Success of GNU/Linux on the Desktop

I was reading an interview with Jeff Zemlin of The Linux Foundation that ended with this:
“Every failure of Linux on the desktop each year has made Linux on the desktop better. If you use a modern linux desktop, it’s pretty good. and that’s the result of a lot of risk-taking over a lot of time.”

My response is “What Failure?” GNU/Linux has been a success on the desktop with every distro I have tested since 2000: Caldera eDesktop, Mandrake, Slackware, K12LTSP, Fedora, Red Hat, Ubuntu, Debian GNU/Linux and a few others I forget (failure of my memory, not the distros). Government, education, business, individuals, OEMs all use it successfully. Consider what some might call a failure on the desktop, Dell and Ubuntu. Just because Dell.com looks like a GNU/Linux desert means nothing. That’s in the home country of M$, the Great Satan of operating systems. Dell is selling GNU/Linux like hotcakes in China. It’s a wild success. They have 220 bricks-and-mortar stores pushing the product.

“millions of Dell’s products with the Ubuntu operating system will be sold every year, covering laptops, desktops, and servers.” see China Tech News.

Success of GNU/Linux is not limited to China, nor to commercial ventures. Many governments and educational organizations are adopting GNU/Linux on the desktop.

If Zemlin believes GNU/Linux has failed on the desktop, perhaps he should do something about promoting it instead of pronouncing “failure”. If he were looking for success rather than failure, perhaps he would have noticed the largest deployment ever, 500K in Brazilian schools. That’s from his own website, for Pity’s sake. Clearly, GNU/Linux is succeeding without much help from Zemlin.

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology. Bookmark the permalink.

31 Responses to Success of GNU/Linux on the Desktop

  1. Dr Loser says:

    @oiaohm:

    “Dr Loser QNX and Real-time Linux branch. Provide the same level of real-time. The quality of QNX has dropped a little since RIM took it over.”

    Oh yes?

    Your ignorance knows no bounds, does it, oiaohm?

    If you don’t mind, I’m just off to post this on TMR.

    Good Golly, you have outdone yourself.

  2. Yonah says:

    Urgh…. I had a feeling you’d do that right after I posted. Let me clarify my request. Please show me laptops being sold in any of the 220 bricks-and-mortar stores.

    Also, the SEC report does not even include the word “Linux”, or even “Windows” for that matter. Dell obviously sells systems with Windows per-installed. So, of that 14% increase, how much of that was due to sales of PCs and Laptops running Linux?

  3. Yonah wrote, “I searched for articles that confirm the laptops are on-sale now and are “selling like hotcakes”, but I didn’t find any.

    Robert, can you give me the articles you found?”

    see china.Dell.com
    Ubuntu Linux 10.10

    That’s a 14inch notebook, not a netbook…

    see also Dell Sells Ubuntu Linux-Loaded Computers in China in 220 retail stores.

    see SEC“Total revenue from Brazil, Russia, India, and China, which we refer to as “BRIC,” increased 14% year-over-year to 14.4% of our total net revenue for the third quarter of Fiscal 2012. We are continuing to expand into these and other emerging countries that represent the vast majority of the world’s population, tailor solutions to meet specific regional needs, and enhance relationships to provide customer choice and flexibility.”

  4. Yonah says:

    “Dell is selling GNU/Linux like hotcakes in China. It’s a wild success.”

    The article you gave us from October of last year only shows an announcement stating, “some 220 retail stores in China *will* feature Dell laptops running the Ubuntu Linux distribution.” I searched for articles that confirm the laptops are on-sale now and are “selling like hotcakes”, but I didn’t find any.

    Robert, can you give me the articles you found?

  5. oldman says:

    “The simple fact is the more you learn about MS Windows the more you hate it. The bugs are many. Traps in licensing is many”

    Very few people give a crap about the esoterica that you think you know sir. Even less will give a crap when they find out that they have to go through the dislocation of a new system/software and the result may just “suck less”.

    If you wish to waste your time hating, go right ahead. I am too busy getting work done and composing music to care.

  6. oiaohm says:

    The simple fact is the more you learn about MS Windows the more you hate it. The bugs are many. Traps in licensing is many.

  7. oiaohm says:

    Dr Loser QNX and Real-time Linux branch. Provide the same level of real-time. The quality of QNX has dropped a little since RIM took it over.

    If know your real-time OS’s the one you would have said is vxworks. That bugger beats both QNX and Linux kernel and yes when QNX is not good enough you are forced to migrate to vxworks. Even better a project started on Linux that Linux rt is not good enough you can just swap the vxworks kernel in. No code rebuilding in most cases.

    Anyone working serous-ally in real-time is going completely laugh at a statement like yours Dr Loser.

    Besides there are other options before Linux project has to migrate to vxworks.

    oldman particular with people like Dr Loser who repeatedly proves he does not know anything. So yes if you are wanting something that needs some real-time you will start with real-time Linux since migration up to more powerful real-time is dead simple.

    You have been no better oldman. Reactos developers have on going study on all versions.

    Last drivers I have build was for Windows 7. I have built drivers for Vista XP 2000 and NT as well. Also done some prototypes for Windows 8. So yes I know the internals. Reactos allows you understand deep way particular things in windows does not work how they are documented.

    Maybe I am giving FOSS an advantage because I know my stuff has this not crossed your mind oldman.

    Linux has me on path for fast and simple access to full blown real-time if the project needs it. Also cost effective if the project turns out not to require it.

    jackaudio is a particular case where the real-time does work.

  8. Kozmcrae says:

    “(Perhaps you could extend that to Oldman?)”

    I’m beginning to believe he likes being abused.

    “Is there any reason you think I have an agenda?”

    We all have agendas.

    ““Linking your argument to facts has been proven to have quite the opposite effect in changing someone’s mind in debates.”

    Really?”

    Yes.

  9. Linux can be configured to do things in real time. To a user, real time is “no perceptible delay” such as typing more than one character ahead or “no dropouts” having clicks and pops in multimedia outputs. GNU/Linux can do that. Lots of people use it for typing and multimedia.

    Unlike normal desktops that idle a lot, I have worked with GNU/Linux terminal servers and marvel at the illusion it gives multiple users of having total control of all the resources. I think I have had as many as 45 users on a single terminal server with a bunch of them in the same room as I and no one noticed the load that I could detect by running “top”. Modern hardware is powerful and GNU/Linux allows users to use it all. I once used an old machine as a terminal server and had 20 simultaneous users and a long wait-queue and it was still faster than XP on the old hardware that I was using as thin clients. The terminal server was 4 years old and the clients were 8-10 years old and the system ran like a Swiss watch.

  10. Dr Loser says:

    @Robert:

    “Dr Loser wrote, ‘Linux is not a real-time kernel’.

    I know what you mean by that but does it matter?”

    No, it doesn’t matter at all.

    But if people like oiaohm are going to throw phrases like “real-time” in, it would be good to have either

    * a definition of real-time

    or

    * a particular case where “real-time” Linux actually works.

    I recommend RiM/QNX, by the way. Unlike Linux, it does its job properly.

  11. Dr Loser says:

    @oioaom:

    “Dr Loser this is why you are screwed trying to convert me.”

    As Oldman says, why would I waste precious seconds of my life trying to do any such thing?

    I really hate to break it to you, kid, but you are one amongst billions. Not even a particularly important one — feel free to waste your life on Linux, it makes no difference whatsoever to me.

    Nope. What I am trying to do is to have a rational discussion.

    Try it some time! And let me know if I have converted you to the concept.

  12. oldman says:

    “Sorry Dr Loser the reason why I can hit windows I did spend time with Reactos developers studing windows internals.”

    It would seem to me that unless folks at reactos have access to windows code, anything that they and you might have come up with is pure speculation. It also seems to me that even if they gleaned some information by indirect study, that information would live only as long as the version inspected. All bets are off.

    “Its always fun watching Wintards talk they know crap about windows yet they think they are going to convert a person who has built drivers for the Windows kernels and knows the bugger inside and out.”

    I doubt sir that Dr. Loser is looking to convert you. You have made it quite clear by the tone of your posts at this site that you WILL always give linux and FOSS the benefit of the doubt because of your very real bias towards Linux/FOSS.

    As far as your alleged experience with windows internals is concerned, since when is writing a device driver make you anything other than an expert on the version that you are writing device drivers for?

    While it is possible to glean some information beyond that needed to write device drivers from MS rechnical and design docuiments. The fact remains that just as in the case of FOSS, unless you have access to the source, IMHO you are just whistling in the dark.

  13. oiaohm says:

    D-bus and Policykit work as one. D-bus is the mechanism Policykit is the policy files D-bus refers to if its allows something. So yes D-bus is policy controlled. It is a secure bus you cannot transfer data threw it unless policykit allows.
    http://en.wikipedia.org/wiki/PolicyKit

    Basically D-bus and Policykit are two sides of the same coin.

    Dr Loser
    “(1) Pre-emptive multi-tasking.
    (2) Hard priorities (implemented by the OS)
    (3) Guaranteed latencies (implemented by the OS)”
    These all exist in the real-time branch of the Linux kernel. Yes there is a interface in d-bus to control what applications get allocated. Guaranteed latencies is not as tight on a standard kernel.

    https://rt.wiki.kernel.org/ Guess what I use for audio work. We don’t use the standard kernel.

    By the way the goal is to merge all the rt tree into mainline at some-point.

    Default priorities in the Linux kernel are finer than Windows 7 and 8. Guaranteed Latencies in the real-time kernel are finer than Windows 7 and 8 with the default kernel. Real-time kernel allows more Guaranteed. Basically Windows 7 and 8 are beaten by the default Linux kernel.

    Besides there is no value to having all this if you cannot control it. ulatencyd the dbus service comes to the rescue here.

    ulatencyd is not like you normal dumb task managers it can be given exact orders to terminate applications in case of resource starvation so real-time requirements are meet. This thing can truly react to what is going on. Linux with this performs way differently. Windows boost for active window can be replicated by this.

    Also real-time kernel is not the most scary Linux kernel in existence.
    http://www.yl.is.s.u-tokyo.ac.jp/~tosh/kml/
    Real-time kernel + kml is the most scary Linux kernel in existence. Yes real-time process with no context switches. Yes it is possible to run lots of native Linux programs either in kernel mode or user-mode. So yes I could run dbus in kernel mode if there was some special need.

    upower is power management side to dbus allows users without privilege to control this.

    The issue has not been the Linux kernel ability to do stuff like real-time. But how to you expose the means to control the ability without breaching the complete secuirty of the OS.

    Only the system control applications have to use d-bus.

    Basically you want to keep on spreading myths what Linux can and cannot do. http://widefox.pbworks.com/w/page/8042320/Real-Time

    Dr Loser this is why you are screwed trying to convert me you keep on saying bull crap about what Linux can and cannot do. Please spend some time reading widefox before commenting. Yes Linux does have weaknesses but they are not were you are pointing.

    Basically no statements by idiots about Linux will make the Linux real-time kernel magically disappear. Dr Loser.

    By the way XML is only the interface defines of Dbus. The raw wire protocol contains no XML. IDL on windows use XML as well. So be very careful before you start throwing stones. Because if I believe your statement Windows is crap Dr Loser because it also uses XML interface defines.

    presence is referring to secuirty problem. You sudo you appear as root everything you do from now on you might as well be root logs will say its root.

    Something triggered by d-bus will trace back to you.

    Same issue exists in windows with run as administrator. Telling that X user was running it as administrators is impossible.

    Sorry Dr Loser the reason why I can hit windows I did spend time with Reactos developers studing windows internals.

    Its always fun watching Wintards talk they know crap about windows yet they think they are going to convert a person who has built drivers for the Windows kernels and knows the bugger inside and out.

  14. Dr Loser wrote, “Linux is not a real-time kernel”.

    I know what you mean by that but does it matter? When I was at Easterville, I played with several GNU/Linux terminal servers while students in my lab were using them. I deliberately did huge network file-transfers so the NICs on the terminal servers were maxed out and the CPU utilization rose significantly. Not one of two dozen students noticed any difference in performance. One server was serving 64K context switches per second and hundreds of processes were running. As long as it wasn’t swapping everyone felt like they had real-time control of their own PC. This was a stock installation of Ubuntu with no particular optimization of the kernel at all. It was running on a modest dual-core CPU. There was a significant wait-queue but real-time performance was unaffected. I know systems can be jerky with “sync” NFS and “atime” mounts and disc commits piling up but you have to work some to degrade performance. I put all of that on the file-server and the users’ processes ran very smoothly on the terminal servers. Modern hardware mostly idles with desktop usage and Linux can run it hard with no lagging.

  15. Dr Loser says:

    @Koz:

    “Linking your argument to facts has been proven to have quite the opposite effect in changing someone’s mind in debates.”

    Really?

    Forgive me, but I’ve gone beyond sophomore debates.

    In this particular case, I asked oioaohm for three hundred words and a link.

    He was kind enough to restrict himself to three hundred words (I can’t be bothered to check with wc, but close enough) and not one, but two links.

    Both of which I believe I have demolished.

    Your thoughts on those?

  16. Dr Loser says:

    @Koz:

    Once again, my thanks for keeping this civilised. (Perhaps you could extend that to Oldman?)

    “I see nothing to gain from providing links to my assertions for the most part. People who have a different agenda than you don’t care to see them so why bother.”

    Well, actually, it was oiaohm’s links. I asked him for three hundred words, and the lad obliged, which I think is pretty decent of him. We’ll make a wintard of him yet … (smiley).

    Is there any reason you think I have an agenda? Why? Just disagreeing with Robert doesn’t constitute an agenda. I’m actually all for open discussion, and I’m quite impressed that Robert (unlike most Loons) is up for that as well.

    But … from a purely academic point of view (forgive me O Lord I am an academic and the son and grandson of academics) you really need to provide a link for this stuff, you know. If only for the good of your academic little soul.

  17. Dr Loser says:

    And btw, which is where you are stuffed, oiaohm, despite your best intentions:

    D-bus is “mechanism, not policy,” isn’t it? Which means that every single FLOSS implementation out there (whether it be Nautilus, CUPS, JackAudio, PulseAudio or whatever. For all I know, the first four have embraced it, but I bet the other 3,000 haven’t) is free to ignore it.

    Mechanism, not policy, refutes the idea that Linux will ever be a worth-while operating system. It’s great for one-offs and lousy for anything usable.

    Oh, and that real-time thing? I’ve done it, and I doubt that you have. It requires:

    (1) Pre-emptive multi-tasking.
    (2) Hard priorities (implemented by the OS)
    (3) Guaranteed latencies (implemented by the OS)

    I’ll grant that it would be possible to invert the entire system and build a small Mach-like kernel based on D-bus which then calls into the generic Linux kernel. I’m not quite sure why you would want to do that, and the possible benefits are invisible to me, but whatever. It hasn’t been built yet, and I look forward to that day.

    Meanwhile: Linux is not a real-time kernel, and no amount of silly crap like D-bus will make it so.

  18. Kozmcrae says:

    “Since I have no prior knowledge of D-bus: anything at all, really.”

    That’s seeking knowledge and not seeking proof which is the only kind of “link please” I’ve ever seen on this blog or any other for that matter. I assumed you were referring to the latter. My mistake.

    I see nothing to gain from providing links to my assertions for the most part. People who have a different agenda than you don’t care to see them so why bother. The link to a fact won’t change their mind in the slightest *. Better to tell the to go fetch themselves. If they are really interested in another viewpoint, they will take the time and look with at least a smidgen of critical thought.

    * Linking your argument to facts has been proven to have quite the opposite effect in changing someone’s mind in debates.

  19. Dr Loser says:

    @oiaohm:

    Thanks for the links. You are evidently more professional than Koz.

    To your first link (freedesktop.org): this appears to imply that D-bus is a super-daemon, or, if you prefer, a watchdog over other daemons. The sole “benefit” evident is that it is defined by a single XML initialisation file, which I suppose takes care of all that rubbish under /init or whatever.

    Of all the cretinous things to pick … XML? Dearie me. Still, better than what it putatively replaces, I suppose.

    Other than that (which is actually a throwaway in your link) I cannot see the point of it.

    For those who are not prepared to follow the link:

    “Currently the communicating applications are on one computer, or through unencrypted TCP/IP suitable for use behind a firewall with shared NFS home directories. (Help wanted with better remote transports – the transport mechanism is well-abstracted and extensible.)”

    So, at this point it is a single computer abstraction without a point, aimed at people who cannot be bothered to write their own listen/accept protocol, or even be bothered to copy and paste somebody else’s. Nice!

    To your second link (Collabra, who I’m personally convinced have never made a dishonest dime by shilling this rubbish to the corporate innocents of this world):

    “This essential piece of Linux infrastructure connects everything from chat clients to the Nautilus file manager, the Common Unix Printing System, and music players. This means that businesses can create systems that interact with desktop applications, or create applications which leverage powerful services ranging from power management to real-time communications and presence.”

    It is by no means essential. There is no obvious reason to connect chat clients to Nautilus or CUP or music players. Even if there were, there is no evidence that D-bus is particularly pre-eminent in this way.

    “This means that businesses …”

    And once more we come back to the inconvenient little point that businesses just do not care. This is glue. Glue on its own is utterly worthless.

    And, as two bonuses:

    * power management: Linux has spent years failing at this. It has just about got up to speed by aping the way that M$ does it. I fail entirely to see how D-bus matters in this regard.

    * real-time communications and presence.

    Perhaps you could help me out with this “presence” thing. It sounds rather desperate to me. The immanence of the Lord, perhaps?

    Linux is not a real-time kernel (I don’t hold that against it) and Linux does not magically become a real-time kernel just because the proponents of D-bus say it does.

    Try QNX, for example. That is actually a real-time kernel. No D-bus required.

  20. Dr Loser says:

    @Koz:

    “Care to detail what you’d expect to find in the link that would change your thinking.”

    Since I have no prior knowledge of D-bus: anything at all, really.

    It must be nice to sit back and ignore anything the world might throw at you to challenge your assumptions, Koz. In a way, I envy you.

    But, absent a chemical cosh, I think I would prefer to ask experts (even dodgy experts like oiaohm) to help me out by offering a link.

    Not being a fanatic like you, I don’t expect a “link” to reinforce my beliefs. I don’t expect it to contradict them, either.

    I’d just like to be able to read it and draw my own conclusions.

  21. oiaohm says:

    Dr Loser really asking a lot d-bus is not simple item its a key historic change marker in mind set of Linux Desktop developers. To be correct dbus did exist 2003 but the desktop environments of 2003 were not using it. d-bus as a prototype appeared in 2002.

    http://www.freedesktop.org/wiki/Software/dbus
    http://www.collabora.com/projects/d-bus
    d-bus key reason is that the first of the merged subsystem between every desktop environment on Linux. Second reason is that is a secure bus. As you can see from collabora d-bus is the start of something. The birth of d-bus marks a change in the Linux desktop environment.

    Items already provided by d-bus. A way to set real-time on permitted processes. Unified package management interface not used everywhere yet but it a start there is an agreement between distribution to unify to a point. Unified network control in the form of networkmanager. Secure replacement to hald for plugged in devices.

    The idea to kill of as many sound servers as possible did not appear until 2006 part following what d-bus had started.

    D-bus is a foundation. Just like you cannot build a house without good foundations. Something like D-bus requirement to be a secure desktop and to share control of system-wide items between applications and users.

    Change to the Linux desktop over the next few years will be big. Particularly the target to get rid of X11 once and for all. Not all will be directly linked to what freedesktop started in d-bus but a lot will be. The effects of d-bus on the Linux desktop environment has not finished yet.

    d-bus provides that much opportunity I cannot really predict exactly what is full list of changes to the Linux Desktop will be in the end. Lot has changed a lot more will.

  22. Kozmcrae says:

    “a link or two would be nice”

    Care to detail what you’d expect to find in the link that would change your thinking.

  23. Dr Loser says:

    @oiaohm:

    I know it’s a lot to ask, but could you just take a single paragraph (three hundred words, max; a link or two would be nice) to explain your constantly repeated theory that D-bus is the saviour of tech-kind?

  24. Dr Loser says:

    @Robert:

    “I find it refreshing that some leaders care about such matters.”

    A peach, Robert: one of your better jokes.

    Unlike you, I am not privy to the inner conclave of the PRC, but it boggles my mind that the Chief Cadre In Charge Of Damming the Three Rivers might have a sensitive, caring, and above all Free side that causes him to mandate the use of GNU/Debian.

  25. Phenom wrote, “I always find it amusing how loons approve all anti-democratic measures of Putin, Chineese Communists Party and the alike to enforce Linux on they administration.”

    You mean, the same way leaders of democratic countries impose that other OS on their workers, by purchasing computers with it installed? I have never been in a workplace where the workers were consulted about what software would be on the PCs. I find it refreshing that some leaders care about such matters. Usually, they come to that point after years of lobbying by workers who do have some say.

  26. Phenom says:

    NT Jerkface, I always find it amusing how loons approve all anti-democratic measures of Putin, Chineese Communists Party and the alike to enforce Linux on they administration. Of course, it always end up with nothing but noise, and administration in these countries is still inefficient and highly corrupt. But that would never stop the loons.

  27. oiaohm says:

    notzed things have changed over the years.

    Remember 2003 there were 6 incompatible sound servers fighting for access to audio.

    Today we are down to 2 sound servers that only 2010 got to the point they can kinda get a long. jackaudio and pulseaudio are the last 2 sound servers on Linux. Recently they have got a workable relationship. Start jackaudio and pulse-audio gives way. No more having to mess around with your sound servers just because you want to do some proaudio work. When you stop using jackaudio pulse-audio nicely reconnects directly to the devices.

    Having dog fight between sound-servers leading to screwed up audio we can tick that off the list as fixed. Stability of pulseaudio that has to be left as work in progress. Stability of Jackaudio almost perfect. This is a vastly different picture to 2003.

    2003 arts esound nas… and others all not on speaking turns no automatic give way or talk to each other. So run a gnome application and esound could start grabing your audio now run a KDE application it would attempt to start arts result was no sound from that application if you were lucky or the application crash due to not being able to put out audio. Reverse happened as well. On top not one of those servers from 2003 was near stable.

    2003 audio on Linux was disaster zone looking like someone had been very busy with a wrecking ball. 2011 audio on linux looks a little like a construction site the buildings not fully complete but you can see form.

    Yes from a useability point of view Linux audio has improved leaps and bounds. Good enough maybe or maybe not.

    Don’t dispute from the outside some of the issues still look like what they were. Numbers of the audio issues have been reducing all the time from 2003 to now.

    If I go through pluggable buses the same is also true. You can see a plan that will deliver what is require where there was no such plan in 2003.

    Really the kernel is becoming seriously critical. Wayland the interface that is going to replace X11 depends on the kernel to manage the memory of the video card.

    The core of what make the Linux Desktop is changing. 2003 KDE and Gnome the olds of theme doing a joint project on anything was almost nil. Today there are many joint projects. Heck if we could time travel tell some of the 2003 KDE and Gnome people that they would both be working on joint projects with each other they would just laugh at you for stating the impossible.

    The idea that the Linux desktop has not progressed since 2003 is very wrong. Copy paste failures have reduced. Better wording of application descriptions.

    Common handling of networking so starting gnome after running kde does not lead to a networking failure(joys of 2003).

    D-bus did not even exist in 2003. Nothing like it did.

    This is the problem notzed you want to act like the Linux desktop has stood still. It has moved a long way.

    List of changes since 2003 are almost endless.

    The near future holds even more.

    Soon is going to move even more. systemd brings a big change of being able to trace what session an application is running it.

    KDE reducing there libraries also will see a lot of common faults go by by.

    Size of the trouble is reducing notzed at an accelerating rate. This is why I want to see how KDE 5.0 compares to Windows 8 when both are released in a formal paper.

    Simple truth notzed if someone did publish a formal paper right now comparing Linux 2003 to Linux 2011 you would look like a complete idiot due the the level of change that has really happened.

    Yes there has been improvement. Compare to Microsoft improvement over the same time is the interesting question. Is Linux catching or is Linux getting behind or worst possible that Linux has beat Microsoft and you guys have not noticed. Only one way to find out run a proper usability study.

    The past 2003 study Linux beat Windows in one section has Windows improved so it beats Linux completely or has it lost more sections to Linux.

    Yes a lot are saying the desktop is dead. To linux this is more than real. qt and GTK both support html5 application providing.

    The Linux world is changing remote desktop might interfaced by web browser in the near future.

    Let the battle for the cloud begin. Desktops on private cloud servers. I guess this sounds out-there but its where we are heading.

  28. notzed says:

    I still think around 2003 was the peak of stability and usability – all the faffing about since then hasn’t really moved anything forward, just sideways and often backwards. And I don’t just mean the ‘desktop’ applications: even linux has stability and compatibility issues from time to time.

    But for the head of a so-called ‘linux’ organisation to be talking about failure is total crap. OTOH linux is only a kernel so is irrelevent to a desktop beyond providing basic functionality, which it seems to have trouble with still: e.g. sound, and pluggable busses.

  29. Dr Loser says:

    I prefer the Charlie Brown take on that:

    “You learn from your mistakes? Really? Then I must be the smartest person in the world!”

    I believe Mr Brown is very close to perfecting the Linux Desktop, even as we speak.

  30. NT JERKFACE says:

    Every success is actually more of a success than anyone realizes.

    Every failure is actually a success.

    We never do anything wrong.

    Everything is going to plan.

    – Soviet Union 1922-1991

  31. Andrew says:

    He (Zemlin) doesn’t want to offend microsoft. In our home country (brasil), due to our president’s initiatives, there been a quiet uptake in GNU/Linux, specially in areas proprietary software vendors would not even venture to.

Leave a Reply