Slashdot Mirror
Trouble on the Debian Front?
Linux.com is reporting that Matthew Garrett, one of the more active Debian developers, has called some ongoing problems with the Debian project into focus with his resignation. While he didn't hold any actual office, many prominent Debian developers described Garrett as "high profile". From the article: "In his own blog, Garrett relates his gradual discovery that Debian's free-for-all discussions were making him intensely irritable and unhappy with other members of the community. He contrasts Debian's organization with Ubuntu's more formal structure. In particular, he mentions Ubuntu's code of conduct, which is enforced on the distribution's mailing lists, suggesting that it 'helps a great deal in ensuring that discussions mostly remain technical.' He also approves of Ubuntu's more formal structure as 'a pretty explicit acknowledgment that not all developers are equal and some are possibly more worth listening to than others.' Then, in reference to Mark Shuttleworth, the founder and funder of Ubuntu, Garrett says, 'At the end of the day, having one person who can make arbitrary decisions and whose word is effectively law probably helps in many cases.'"
Especially when Ubuntu was released, everyone thought Debian was "dead" or "irrelevant", despite Ubuntu being based off Debian.
However, Debian's release cycle is picking up the pace, as Etch is set to be released soon (a quicker release cycle than Sarge's). Things are looking good as far as a mere user like me is concerned. There are a lot of hardworking people working on Debian, and the politicking is nothing new.
Lack of leaders is not the same thing as a lack of rules, and I expect that the real problem with the Debian project is that they haven't yet gotten to the point of fully defining rules that enable decent and useful conversations while discouraging the less productive kinds of conversations.
Sometimes boldness is in fashion. Sometimes only the brave will be bold.
"From the article: "In his own blog, Garrett relates his gradual discovery that Debian's free-for-all discussions were making him intensely irritable and unhappy with other members of the community. He contrasts Debian's organization with Ubuntu's more formal structure. In particular, he mentions Ubuntu's code of conduct, which is enforced on the distribution's mailing lists, suggesting that it 'helps a great deal in ensuring that discussions mostly remain technical.' He also approves of Ubuntu's more formal structure as 'a pretty explicit acknowledgment that not all developers are equal and some are possibly more worth listening to than others.' Then, in reference to Mark Shuttleworth, the founder and funder of Ubuntu, Garrett says, 'At the end of the day, having one person who can make arbitrary decisions and whose word is effectively law probably helps in many cases.'""
Wow! Sounds a lot like the cathedral model. Doesn't it?
It's kinda interesting, the last comment regarding having a single individual who's word is basically law in a project. It's worked for the Linux kernel, and the longest surviving Linux distribution (Slackware).
I was never a fan of the political backend of Debian, but I recognize the developers' contributions to the distribution. Maybe now that Ubuntu is popular and succeeding, a change in the way politics are done at Debian is on the horizon?
NetBSD is dying
What you have here is someone who has taken an either/or position on formal structure. This is a fallacy that is refuted every single time it is used. 'You're either with us or against us', 'Emacs, not vi', and 'cathedral, not bazaar'.
What is necessary is not a central bureaucracy that keeps people in line. Nor is it absolute freedom that allows any idiot to speak with equal stature of someone with multiple credentials. There are no hard and set rules that will make one project more successful or attractive than another. The best you can do is to take care of the community members that are productive and useful and try to avoid those members who are more prone to religious wars than code reviews.
I doubt many of Debian's greatest contributors would have been there building stuff for these years had the organization been different. Much of Debian's beauty and attractive for many is based precisely on its 'loose' or rather free structure. If it survives, or if it will disappear we do not know yet, but it right now its offspring have shown they are really strong and effective, and I guess thats one of the main reasons-to-be for almost any entity, be it living or algorithmic.
... from the forgotten corner in europe
...with great power comes great responsibility.
I believe that Ubuntu is on the right track because of the rules they have in place. Some open-source advocates confuse structure with lack of innovation, or with coerciveness, and thus eschew these rules which, in the long run, will hurt their cause. Anarchist behavior appears to be a good thing only in fiction. In real life it leads to erosion of the institutions that harbor it.
The open-source community wields great power now that our software is being adopted for solving a wider range of problems. Our responsibility is to create an environment that will promote cooperation and the continuous evolution of our products and services. An environment where flamewars and egos are flaring all the time will always end up hurting the projects until they wilt and die. This hurts our collective credibility and hinders our ability to bring more open-source projects in-house.
Cheers,
E
http://eugeneciurana.com | http://ciurana.eu
Welcome to the light.
i feel like a traitor, but should i at least look at ubuntu?
I love the wide support of Debian. That's why i've been using it for so long.
Have you read my journal today?
...kind of like the USA President .... oh.
Lack of leaders is not the same thing as a lack of rules, and I expect that the real problem with the Debian project is that they haven't yet gotten to the point of fully defining rules that enable decent and useful conversations while discouraging the less productive kinds of conversations.
The sad bit is that you usually need a leader to help make rules; when it comes down to it, the top couple of people most interested/involved/popular/whatever set some basic rules. Too many cooks etc. Add in egotistical or socially clueless people...and the number of practical cooks drops. Radically.
The really sad bit is that "just enough" of the people left out will devote endless amounts of time to arguing about said rules. BTDT in many clubs, for example. The best approach is to write the first draft of rules to be simple, un-evil, and able to be modified in the future, but not too easily.
Please help metamoderate.
Why doesn't he simply drop it and join Microsoft. Microsoft has all he needs!! Microsoft Windows has even more like WGA, Microsoft Office and Microsoft Visual Studio, which comes with my favorite reading, MSDN, which will help you out with your way into .NET 2.0 Final RC Beta 32 bit AND x64 versions. And you'll always have Microsoft DirectX for all your 3D needs! Even hits 4th on Google if you query "best operating system ever"!
I'm kind of getting sick of this. Every once in a while a story gets approved that is little more than some developer angsting away on his blog about some scene drama. It's not really "news", it's not stuff that matters much.
The irony here is that the blog post in question here is about the virtues of having leaders who can make arbitrary decisions. Apparently this doesn't apply to Slashdot.
Put a group of alpha geeks in a room and start a discussion. Inevitably, they spend more time trying to prove to each other who is the smartest than they do actually pushing forward the discussion. Why is that?
"I'd rather be a lightning rod than a seismometer." -Ken Kesey
Netcraft confirms it, Garrett is whining.
random underscore blankspace at ya know hoo dot comedy.
Without Debian, where would we have the amazing, huge codebase for every Ubuntu, Jibbajabba, or Lilixinidros distribution out there? Debian is the closest thing to a "standard Linux" if there ever was one. Slackware is a good candidate, too.
But NOT Fedora. That's a commie plot, I tells ye.
random underscore blankspace at ya know hoo dot comedy.
It's also worked for Apple's OS X, which claims to be the most widely distributed desktop version of *nix ever.
I tend to agree that there needs to be somebody to make final decisions on matters of wide questionability. Just the other day I compiled an app on Ubuntu and moved it to RHEL3 only to find that the static libraries were in a different location. I praised Apple's build system as well as the efforts of LSB and gave up on my quest to run hacked code on RHEL3 since I'm nowhere near a guru developer. (The app compiled and ran flawlesly on OS X and Ubuntu using debian packages.)
I disagree with this developer's comments, and suggest that perhaps his way of thinking is perhaps not suited to a meritocracy. Perhaps he needs an authority to appeal to in situations of disagreement.
While having one point of authority is good if you are looking to conduct a project under corporate type structures, it is undesirable if you are looking to adhere to principles of community involvement and community focused agendas.
I agree that it must be acknowledged that not all developers are equal, but disagree that this must be explicitly stated somewhere. In an open, meritocratic forum, relative skill levels become apparent fairly quickly, and if you need full and formal recognition of your work, then you are out of place in the open source community.
I have found the Debian mailing lists to be quite helpful, and if there genuinely is a lack of an appropriate forum for technical discussions, then this is a minor administrative problem (i.e., get a moderator to keep discussions on topic in the developer lists), not an intractable structural problem.
In any case, the proof of the pudding is in the eating, and I find it difficult to accept that the "Debian Way" is broken when the project is so old, so well regarded, and so successful.
Garrett: If you are unable to work in the Debian project becuase your ideas conflict with it, then don't be blaming the Debian project. It may simply be the case, as with many relationship breakdowns, that your ideals and theirs are simply incompatible.
I hate printers.
First of all whats with the hostility with Ubuntu being a Windows replacement? You got something against technology being accessible to regular people? Does it diminish your value as a geek when people no longer have to defer to your black magic ways?
Anyway yes Debian is stable because it doesn't integrate new stuff too often. Its also antiquated. If you want really stable you could always use Windows 3.11 for Workgroups but that would suck wouldn't it? Do you really have to wait 2 years before updating a package before you will trust it?
Your attitude concerning Debian's rate of development is one of the _core_ issues that caused Matt (and many others) to leave. It used to be funny how old th estuff in Debian stable was. Why is OpenBSD able to use modern packages yet remain stable and secure while Debian is not?
Mac OS X and Windows XP working side by side to fight back the night.
That is nothing new. Linus fills that position for the kernel. Mark has the potential to be that for the whole Linux OS. Lets hope he is into Ubuntu for the long haul.
This is one of several reasons I switched to OpenBSD after nearly a decade of Linux (first Slackware, then Debian). Nowadays I won't touch Linux at all unless it's mandated by the project leader or customer.
Wow, this wasn't even off topic. Try -1 nonsensical
"THERE IS NO JUSTICE, THERE IS ONLY ME." -Death
Besides, Ubuntu would not exist and remain in existence without Debian.
Errr, what would keep Ubuntu from continuing if Debian simply and abruptly came to an end? Perhaps it would somewhat affect the course of Ubuntu's development but it wouldn't spell the end of Ubuntu or any other successful Debian-based distribution should Debian itself become defunct.
H. Sapiens remain in existence today despite the fact that H. Erectus ceased to exist long ago. Perhaps Debian is reaching the end of its predominance and the frontrunning Debian-based offshoot, Ubuntu, is finding its place as a replacement. It really looks to me like evolutionary development occurring within the Free Software ecosystem--Linux went from being a student hacker's experiment, to a hobbyist/enthusiasts toy, to a few rough-around-the-edges distributions managed usually by individuals (eg. Slakware), to full-fledged community-driven collaberative efforts (Debian) and commercially-driven products (Red Hat, SuSE).
Since the commercially-driven efforts continually evolve (Red Hat dropping consumer-level products and establishing Fedora, Mandrake and Connectiva merging and re-inventing their businesses, SuSE being bought by Novell and releasing a community edition of its own) what should keep purely community-driven efforts from evolving as well? Ubuntu is a reponse to influences and pressures of the Free Software community--it shares the same technology, much of the same content and has some common roots in its founders and contributers. It keeps Debian's strengths (package management system loved by many, lack of direct corporate influence and commitment to the concept of Free Software, relatively high commitment to stability etc.) and abandons other characteristics that are weaknesses (lack of organisational structure, political disputes impeding on technical progress, slow pace of development at times, unpredictable release cycle).
This is exactly what makes Free Software so valuable--even if Debian were to disintegrate as a project there will be nothing to keep Debian's code and heritage from living on in new projects that pick up the pieces and move forward in great and exciting new directions. I have personally seen a couple of closed software applications of great value pretty much die because the companies responsible for development went insolvent, and for what I can only think are financial reasons nobody ever let the code go Free (perhaps doing so would make the intellectual property asset worthless from a balance-sheet perspective--in one case the receiver sold all IP to a competitor and all that remained of its applications were what was incorporated in the competing product. In the other case much of the software became abandonware).
So while this news may be cause for sadness towards a legendary Free Software project, it is far from cause for alarm. Debian itself will evolve into something better, or perhaps go extinct while its resources fully migrate over to a new project, likely Ubuntu. In the end we'll all get better software as a result.
Things are looking good as far as a mere user like me is concerned.
Exactly. What problems are actually showing up in software?
A developer is leaving, that's a problem. It's sad to see a talented developer go, but someone else will step up the the plate and prove that every developer indeed deserves a voice.
A developer claims that mailing lists made him irritable. That's a problem that has one of two causes, the lists have been infiltrated by trolls or he needs to more tolerant and less easily bothered. The solution treats both causes. Realize that some people on your list are intentionally provoking you and ignore them. Realize also that differences can always be worked out and that not everything has to go exactly your way. If you are right, the project will get back to your way even when it makes mistakes.
Free software has enemies, that's a problem. Back in 1998, Microsoft declared war on free software with their Halloween Document and targeted the user community. Trolling lists is something they have been doing all the way back to Steven Barkto. It disrupts useful activity, promotes ill will and distrust of your neighbor and can even move organizations to the wrong conclusions and in the wrong directions. Eventually, the truth comes out so the strategy is ultimately wasteful. There is nothing M$ can do to make non free software competitive and they can't really shut down free software. There are far too many projects and damaged communication channels are routed around. The co operative spirit of free software depends on good will, but free software creates that good will in abundance.
The answer is not to make a king. If you think your peer is annoying now, imagine them with the king like power to make decisions you want for yourself.
None of these problems is an actual software problem. The kind of people who pretend such things are a big deal are the kinds of people that said free software could not make a friendly user interface, usable documentation, a coherent distribution, a kernel, a compiler, a text editor, etc. Etch is a fantastic distribution that shows that things are working very well.
Friends don't help friends install M$ junk.
... only problem with Debian is the amount of politics involved. X is severely broken for weeks, but let's discuss politics and ethics instead. Nice goin'.
-- The Online Photo Editor - http://www.phixr.com
i feel like a traitor, but should i at least look at ubuntu?
That depends on what you want to do. If you want to play games with accelerated graphics or watch YouTube or other flash stuff, you need Ubuntu's non free goodies. If you want a sane place to put your email, web research and 95% of what computers do for people, you want Debian's free goodness. Debian runs well and upgrades gracefully. A simple rule might be: Stable on the server, Testing on your desktop, Ubuntu, Mepis, Xandros, Linspire, etc on your toybox. If playing with the software itself is your thing, go for Sid. Give the people what they want. That includes yourself.
Friends don't help friends install M$ junk.
posted by Charles... Maybe this is OSS evolution in action.
Who would've thought that civilized, focused discussion would be more productive than a free-for-all...
I love Debian, but I've long had the suspicion that part of the reason Debian has such a long time between releases (which I view as a mostly good thing) is because they've got too much of a "free form" development process. That's good for small projects, and it served Debian well in the past, but Debian's scope has broaded so much in the last 5+ years that new considerations should be made...
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
So, if someone leaves Canonical/Ubuntu, lets say jdub (Jeff Waugh), then they are in troubles too?
P.S. i give you that Garret and Waugh resignation motives are different...
Linux is dying...
I disagree with you. Even a meritocracy needs some method of breaking deadlock. The challenge is to find what I'd term an 'enlightened' leader. To define the term, it's a leader who him/herself is as ego-free as you can get it, sets achievable aims and who can balance out discussions in a fair but focused and constructive way. I certainly class Linus Torvalds and Mark Shuttleworth in that category, Linus by reputation and what he does, and Mark because I know him (and again, by what he has done and is doing). BTW, 'enlightened' does not mean 'perfect' - we're all human :-).
:-) - you don't abuse the priviledge of people wanting to work for you and it's a two-way street.
I have led tech teams myself, and I can still call any of the people I've worked with and ask them to come and work for me - out of a 100 people I would be disappointed if not at least 90 would want to (barring personal circumstances). THAT is a matter of personal pride to me. Not that I screwed some more hours out of a poor slob, and I've had to re-educate quite a few managers on that topic. I've had often enough that I had to instruct team leaders to drag their team off their chairs into a restaurant or cafe because they were working too hard (yes, company paid
Leadership is making a team of 10 think as if they were 20 strong - it sort of 'sings' and is one of the best working experiences you can have. Few teams do that by themselves because it's a different skill..
Insert
I believe to correct quote was "With great power comes great Ownage.".
As I posted earlier, even in a meritocracy you need leadership. As long as participants (a) recognise that leadership is a skill in itself (b) respect the person herding the cats for doing that job and (c) the leader has the trust and support of the group (which is where the authority comes in, as well as strong personal ethics and honesty) it'll work.
/know/ they are good - the challenge is to make them see that they can even be better by allowing the team to work as a team. And I know from happy experience that it only takes one team session where everyone works as a team to convert those people forever to the idea. /THAT/ is leadership in my book.
The trust is the hard bit - it's a special skill to manage a number of often quite strong personalities. People that are good at what they do
Insert
Hello, Testing, just because it looks like I can never post. Merc.
Hello,
I can see two problems with the way people are interpreting what happened.
The first one is that a lot of people are implying "One developer has left. Big deal. Somebody will step in". FALSE. A single, skilled developer can make the difference between a successful project and an unsuccessful one. As many good manages know, replacing a good worker is _very_ hard - sometimes impossible.
The second problem, is that a lot of people here have written comments without reading the mailing lists. Somebody implied "oh, it's the developer's fault, he shouldn't have been bothered in the first place". FALSE. Garrett really cares about the debian project; I generally agreed with what he said; lately, I was thinking "Geee, if I were him, I would quit". He found some of the tones grating as you guys would have if you cared about the project - and, above anything else, if you had read some of the messages in the mailing list. Accusatory. Unnecessary. Excruciating. Always coming from the "usual suspects" - who nobody seems to be able to shut up.
More and more people will leave, unless things change - rapidly.
Merc.
Editor In Chief
http://www.freesoftwaremagazine.com/
depending what type of embedded systems you talk about. my experience in telecom/datacom is exactly opposite. cos move from vxWorks to Linux, not to WinCE. and may be eCOS, but again, not WinCE. i worked with WinCE development environment is bad. i mean truly BAD.
A lot of the comments here seem to be along the lines of "use Ubuntu for as a Windows replacement, use Debian for servers". Not necessarily. Ubuntu installs Gnome by default, which is not as much of a Windows replacement as KDE (as used in Kubuntu). For working servers, though, I've done server-only installs of Ubuntu, which leaves off X-Windows entirely, giving me a lightweight install I can customize from the command line. Then, as a test, I've added the Xubuntu components to one server, which adds Xfce4 and some X apps like the Synaptic package manager. I don't need Gnome or KDE on a server, even Xfce is a luxury I could easily do without, and Ubuntu lets me decide which way I want to go.
My point is that the monolithic Ubuntu Gnome install that gets the most publicity is not the be-all and end-all of the Ubuntu project. In particular, this "Windows replacement" talk is misplaced, if you stick with the Gnome defaults, instead of KDE.
(this is not a
"Indeed, it has been said that democracy is the worst form of government except all those other forms that have been tried from time to time."
That's what Churchill had to say about such matters. Indeed many people still don't know how to deal with trolls. Some people just like to get all up in arms from time to time I suppose. Other than that maybe he should have just announced that he was ignoring some people and that replies to those people should be marked somewhere so that he can sort them as well automatically. So that those people that like to respond to trolls can do so and don't confuse the ones that don't.
Although no one's come and out explicitly mentioned it yet, it strikes me that Mr Garrett is saying that traditional corporate structures work best when developing software.
Enforced rules of conduct, a formal structure, an acknowledgement that not everyone is equal is skill or knowledge and a single leader who has the power of final decision. Strip out the jargon and it sounds pretty much exactly like a traditional office environment.
Does this mean that while OSS has made many people rethink distribution and revenue models, open source development will mature into exactly what we have now?
For him, I would go -1, jackass
Think of any group activity in which you are involved. Being in a musical group is a great example and one which I I have quite a bit of experience. In the end, SOMEONE has to be in charge and perform the vital role of focused leadership. I have my problems with the peacock strutting attitude of Ubuntu, but I think they will succeed because of the organizational structure. Debian? I hope so too, but Ubuntu will get to the point where it doesn't need Debian. Though I suspect that will require Canonical hiring some key developers to really keep the process moving properly.
I am tired of these motherfucking Muslims on this motherfucking plane!
- overheard on flight 93
I have to agree with you - Debian is still live and kicking. Still I also think that Debian, like everything else, will have its day and then pass into history.
That's what the ensuing flameware will be about if you boil it down. How fast is Debian dying.
All this developer is saying is that he personally feels that the egalatarian/authoritarian balance is probably skewed in favor of the former in Debian.
And I have no opinion re Mark Shuttleworth, but ask all students of history: When does a benevolent authoritarian run a more efficient state than a republic/democracy? Every time. The trick is how to keep a succession of benevolent authoritarians...
Using plain ol' text since 1968
Overseeing a committee, is most likley the best way to run things ( even countries ). As long as the guy at the top isnt into a power trip.
---- Booth was a patriot ----
Just because having one person at the top (fascism) is simpler in some ways for some issues, doesn't mean it's the best way, the right way, or the even the easiest way. It's not like the debian project couldn't VOTE to administer a code of conduct. It's not like the chaos method they've been using the whole time hasn't resulted in one of THE BEST OSES IN HISTORY. Can we get more news, and less arbitrary opinions from nut jobs who admire low level thinking as it's own virtue? That goes for the government here, as well as for /..
I hold very few opinions. I hold information based on observation and fact. If you wish to disagree, please use facts.
Interesting, in light of some of the NetBSD issues posted by Charles... Maybe this is OSS evolution in action.
The real story behind Charles Hannums little rant has now been revealed - see the NetBSD-users mailing list. It turns out that Hannum had fucked up the day to day running of The NetBSD Foundation, to the point were it was not conforming to the regulations in Delaware where it was incorporated. Christos Zoulas and others sorted the mess out, but Hannum was totally uncooperative (the "fraudulent coup" crap that he was harping on about was him being ignored after being a jerk).
godwin's law, within the first 10 posts =/
Well, yes, leadership works and is needed for most projects. But what one must always have in mind is the difference between Dictatorship and Democracy.
In a large corporation, from the point of view of the employees, the manager is a dictator with almost unlimited powers. This is bad for corporations, because dictatorships are usually less efficient in the long run than democracies. It's good for the leader to be forced to listen to other people, absolute power normally makes a person so arrogant that he becomes unable to see his own mistakes.
The role played by Linus Torvalds, Mark Shuttleworth, or Patrick Volkerding in their respective organizations is "benevolent dictator", which means they have the power to make a decision when the discussions become stuck at a dead end, but they will make efforts to not abuse that power. In the end, the success of a benevolent dictator will depend on his practical management of day-to-day issues.
In this respect, I had misgivings about Debian from the start, because of that "GNU/" that they insist so much on prepending to the "Linux" name. Perhaps Linus Torvalds would have done the Debian team a favor if he had exerted his trademark ownership and prohibited them from using the name like that, but it's too late now. Having principles, specifying them in clear rules, and enforcing those rules, that's good and I admire that. But principles are, by definition, basic and fundamental. To start nitpicking about the name in the very first paragraph of their home page shows a lack of clear vision on the project's purpose and ends. In the same way that arrogance is the downfall of dictators, endless bickering on small issues is a poison for democracies.
This discussion sounds a lot like the divisiion between Marx's authoritarian communism and Bakunin's libertarian socialism.
Garrett's comments can be summed up as: "I don't develop for Debian because people don't treat me with the respect I think I deserve. Debian needs a dictator to make everyone be nice and make me feel happy."
So Garrett didn't fit in with the rest of the Debian community, despite his technical aptitude. There are plenty of social specs that take priority in communities, even mailing lists, which are often independent of technical qualities. Garrett apparently didn't like the Debian "anything goes" style in developer discussion, so he left.
No problem. He can switch to Ubuntu's team. Sounds like they'll be glad to have him. And interested people in the Debian community can still use Garrett's Ubuntu work to improve Debian, as it's all GPL. This is the strength of openness, both in the software and in the groups of people. When we can choose how and with whom (and with what) we work, we can work the way most productive for us. And thereby, for everyone else in the cycle.
--
make install -not war
It is now official. Netcraft confirms: Debian is dying.
One more crippling bombshell hit the already beleaguered Debian community when IDC confirmed that Debian market share has dropped yet again, now down to less than a fraction of 1 percent of all servers. Coming on the heels of a recent Netcraft survey which plainly states that Debian has lost more market share, this news serves to reinforce what we've known all along. Debian is collapsing in complete disarray, as fittingly exemplified by failing dead last in the recent Sys Admin comprehensive networking test.
You don't need to be the Amazing Bruce Perens to predict Debian's future. The hand writing is on the wall: Debian faces a bleak future. In fact there won't be any future at all for Ubuntu because Debian is dying. Things are looking very bad for Debian. As many of us are already aware, Debian continues to lose market share. Red ink flows like a river of blood.
Debian is the most endangered of them all, having lost 93% of its core developers. The sudden and unpleasant departures of long time Debian developer Matthew Garrett only serve to underscore the point more clearly. There can no longer be any doubt: Debian is dying.
Let's keep to the facts and look at the numbers.
Debian leaders state that there are 7000 users of Debian. How many users of Kubuntu are there? Let's see. The number of Debian versus Kubuntu posts on Usenet is roughly in ratio of 5 to 1. Therefore there are about 7000/5 = 1400 Kubuntu users. Xubuntu posts on Usenet are about half of the volume of Kubuntu posts. Therefore there are about 700 users of Xubuntu. A recent article put Ubuntu at about 80 percent of the Debian market. Therefore there are (7000+1400+700)*4 = 36400 Ubuntu users. This is consistent with the number of Usenet posts.
Due to the troubles of Walnut Creek, abysmal sales and so on, Knoppix went out of business and was taken over by SCO who sell another troubled OS. Now SCO is also dead, its corpse turned over to yet another charnel house.
All major surveys show that Debian has steadily declined in market share. Debian is very sick and its long term survival prospects are very dim. If Debian is to survive at all it will be among OS dilettante dabblers. Debian continues to decay. Nothing short of a miracle could save it at this point in time. For all practical purposes, Debian is dead.
Fact: Debian is dying
just get over it, as very few things are ever accomplished solely by individuals
imhe, most people are simply far too egotistical and selfish to work well with others
He was just trying to be a bit witty- no way that was a troll.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
Why is this typical copy of the "BSD is dying" meme, with a few Debian-specific terms substituted in, modded "+1 Interesting"?
- Kevin B. McCarty
I think that this whole issue with Debian not liking Ubuntu is misunderstanding. Sure, Ubuntu is using Debian's work, but Debian agreed to that in the first place, even before Ubuntu was concieved - that's the ultimate point of open source, after all.
Debian creates, Ubuntu releases. That's how I see it and it seems to work; I hope that some inflated e-peens won't break anything.
Debian's free-for-all discussions were making him intensely irritable
Seriously, if that's the only real problem in Debian-land these days then Debian is in really good shape. Two years ago they couldn't get a release out the door because of squabbles over how free was free enough.
Moderating "-1, Disagree" is simple censorship. Have the guts to post your opinion.
Yes, because you'd have to be INSANE to use Ubuntu for E-mail, Firefox and 95% of what computers do for people, such office work, solitaire and software development.
The opposite of the truth is often a lie. Debian is a sane place to put your email and other important stuff. It's stable and upgrades better than anything I've seen. Putting it elsewhere may not be "insane", so speak for yourself not me.
Placing that important work on a system that has non free software is not a good idea and might be worse than you think. Quality is usually less than free equivalents, so your performance will degrade in proportion to the amount of non free software you use. When such an option even exists, non free software is always more difficult to upgrade than free software. More important problems are in the EULA. Microsoft, to name an extreme example, gives itself the right to inspect your files for "copyright violation" and delete them at will with no further obligation to you. Clearly, it's insane to software from a company like that. Others may not be as bad, but you never can tell with non free software. Non free software, regardless of it's function, is dangerous. Non free software is written to do the bidding of it's owners.
You're full of it.
Yes thank you, there's more where that came from.
Friends don't help friends install M$ junk.
If this continues, it's only a matter of time before Ubuntu is thought of as "downstream" and Debian is the "offshoot". I mean, isn't that what would happen if Ubuntu starts getting more actual work done, and developers in Debian are porting the "upstream" Ubuntu stuff to Debian more and more? And look at what MEPIS has done!
... Ubuntu Embedded might be on the way ... maybe they'll call it Bubuntu or something hot like that.
Of course, this would only be from the perspective of a few architectures. Debian is still a big deal for embedded and niche archs. But who knows
random underscore blankspace at ya know hoo dot comedy.
for that post to make sense. I'm conceptually dislexic at times :)
random underscore blankspace at ya know hoo dot comedy.
Thank god the system is working, you have such "insigtful" things to say.
If you want to see what an OSS project where money buys voting rights, go look at Eclipse.
Its a company, not a community. They do some good stuff, with enough full time staff to realease on a regular basis. but it doesnt feel like a place that needs or wants your help, not unless you have $150K and two full time engineers to spare.
the worst part is that in an effort to appear smarter they try to come up with progressively more complex solutions.
Its almost like the purpose of government bureaucracy; take something simple and complicate it.
Seriously... the amount of times I've seen 'alpha geeks' introduce bleeding edge, over complicated solutions into a production environment... makes me shudder.
IMHO 'computer enthusiasts' should be doing R&D work, not operations and engineering. They are (typically) risk takers.
Operations and engineering are for people who don't take risks (with other peoples businesses or property).
In the free world the media isn't government run; the government is media run.
In FreeBSD-land, they knew this 10 years ago already.
We can be glad that at least the kernel is developed in a sensible way (with lots of room for improvement, nevertheless)
Windows 2000 - from the guys who brought us edlin
It's organized using the pyramid-power design. --That is, one all powerful individual at the top, and then cascading levels of management beneath. This is, of course, the standard model for most large organizations in the world, including business, military, government, religion, etc.
The problem is that such systems lend themselves to easy corruption by the forces of greed and self-service.
There is an alternative system for organizing, and it uses a cell-based power distribution system with no one individual at the top and no downward cascading power structure. Organic systems throughout the biosphere use the cell-based method of organization to great effect. --And many open source projects seem to work this way as well.
One of the noteworthy factors about Cell-Based systems is that they are far less easily corrupted by greed and self-serving individuals because everybody has the power to call attention to all manner of problems without the threat of recrimination or dismissal; without having their complaints arbitrarily over-turned by individuals who might be driven by ego and emotional concerns. Psychopaths are well suited to successfully infesting and rising through the ranks of pyramidally based power structures, because they are drawn to power. But when power is evenly distributed as it is in Cell-Based structures, where are they drawn to? --And how much more easily are problem individuals such as the psychpathic or sociopathic personality noticed and weeded out?
--It seems to me that the idea of being of non-self-service, but rather other-serving in orientation, (no multi-million dollar salaries for CEO's), is directly related to an entire pattern of thinking and awareness, part of which is intrinsically linked to the decisions for how the power and 'command' structure of the organization is laid out, either Cell-Based, or Pyramidal.
I think it serves well to be attentive of these two patterns and how they affect our world.
-FL
That Debian has serious problems is proven by the following article:
/boot/grub/menu.lst file now has entries for each kernel version:
/boot/vmlinuz-2.6.8-3-k7 root=/dev/hda6 ro /boot/initrd.img-2.6.8-3-k7
/boot/vmlinuz-2.6.8-2-686 root=/dev/hda6 ro /boot/initrd.img-2.6.8-2-686
/usr/share/doc/kernel*. Unfortunately, most of it was for the 2.4 kernel and I could not find where it explained how to reassemble the kernel sources.
/usr/src/.
/usr/src/
/usr/src/)
Extra Help for Debian Users.
Although, Debian has worked hard to make the addition of pre-compiled packages easy, they have gone out of their way to make compiling your own programs, a nightmare.
Unlike SuSE or Mandriva, the Debian packagers have not included the kernel sources for the default kernel images (those automatically chosen when you install Debian).
For example, the kernel image chosen for me, was:
kernel-image-2.6.8-2-686 (2.6.8-16sarge1) Kernel image for PPro/Celeron/PII/PIII/P4.
However, the only kernel source packages are:
kernel-source-2.2.25 (2.2.25-4) Kernel source for 2.2.25
kernel-source-2.4.27 (2.4.27-10sarge3) Kernel source for 2.4.27 with Debian patches
kernel-source-2.6.8 (2.6.8-16sarge4) Kernel source for 2.6.8 with Debian patches
none of which match the kernel image. Apparently, 2.6.8-16sarge4 means a 2.6.8 kernel, with general patch level 16 and security patch level 4 (20 levels of patching overall). The closest source package is kernel-source-2.6.8, however, it has a different security patch level (patch level four, instead of one). There are two ways around this, you can either roll back the patches to level one (hard) or choose a new kernel image which matches the available sources (easy). So, I chose the new kernel (which was a better fit than the automatically chosen kernel, anyway):
kernel-image-2.6.8-3-k7 (2.6.8-16sarge4) Kernel image for AMD K7.
After installing the new kernel image with synaptic, the
title Debian GNU/Linux, kernel 2.6.8-3-k7
root (hd0,5)
kernel
initrd
title Debian GNU/Linux, kernel 2.6.8-2-686
root (hd0,5)
kernel
initrd
Choosing Debian GNU/Linux, kernel 2.6.8-3-k7 from the grub menu, boots the 2.6.8-3-k7 kernel.
Choosing Debian GNU/Linux, kernel 2.6.8-2-686 from the grub menu, boots the 2.6.8-2-686 kernel.
Unlike SuSE or Mandriva, the Debian packagers have split the kernel source into a number of separate pieces.
For the 2.6.8-16sarge4 kernel, the sources come in the following three pieces:
kernel-source-2.6.8 (2.6.8-16sarge4),
kernel-headers-2.6.8-3 (2.6.8-16sarge4),
kernel-headers-2.6.8-3-k7 (2.6.8-16sarge4).
The installation does not reintegrate the sources, even though it easily could have done so, but leaves you with one unopened tar archive and two separate directories containing header files. Never before, have I seen a package install, that does not even bother to unpack the archive. This is the Debian packagers, deliberately making it difficult for you to compile your own kernel.
I looked around for a README, or some help, and was not surprised when I didn't find any. I eventually found lots of documentation in the directories
After installing the three, above mentioned, kernel packages with synaptic, you have the two directories of headers files:
kernel-headers-2.6.8-3
kernel-headers-2.6.8-3-k7
and the kernel-source-2.6.8.tar.bz2 archive in the directory
To sort out the install mess, execute the following commands:
cd
mv kernel-headers-2.6.8-3 linux-2.6.8-3-k7
tar cf kernel-headers-2.6.8-3.tar linux-2.6.8-3-k7
rm -fr linux-2.6.8-3-k7
(go to the directory
(the move ensures the tar archive will unpack into linux-2.6.8-3-k7)
(creates tar archive named kernel-headers-2.6.8-3.tar)
(removes
many people seem to think that software development should be some kind of democracy, and that whatever side has the most posts should win. This is clearly not the case as the majority of people in any community are not competent to make most kinds of technical decisions. In a smaller comunity, people who are experts in a field are pretty easy to identify, and decisions are generally made by having a short discussion with that person. In larger communities, such people aren't so easily identifiable, and their input will likely be ignored.
The solution generally employed is to identify certain people responsible for a project, and then to build a trust network rooted at those project leads. The leads are responsible for inviting knowledgable parties to any discussion, for moderating any discussion, and for making decisions in places where concensus can't be achieved. Practically speaking, it's usually more important that a decision is made and that people follow that decision, than that the right decision is made. In the worst case, the work can just be done again if a decision turns out to be so bad that it leads to unsurmountable technical obstacles.
As a simple example, if some piece of software must be written by a number of developers, a language must be chosen. One can debate endlessly the merits of various programming languages, and often people do. However, what one cannot do is choose more than one (in *most* cases). For most software projects, choosing multiple languages would make the code unreadable and create a number of technical hurdles (especially on linux where COM is non-existent and CORBA is... CORBA).
The best path to success is to reduce the conversation to just those who are competent to understand all the relevent factors,which is usually vanishingly few. This reduces time to make decisions, which increases time spent working, and it reduces the chances of crappy decisions being made.
The departure of this debian developer could be an occurance of organizational disruption, brought on by a guerella organizational disrupter.
Several recent marketing books offer "guerella marketing" as the way to promote products with less conventional advertising.
There is a corresponding discipline, "guerella organizational disruption" used in power politics and power business as a less than ethical set of methods to diminish the oponnent. Provocateurs are surmised to exist and not often acknowledged.
Debian is capable of being disrupted by deliberately fomented discord. Debian has a wide open organizational structure, a relatively small number of contributors and a determined style to accept a wide range of people in their community.
My brief contact with #debian on IRC was an unforgettable piece of brisk rudeness. (I went elsewhere and stay away from debian after that kind of humiliation).
Suppose what happened to me was deliberate guerella dismarketing. Suppose persons on the debian developer channels are actually practicing guerella organizational disruption?
How would you discover or map out the individual who consistently and plausibly practices organizational disruption?
I can see a way to start: Get an archive, rate every message, group rated messages by sender, sort to find the person who sends the consistently most disruptive messages.
I think it would be very interesting to interview one of these disruptive people and find out how they reached their state of mind. Is the person in some sense sponsored or involved with a competing operating system vendor? That would be interesting to find out.
(go to the directory /usr/src/)
/lib/modules/2.6.8-3-k7 /usr/src/linux-2.6.8-3-k7 boot /usr/src/linux-2.6.8-3-k7 source
(the move ensures the tar archive will unpack into linux-2.6.8-3-k7)
(creates tar archive named kernel-headers-2.6.8-3.tar)
(removes unneeded directory and contents)
mv kernel-headers-2.6.8-3-k7 linux-2.6.8-3-k7
tar cf kernel-headers-2.6.8-3-k7.tar linux-2.6.8-3-k7
rm -fr linux-2.6.8-3-k7
(the move ensures the tar archive will unpack into linux-2.6.8-3-k7)
(creates tar archive named kernel-headers-2.6.8-3-k7.tar)
(removes unneeded directory and contents)
tar xjf kernel-source-2.6.8.tar.bz2
mv kernel-source-2.6.8 linux-2.6.8-k7
(unpacks main kernel source)
(move main kernel source to linux-2.6.8-k7)
tar xf kernel-headers-2.6.8-3.tar
tar xf kernel-headers-2.6.8-3-k7.tar
(unpack kernel-headers-2.6.8-3.tar into linux-2.6.8-k7)
(unpack kernel-headers-2.6.8-3-k7.tar into linux-2.6.8-k7)
In summary, the directories of header files are repacked so that they will unpack into linux-2.6.8-3-k7. The kernel-source-2.6.8.tar.bz2 package is unpacked and moved to linux-2.6.8-3-k7. The tarred kernel-header directories, also unpack into linux-2.6.8-3-k7 and the sources are reintegrated.
By the way, the name linux-2.6.8-3-k7 was chosen simply because the command uname -r returned 2.6.8-3-k7.
Well, fine and good, so far. But the nightmare is not quite over.
The Debian packagers have "accidentally" forgotten a necessary link and have another pointing to some random place. To fix these "oversights" do the following:
cd
rm -fr boot source
ln -s
ln -s
Now you can continue with the ATI driver installation as above (tested on Debian Sarge).
Of course, you will have to adjust the kernel versions in the above to reflect your situation. You may also ask some Debian person for the name of the script (assuming it exists) that reassembles the unnecessarily split kernel sources. Of course, this script should be automatically run, when the kernel sources are installed.
Remember, that with SuSE or Mandriva, the kernel-sources can be installed with one or two clicks.
You may wish to speculate, as to the reasons the Debian packagers are deliberately making things difficult for you.
From http://m.domaindlx.com/LinuxHelp/ati/ati.htm
I sympathize, and agree that you must do what is necessary to protect your mental well-being.
.Net replacement unusable in business, through overly-strict licensing.
It's unfortunate, but probably true, that part of what has driven you to take this step has been activity by Microsoft astroturfers.
We all know that Microsoft is a company that will do anything to protect their monopoly position. That includes breaking the law by, among other things, committing fraud and sabotage. Consider, for example, Microsoft's past involvement in undermining WordPerfect, or the OpenGL standards development. Or, look at Microsoft's attempt to sabotage Java, in their own words, to "Kill cross-platform Java by grow[ing] the polluted Java market." Or, look at Microsoft's payments to SCO, and various "pretending to be independent" political action groups, such as the Alexis de Tocqueville Institute.
So, years ago (probably 1998), when Microsoft held a strategy meeting on how to stop Linux, someone would have proposed the following ideas:
> Infiltrate Open Source Development Teams and...
> - submit polluted (e.g. broken) code
> - push complexity and bad ideas
> - create dependencies on proprietary software
> - slow down progress in general
> - make life miserable for other developers
> - argue every minor detail, disrupt discussions
> - be pedantic about OSS concepts
> - polarize every discussion -> no common sense
> - hound and drive out the best developers
We know that Linux threatens to take a multi-billion dollar market away from Microsoft.
And we know that Microsoft has no morality, and little fear of the law.
And so, Microsoft would have chosen not to follow the above strategy... why? Because it's not nice?
We see Microsoft astroturfers at work all the time, writing FUD articles, or posting obvious lies in various forums.
Thus, I take it as a given that Microsoft has people involved with the most major Open Source projects, trying to do exactly as suggested above.
As some have pointed out, it can be difficult to interfere with a project such as Ubuntu, where a benevolent dictator can identify and kick out the troublemakers.
But it is relatively easier to infiltrate a highly democratic project, such as Debian.
Likewise, I expect that Microsoft has people in the FSF. Those people would not be arguing against Free Software, on the contrary, they would appear as the most extreme supporters of FSF principles. They would attempt, for example, to make GPL3 so extreme as to prevent its success in commercial venues, as they may have previously succeeded in making the FSF's
Unfortunately, sabotage has always been one of Microsoft's most effective weapons. While Microsoft's technical incompetence is such that they have never defeated a strong competitor through actual competition, various acts of sabotage have allowed Microsoft to come out ahead against DR_DOS, Geoworks, WordPerfect, OS/2, Go, AmiPro, Netscape, and Java, among others.
Also unfortunate is the fact that it is hard for a democratic group (political parties, unions, Debian, etc.) to prevent this sort of infiltration. After all, you can't go around accusing everyone of being an agent, or you create an atmosphere of paranoia, which just plays into the infiltrator's hands.
Thus, what we are usually left with is simply to continue to fight the good fight. Keep arguing with the jerks until it becomes obvious that they are jerks. Make them ineffective, and force them to start again with a new identity, and no credibility. And continue to push the project forward, even if at a slower pace.
Note that I am not trying to talk you into changing your mind. If you feel the way you do, then you should rest. Let the young idealistic hotshots carry the torch. As for you, if you still want to be involved with OSS, then choose your own project, where you can avoid the astroturfers and the jerks
Before I forget, I stupidly forgot the magic ingredient: having a VISION to follow, and I don't mean that in the 'I seeee a siiign' way although that too matches the requirements :-). Duh.
Look beyond the bluster and the, um, somewhat socially inept commentary (not unusual for tech people). He knows what he wants, says it the way he sees it and in general leaves no doubt what it takes to work in his sphere. Is he always right? No. Does everyone agree with him? No. Yet he continues to have a loyal following because he knows what he's doing, has a good track record and is consistent. This also gets me to another challenge for leaders: knowing how to be wrong.
If you have ran teams like I have with absolute stars in their field you may know as much as you want - an experienced hack is likely to have come across another way of doing things that may help move things along faster. That's why you employ them, so not letting them speak is stupid IMHO. I've had to knock sense into countless managers who somehow seemed to consider changing an approach because it's someone elses idea an attack on their authority. In my book the roles are pretty accurately defined: my team is supposed to give me the advice I need to decide what we're going to do, but after listening to everyone it's still *my* decision. That works both ways too: it's my neck, not theirs, and God help anyone trying to yell at my team without coming through me - it would break my 'praise in public', 'discuss in private' policy. I'm not a good guy to have a problem with, there's more to social engineering then just getting passwords .
I don't mean non-ego in the 'not having an ego' sense, I mean non-ego in that a decision process of a good leader should be divorced from whatever ego the person possesses. If you can manage that you're a leader, especially if you can add people management skills to that. And yes, I wouldn't rate Linus very high on the people skills, but for the people he deals with it's almost not required.
I must admit, though, that his habit of totally refusing the play the normal media games (i.e. thus annoying the hell out of people who are looking for some sort of sensational statement they can print) never stops amusing me..
Insert
"Equality is death!"
- Baron Rudolf Sebottendorff
"The only clear view is from atop the mountain of our dead selves." - Peter Carroll
Good riddance to M$, RIAA, MPAA, infiltrator Matthew Garrett:
Matthew Garrett is one of the M$, RIAA, MPAA, infiltrators described in the above post.
He has deleted the copy of the above article on M$ infiltrators from the main body of his blog, so that it will not attract the attention it deserves.
It is now a "(Screened Post)". Look for yourself: http://mjg59.livejournal.com/66647.html
When the article was first posted he collapsed the thread to hide it in a link that most would not click (previous to this the thread had been flat, ie, all posts were displayed).
Because he hid the article in this way, I posted it again to the main body of the thread, where it would attract attention, so he just deleted it.
Clearly, Matthew Garrett is one of the M$, RIAA, MPAA, infiltrators described in the post that he deleted.
I can't imagine running any other distro on a server. Why does a server needs X? Most distros install X by default. This sounds like Windows to me. You say you can't run anything new in Debian. You certainly don't know the bauty of the apt-get suite. Anyway, easier, just upgrade to testing. When I got mad at Woody for it's old packages I moved to Gentoo. Took a whole lot of time to have a system running. After a while I went back and moved to testing. It is working fine for me.
The metaphor may be applied to any level of the creation of software. Project management is a part of that and the bazaar-inhabitants have agreed to be at the bazaar to have their wares available.
The Third Way of collaborative software development is the Supermarket where an organised team facilitate a bazaar-like pattern within a cathedral-like administered setting: someone may develop a Debian package for Foo (which replicates work already from the bazaar in a different style) but makes sure that it works with the other Debian packages in the bazaar, while also aware of a the organisation calling for an off-the-shelf available stable edition of Foo for turn-key deployment.
The reason this is applicable is that very few people download their GNU/Linux packages from the original sources (although that could be an interesting USP for a distro), instead collecting a branded collection of sources (e.g. Gentoo) or binary packages (e.g. Fedora/Ubuntu/Debian/Suse etc.). The very notion of a GNU/Linx distribution is cathedral-like; the very best distributions allow their users to roam the bazaar, albeit their own-brand edition of the worldwide bazaar. This permits a user to administer the software they choose to use while supported by the standardised quality of one source of software packages.
How the system is governed is always going to be an interesting question. With a certain sort of person in a given supermarket software project, you will need strong leadership; with another sort of person, you will need collaborative decision-making powers. It's a battle fought throughout the world in a variety of different settings because people are just people, and I doubt that segregatng by people-type will be helpful in the long-run as tolerance and flexibility are needed.
Ubuntu patches are passed back up to Debian, so everyone benefits.
On the same blog, Alan Levin asks how such entrepreneurs can change the world. It is hard to give - sometimes harder than taking. In Africa, it is about transfer of ownership - can you persuade the recipient that you are not wasting their time ?
Andy Rabagliati
The biggest risk to Ubuntu is that the Debian developers start viewing it as a competitor and start screwing up.
What Ubuntu really is is an important sub-project of Debian, something that performs the final packaging and integration steps necessary for producing a high quality desktop distro, and something that Debian has never had the resources for. Nevertheless, it is important that Debian itself continue to produce their own distributions, however irregularly, because Debian as an organization also needs to go through that exercise.
But since a lot of people do open source also for the publicity, maybe Shuttleworth could help by featuring Debian more prominently as part of the Ubuntu project, perhaps going as far as calling the distribution "Debian Ubuntu" or "Ubuntu Debian", or at least using "...based on Debian" as a tagline.
That seems to be the problem with most OSS. Look at the descriptions (DESCR) for pkgsrc packages, or the main page for a project on SourceForge. You get very very little information on what the hell the software _does_. Sure, I can find out what lines changed in each source file, but how about a simple, maybe 3 line explanation for what the software does and why I'd use it?
</RANT>
Awesome furniture, accessories and cabinetry in Santa Rosa, CA: http://humanity-home.com/
But can we fairly characterize the early 'net as having a lack of fixed leaders? Usenet may have been chaotic, but to those who followed the RFCs, we had certain leaders. Jon Postel and Vint Cerf leap to mind as ones who led by example of patience, intelligence and reason. They may not have been "Self-Appointed Benevolent Dictators for Life," but they did provide leadership of a sort that effectively mitigated some of the anarchy we have seen with the Debian lot. (Disclaimer: I love Debian, have provided some modest feedback to the maintainers of their documentation over the years, and am currently an Ubuntu user and evangelist.)