

I think we have to be careful about drawing analogies and using terms national politics when we talk about group dynamics. We shouldn't let our justifiable aversion to political dictatorship poison our attitude towards the idea of strong leadership, which is a completely different thing.

What we call "benevolent dictatorship" in a group like this means decisive leadership with power make its decisions stick within the confines of the project. This is a very different thing than a political dictatorship (benevolent or otherwise), because in national politics there are no practical limits to which the power of a dictator is confined.

Philosophically, strong leadership within a voluntary project is consistent with the status of an individual contributor as a rational being. The contributor has the practical option of "voting with his feet", by jumping to a different project. A wise leader who relies upon voluntary contributors keeps the best of them happier contributing than leaving. A true dictatorship is inconsistent with the status of an individual citizen as a rational being, because a citizen cannot vote with his feet. The dictator can choose to make any trivial dissent a matter of life and death, and choose in a completely arbitrary way.

I once heard Prince Bandar claim that absolute monarchy wasn't really different than democracy, only in a monarchy people "vote out" the government by taking to the streets. While this might be wise for abolute monarchs to bear in mind, it conveniently ignores the fact that the subject of a true monarchy must be willing to risk everthing, his life, the life of his friends and family, in order to act in accordance with his reason or conscience. It sidesteps the question of whether it is necessary or beneficial to make the exercise of individual reason a life or death decision.

In the Debian case, anybody unsatisfied with Debian can join a different Linux distro project. Not only that, they can walk away still in posession of their entire body of contribution, as well as the entire contribution of everybody else. The only things the can't take with them is the community (they'd have to build their own or convince others to move with them) and the name.

The irony here is that the apparent anarchy of the free software paradigm makes it possible to exercise extremely decisive leadership with little or no ethical risk. There is nothing a free software leader can do to a contributor, other than refuse to take his contributions.

The utter inability of a project leader to inflict meaningful harm on a contributor makes elaborate safeguards for the dignity of the individual redundant. That respect is built into the software development paradigm, not the organization.

It's very important not to confuse strong leadership with dictatorship. Leadership is securing the voluntary cooperation of individuals, sometimes to a course of action that the individual does not favor. While leadership might be useful in a dictatorship, the key job skill for a dictator is the infliction of fear.

Ian Murdock to join Sun

Hi all,

It's being announced today that I'm joining Sun as chief operating
platforms officer, which basically means I'll be in charge of Sun's
operating system strategy, spanning Solaris and Linux. I just posted the
announcement on my blog (http://ianmurdock.com/2007/03/19/joining-sun/),
and it'll likely be making the rounds soon. Just wanted to
make sure you heard the news directly from me and to introduce myself.

First things first: I'm a long time Linux user, developer, and advocate.
I founded Debian in 1993, co-founded a Linux distribution company called
Progeny in 1999, and most recently served as CTO of the new Linux
Foundation, where I was (and still am) chair of the LSB, the Linux
platform interoperability standard. I'm also a long time Sun fan.

As for what I'll be doing: While I'm coming in with some fairly formed
opinions about what Sun/Solaris/OpenSolaris ought to do (peruse my
blog a bit to learn more), I'm also a big believer in listening
before talking, and I have a lot of listening to do in the weeks
to come. So, please, feel free to drop me a line if you have
anything to tell me. And, please, be gentle while I get settled. :-)

Gotta get on a call in a few minutes. In the meantime, I just wanted
to say hello, and to make sure you heard the news directly from me.

Later,

-ian
--
Ian Murdock
http://ianmurdock.com/ [ianmurdock.com]

Wait a minute..

by Anonymous Coward

+

I'm a long time Linux user, developer, and advocate.

= WTF? Not every linux dev is a slashdot user?

"the wreck that McNeally created"

The "wreck McNeally created" went from startup to $18 billion on his watch -- and yeah, back down to $13 billion. As soon as you get that $13 billion company of your own going, I think you're safe to criticize McNealy for his failings. Heck, check in at two billion and we'll give you a listen.

He also correctly identified Microsoft as Sun's up-and-coming competitor years before anyone else got it, and then correctly identified that the level of anti-MSFT rhetoric was causing major problems and cleaned that up, netting Sun a nice $2B in the process. Maybe slow to get on the x86 bandwagon, but he got there, bringing back one of the industry's best system designers in the process. He groomed two successors, one of whom now seems to be the real deal, but in many cases is getting credit for a lot of things McNealy had already set in motion. (And the other one is off perhaps tanking another company -- maybe this is where the "wreckage" came from?)

Sun *is* "selling like it once did." It's the 3rd largest server vendor in the industry. It's the 5th largest x86 server vendor in the industry -- again, something McNealy set in motion.

There's a lot of things he did wrong, but there's a lot more he did right. Sun went from an engineering workstation company to the third company regularly mentioned in the same breath as HP and IBM, two much older and more well-established companies.

This is coming off like a gush about McNealy and Sun, but really, consider it more a rant against calling something a "wreck" when you have no idea what you're talking about. Get picked for the board at GE, then you get to talk about someone else being a "wreck."

It's sad

Unfortunately, Debian has suffered from a concatenation of problems this year. Dunc-tank (a scheme to pay some developers) sapped a lot of good-will and motivation, and made some developers actually work to hold back the release in protest, and as a result it's another "who knows when it'll happen" Debian release. There has been a lot of bickering on other topics - Debian should never hold face-to-face meetings, something bad always happens - and unfortunately the current DPL hasn't been able to rally the troops or lead effectively in any way I can see. I hope they recover, I think they are still our best hope among Linux distributions.

Bruce

Re:It's sad

Sorry for the naivete, but I don't plain understand the rationale behind the DuncTank failure.

I mean, even if I'm a non-paid developer, what's bad in having me collaborating with payed developers if it helps getting the work done? Isn't it a bit like the GSoC? People in GSoC-funded projects should whine and hold back releases because "hey, why is he paid and I am not?" I just don't understand it, but I don't know the exact story behind the Dunc Tank collapse, so I'd like some enlightenment.

Re:It's sad

I'm afraid I'm 'immature' then.

If I was helping create a distro, and nobody was being paid... Then only a few people got money for doing exactly the same thing as before, exactly the same thing as I'm doing... I'd be upset, then disgusted, then I'd probably quit. (I wouldn't be so immature as to remain and hold back the project, though.) Then I'd either find something else to do with my life, find another distro to help, or make my own.

Yes, there's ego involved... Everyone on a 'team' wants to feel like their at least equal to everyone else. With some people being paid and others not, it draws a very clear 'you're not as valuable' line. This is exactly the reason that many businesses make it a fire-able offense to discuss wages with other employees. And I whole-heartedly agree with that policy.

Re:It's sad

While what you say is true, the problems stemming from paying some developers should have been anticipated.

I think it was anticipated. I don't think the full fury of their immature response was, however. I don't think they believed that people would put their effort into making a project succeed, then turn around and put it into making it fail by deliberately holding it back.

I think that while their contributions may be highly valuable and are in any case appreciated, the people who would do such things should be removed from the project. Their actions prove that they are interested more in their own reputation than in actually making the project succeed. If their efforts in some other place can assist the project, then so be it, but I think that keeping them around where they can enjoy further self-aggrandizement is only rewarding bad behavior, which encourages more bad behavior. When my little parrot squawks at me over and over again, fit to burst my eardrums (what bird experts typically refer to as "inappropriate vocalization") I don't yell at her; that just gives her attention. I cover her cage with a sheet, and wait for her to calm down. Perhaps the same response is appropriate in this situation.

I'd say that our continuous tendency to reward bad behavior is the biggest problem in the world today.

If you and I are both working on a project for free, and the organization running that project decide to pay you but not me, what they are essentially saying is "tyke, you're contribution is not as important as drinkypoo's". That is a slap in the face, especially if I think my contribution is as important as yours. True, I haven't lost anything, but you can't overlook the de-motivational impact of rewarding some people but not others.

It IS saying that. In so many words! And as a contributor to the Debian project, these people have to decide what is more important; their own ego, or the Debian project. If they feel the former is true, then rather than deliberately holding Debian back, they need to go somewhere where they will receive the appreciation they feel they so richly deserve. Because there should never be room for someone whose ego is larger than the project.

Because let's face it, I might HAVE a greater contribution to make than you do, and there is only so much money to be shared. So I might get that money, and you might not. My contribution might BE more valuable than yours is. Does that mean yours is not valuable? Of course not. Does it mean that YOU are not valuable? By the same logic, it cannot mean that. It can only mean that I am more critical to the project than you are, and thus it is worth it to pay me to be sure of retaining me. In this world we all have to accept that we are not at the pinnacle of every scale, not least because many are contradictory. Even if I were the most badass programmer to ever have lived (which I clearly am not, but bear with me) I would probably not be the best person on the planet. There's only so much of each of us to go around, if you are spread thin then you never reach much of a height in any category.

It sounds to me like some of these people are good programmers, but not very good people.

Re:It's sad

Debian should never hold face-to-face meetings, something bad always happens

The lack of social skills is a really sad aspect in our community. I suspect it is at the root of your comment. Some geeks take a long time to mature (and some never do ;-). You had your shot as DPL, and the recent voting for SPI director suggests there is still room for improvement. Claiming that we should never meet seems defeatist to me. Meeting together and working on our social skills looks like a better choice to me. If we start out by recognizing we need to practice our social skills, we can improve.

Times have changed. The old joke about no one on the internet knowing you are a dog still applies. But our respect is still based upon skills and knowledge. We just need to augment our view of what a person accomplishes to contain a social aspect too. The process may not be pain free. If we must deal with expulsion requests [debian.org] or a myriad of flamefests, then so be it. The Debian core values remain intact. We need to learn how to scale to larger numbers without diluting them.

Ian Murdock...

Ian Murdock was the reason I first tried Debian, after disastrous experience with early RedHat builds. Read an interview, he seemed like a good guy and knew how to run a project.

Debian's meteoric rise in suckitude correlates very well with Murdock's departure and the further stepping away from the way he ran things.

Ubuntu is the new Debian--even despite its often-busted packages and all.

The problem with Debian

is Deb and Ian. That's what an IBM guy told me at FOSE a few years back.

IDNRTA

I did not read the article, but here's my two cents:

Ubuntu is trying to be a Windows killer. And it could be. Wine is "good enough" with the right settings for 90% of what most people want to do coming from a Windows world. Drivers exist. No, they're not FOSS, and I understand why people want FOSS ones, but....

Why doesn't Ubuntu seal the deal?

With beryl, good drivers, and built in FOSS apps that beat MS at every turn (Firefox > IE, Beryl > Aero, Thunderbird > Outlook, and VLC > WMP), it seems like the win would be fast and clear. Nobody wants Vista, especially when you have to pay. Ubuntu comes preconfigured in a way that is over all superior to every Windows that has ever existed. It's more solid and reliable, it has four desktops (though they moronically all have the same wallpaper by default, and it happens to be shit brown), it has a very nice user interface (though *i* and many others feel it could take some design cues from Windows 98 with regards to menu structure and some other minor details), and it's free. Oh yeah, and it's open source, so anybody who doesn't like part of it can fix it themselves.

But nobody has. It's like people take pride in allowing the world of uneducated masses sucking on the corporate tit of MS. I just don't understand it.

Feisty could win the OS wars decisively, but given the over all FOSS community attitude towards ordinary people....

Damn, gotta catch my plane.....

Sad?

rhY

Re:IDNRTA

But nobody has. It's like people take pride in allowing the world of uneducated masses sucking on the corporate tit of MS. I just don't understand it.

Feisty could win the OS wars decisively, but given the over all FOSS community attitude towards ordinary people....

Um, if this attitude was such an obstacle, then Ubuntu wouldn't exist in the first place. If anything, Ubuntu is proof that there is a significant portion of the FOSS community that wants to bring FOSS to "ordinary people". Sure there are people who don't, and they're running Slackware.

So given that, I must have completely missed the part where you specified what it is that is preventing Ubuntu from winning the OS wars decisively. You say it's comes preconfigured in a way superior to Windows. Personally I think Ubuntu, and Linux in general, has a ways to go before it's really an "ordinary people" as in "Windows replacement for everyone" kind of OS. I think they're a long way from winning the OS wars decisively or otherwise. But it is getting there, by leaps and bounds. You seem to think it's even farther along this path than I do, poised and ready to claim victory, so I'm again left wondering what it is you think is holding Ubuntu back.

Re:IDNRTA

I'll explain it to you.

Thunderbird Outlook and in some cases, *nothing* = Outlook for calendaring, contact management, etc. When Linux has a drop-in replacement for Outlook that connects to Exchange Servers and can handle PSTs, they'll have the killer app needed to crush Office. Until then, it'll be no sale. Believe me, programmers would probably love to switch but they still need to get email at work from the Exchange Server.

And no, solutions that require interdiction with Exchange administration do not count. Drop-in replacement is exactly that, just your Windows domain username and password.

-BA

Re:IDNRTA

Why doesn't Ubuntu seal the deal?

With beryl, good drivers, and built in FOSS apps that beat MS at every turn (Firefox > IE, Beryl > Aero, Thunderbird > Outlook, and VLC > WMP), it seems like the win would be fast and clear. Nobody wants Vista, especially when you have to pay. Ubuntu comes preconfigured in a way that is over all superior to every Windows that has ever existed. It's more solid and reliable, it has four desktops (though they moronically all have the same wallpaper by default, and it happens to be shit brown), it has a very nice user interface (though *i* and many others feel it could take some design cues from Windows 98 with regards to menu structure and some other minor details), and it's free. Oh yeah, and it's open source, so anybody who doesn't like part of it can fix it themselves.

rant:
Thunderbird > Outlook? Seriously? Outlook is one of the very, very few apps that Microsoft got somewhat right. As opposed to Thunderbird, it can be used to share calendards, contacts and stuff easily. Thunderbird is just an E-Mail app. Outlook isn't.
VLC > WMP? For some values of VLC, that is correct. But the userinterface is better on WMP. How on earth do you get a slider in fullscreen mode on VLC?

And your statemend about open source is just plain wrong. I can't see my mom "fixing" the freaking lameness that is "cut and paste" in gnome. It's simply broken, it doesn't work. When it does work, you have to try pasting in three different ways! Open Source doesn't mean anybody can fix. It means that the knowledgeable *may* fix stuff that they find annoying. Even then, it might not go upstream so other users can benefit from it.

I'm an ubuntu-only user, so I think I am semi-qualified to know what I'm talking about. I dig linux. I've been digging linux since '93. I've had windows too periodically, but linux usage far outweighs windows usage.
Linux sucks, unless you're somewhat skilled. Take the Gnome copy-paste dysfunction for example. When copying in the terminal, *sometimes* it picks up what I've marked with my cursor, so that I can just press shift-insert. Sometimes it doesn't. WTF? WHY NOT?. Oh well, then I have to right-click to make it put the text on the clipboard. So... now I've got the text on the clipboard, everything should be fine and dandy, right? NO! I still can't use shift-insert in a sane way. Sometimes it works, sometimes it doesn't and I have to rightclick *again* to paste! WTFISTHAT? I've switched to Kubuntu not long ago, and thank god... The copy and paste functionality appears to actually WORK AT ALL. It works pretty good. The even better part is that you can predict if it works or not. With gnome you just can't.

How do you suppose I fix that? It's open source isn't it? Then I should be able to fix that easily!

To all the proponents of Linux On The Desktop:

1. Please stop flounting linux as totally superior. Be realistic. It sucks in many ways, but it sucks in other ways than Windows
2. Make sure that you point out that learning linux isn't as easy as windows. Really. Do it. Please.
3. Make sure you've pointed out 2.
4. Accept that Linux is a Tool, just like Windows. Every tool has its good and bad sides. Windows has a (mostly) coherent user experience, linux has not. Windows has (inflexible) wizards, Linux has extreme flexibility (at the cost of complexity). You can't have it all. EVER. /rant

You can mod me down now. Just had to get that out. Should be incoherent enough to make it hard to read :-P

The no RC bug ideal...

...is sorta like the "no deaths in traffic" ideal, nice ideal but if you live it to the letter everything wlll stop. What gets Debian every time is the long tail of RC bugs, some long-lived bugs in e.g. the kernel linger on while less critical software go through many cycles. They go into a sort of meta-support stage where they're busy backporting fixes to etch, before it's even released. Sure every distro has those but for Debian it seems to go on for months and months.

Yeah, 'When It's Ready'

If you're looking for the latest drivers/kernel tweaks, it seems like Debian is perpetually behind. Every so often I try installing it (and Ubuntu/Kubuntu also), but with any new hardware it breaks and I end up re-installing SuSE again. Not that SuSE is perfect but at least it works with my hardware better than Debian/Ubuntu/Kubuntu.

Slowness

The slowness of Debian updates is a feature, not a bug. When you have a server 4,000 miles away from home (where a major OS upgrade can quite easily leave the machine an unbootable lump of metal), having a long time between major releases, and the updates to the current release being rock solid - it's a BIG feature. It's why I run Debian on those servers - because it's a lot less stressful than running a faster moving distribution.

On a point of pedantry, also you cannot have a meteoric rise. Meteors fall!

Different markets

The slowness of Debian updates is a feature, not a bug. ... It's why I run Debian on those servers - because it's a lot less stressful than running a faster moving distribution.

Definitely. I've been using Debian for over a decade, but what I'm seeing now is that Debian and Ubuntu are cooperatively focusing on two different markets. They aren't really duplicating effort, because they seem to be sharing packages and patches back and forth, and even users can setup hybrid systems if desired. But what they are doing is aiming for two different things.

For the moment, Debian seems to be producing a more stable distribution with server packages kept up-to-date and good attention to security fixes. Ubuntu seems to be producing a more user friendly distribution with simpler installation, ease of use, and more up-to-date desktop packages.

I see this as being beneficial so far. Any software developed for one of them can be ported to the other, and so having two separate organizations developing two different lines for two different purposes can make progress and quality better on the whole.

Debian has it's place

I used debian for years on my servers and desktop and really enjoyed it. Then one day I went to install a hauppauge video capture card and a couple other devices that aren't very standard. After weeks of recompiling the kernel, out-of-branch kernel sources, and various other things it became very tedious. A friend gave me an Ubuntu CD to try it out and everything just worked out of the box. Every piece of hardware was configured and working nicely out of the install, and the universe/multiverse feature was nice for getting things Debian normally doesn't carry. So for now I prefer Ubuntu for the desktop, and Debian or Ubuntu for servers. Just my oppinion, but I've had a couple friends switch over too because they wanted more bleeding edge software or wanting things to just work.

When Did NetCraft Confirm This?

Let's get a few things straight.

1. Another post mentions a concatenation of problems. I agree with this post.

2. Ubuntu is not a good server distro!
Stable and well-tested older packages are a strength of Debian. Yes there is a large class of sysadmins that like keeping odd hours running buggier systems. They generally burnout or learn how valuable stable is. To address the rather immature "needs newer packages" complaints, may I refer you to http://www.backports.org/dokuwiki/doku.php [backports.org]

3. Depth of Knowledge
There are still, many excellent Debian sysadmins out there that share and certainly have brought my skills up to a higher level. I don't see the same depth in Ubuntu forums.

4. Ubuntu Money
Mark's bringing money to the table, he gets to call the shots. That's well and good because the honeymoon is on right now. What happens when the honeymoon is over? Debian doesn't look organized compared to a guy calling the shots with his bankroll. It's an apples-and-oranges comparison.

5. Etch
I'm running etch right now on my desktop and in testing. It was ubuntu-release quality months ago.

Am I the only one...

..who looked at the title and for just half a second though, "Who gives a rat's ass what the owner of NewsCorp(TM) thinks about Debian??"

Debian will never die.

Just take a brief look at this list.

http://www.debian.org/users/ [debian.org]

Ditch Stable

They should ditch stable, testing should be given more direct-oversight. Stable is always released way out of date, and all the news is just how out-of-date Debian is. Let organisations that can make decisions take testing and "stabalize" it.

They have the most ears

The Debian society has my best wishes. I just wish they learned to listen to non-developers better.

Stable is the problem

In the last years Debian has become more and more the base for other distros, while Debians very own distro, namely the "Debian stable" branch has faded away into total obscurity. Sure it might still run on a few server here and there and I have it running on my router myself. But its basically a total failure. Why? Simple, sure you might want a distro that doesn't ship the newest stuff of everything and instead focuses on stable software, but that is *not* what Debian stable is doing. Debian stable is simply old software without RC bugs assigned that doesn't get updated any longer. That often means that you won't see the bug fixes that happen in upstream. More then once it has happened to me that upstream already long fixed critical bugs that never made it into stable, would I have run 'unstable' I would have never run into the bugs in the first place. And of course it also happens a lot that I run into software that either isn't available in stable or so outdated that its just of no use any more, so I have to compile it myself, use untrustworthy third party repositories and do all that security update thing completly manually, since Debian stable of course doesn't provide any updates for unofficial packages.

Debian should just give up on the whole stable thing is it is now and instead turn it into a branch that is really worth its name, i.e. only package software there that is really stable and proven to work and not some FooBar 0.0.6 stuff that just happens to be floating around in unstable without RC bugs assigned, since that will be obsolete in a matter of weeks anyway and provides basically zero value at a point where stable is released, let alone years after that.

Debian testing/unstable as become an important core of other distros and they should really focus much more on that instead of trying to pretend that unstable one day will become stable, since that simply isn't what is happening.

Debian Weekly News says it all

http://www.debian.org/News/weekly/ [debian.org] says it all really...

It is *called* 'weekly' news yet, in most cases, it comes out monthly.

Seems pretty inevitable...

Hardly surprises me that every few years, a group so rigorously dedicated to a set of strongly defined principles would suffer a period of amok time. It's simply logical.

Old news

How is this news? I knew Debian was dead the first time I saw Ubuntu Warty Warthog. Feisty is due in less than a month. The only reason Debian still exists is for those crazy free-only people who refuse to install a proprietary driver that's free as in beer, increases the functionality if their computer, and they would never look at the code even if it were open source. I say let it die, who cares?

On the Desktop, who cares? But on the Server...

I don't understand why so many desktop users are flocking to Ubuntu, which seems slow to me and has had quite a few show-stoppers, especially since they are essentially repackaging Deb at this point. On a desktop, Debian "Sid" (the unstable development branch) is quite bleeding edge, yet with less breakage than a "stable" build of Gentoo after a few make-worlds, or even Ubuntu right out of the box. This makes it very attractive to me as a powerful and current desktop for general use and development. Plus, the community itself is very knowledgeable, if not a bit grizzled. What's wrong with Sid on the desktop? For most people, absolutely nothing. For others, the biggest hangup is that there's just not a lot of exposure for it, and a net-inst can be tricky for n00bs. So while that can be improved, it still seems like a walk in the park compared to installing Win95 from floppy.

On the server however, it's a totally different beast. Be it webhost or data center, most admins are understandably wary of "unstable" distros -- as they should be. Now don't misunderstand me, Debian "unstable" is a damn site more stable -- in a reliability sense -- than the so-called "enterprise-grade" product which currently holds monopoly status in the industry. But as an admin, I don't want a system who's core packages change often, even if they are changes for the best. So the release candidate, "Etch", may be reliable and basically finished, but is effectively excluded for server usage en masse because it's not officially finalized. Again, this is because if the server changes and breaks part of the app stack, there's nothing but blood and tears for you as an admin. With that in mind, I hope to see Etch released and accepted abroad, so that Debian can get out of this rut of negative public perception, and consider being one of the best Linux distros anywhere.

I know how you feel, Murdock

Murdock believes that Debian is "process run amok" -- nobody feels empowered to make decisions, leading to the sluggish rate of progress."

That's how apt-get makes me feel, like I am not empowered to make any decisions. So even if I know that a library dependency is met, I am not empowered to override apt-get and tell it to install the package anyway. Apt-get is emblematic of the Debian project's biggest strength and, as Murdock has pointed out, its greatest weakness. "Process run amok" is a good way to describe it. People need to be able to make exceptions, because no one - not even the Debian maintainers - is a God.