Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
Ubuntu Linux Business Red Hat Software Linux

Ubuntu Claims 12 Million Users — Before Lucid 360

darthcamaro writes "It's always a challenge to try and figure out how many users a particular Linux distro has — but Canonical is now providing a new figure for Ubuntu that is 50 percent more than what they were claiming just 18 months ago. 'We have no phone home or registration process, so it's always a guesstimate. But based on the same methodology that we came up with for the 2008 number, our present belief is that it's somewhere north of 12 million users at the moment,' Chris Kenyon, vice president for OEM at Canonical, told InternetNews.com. Just in case you were wondering, Fedora still claims more — actually almost double, at 24 million."
This discussion has been archived. No new comments can be posted.

Ubuntu Claims 12 Million Users — Before Lucid

Comments Filter:
  • NTP-servers... (Score:4, Interesting)

    by beaviz ( 314065 ) * on Wednesday April 07, 2010 @06:16PM (#31768334) Homepage Journal

    We have no phone home or registration process, so it's always a guesstimate.

    I always thought they used their NTP-servers to count installations...

    • by jon3k ( 691256 )
      I hope not, because all our desktops use an internal ntp server.
    • Re:NTP-servers... (Score:5, Interesting)

      by Runaway1956 ( 1322357 ) on Wednesday April 07, 2010 @06:47PM (#31768742) Homepage Journal

      "'We have no phone home or registration process,"

      Actually, there is an "opt-in" phone home process. Ubuntu has an option to participate in a software popularity contest thing. Those who opt in not only can be counted as using *buntu, but the poll tracks which software packages are installed.

      And, it is really "opt-in" because you are asked if you WANT TO participate or not. I'm almost certain that it defaults to "no", you have to click the "yes" button to participate.

      So, if this popularity thing tracks "x" million computers, it's pretty simple to double or triple that number, and claim "x times 3" installations.

      • Re:NTP-servers... (Score:5, Informative)

        by icebraining ( 1313345 ) on Wednesday April 07, 2010 @06:55PM (#31768846) Homepage

        I'm almost certain that it defaults to "no", you have to click the "yes" button to participate.

        It does, at least in Debian.

      • Re: (Score:2, Informative)

        by alex-tokar ( 1727590 )

        I'm almost certain that it defaults to "no", you have to click the "yes" button to participate.

        This [ubuntu.com] page says that the package is already installed on the system, but is disabled by default:

        This means that all you need to do is enable it.

      • by steveha ( 103154 )

        I'm almost certain that it defaults to "no", you have to click the "yes" button to participate.

        That is correct, for both Ubuntu and for Debian (upon which Ubuntu is based).

        In fact, in Ubuntu, the checkbox for this is only available if you click on a button that says "Advanced".


  • ... a few times to at least get some data to work with? If you're doing something like canonical is doing you think they'd want metrics.

    • by gstoddart ( 321705 ) on Wednesday April 07, 2010 @06:59PM (#31768912) Homepage

      ... a few times to at least get some data to work with? If you're doing something like canonical is doing you think they'd want metrics.

      And, if they had installed some phone-home pingy-thing, they'd be pilloried in the town square by people screaming about that when it came to light, and they'd be decried as violating people's privacy. Geeks on Slashdot frothing at the mouth and wielding torches, cats living with dogs, that kinda stuff.

      (And, before I get modded flamebait ... that's also half of the amusement of being here. =)

  • by electrosoccertux ( 874415 ) on Wednesday April 07, 2010 @06:20PM (#31768384)

    I'm not getting my hopes up until Netcraft confirms it.

    • Well, they did say it was *before* the new, just as ugly [slushdot.com], look.

      They said on their blogs that they had some professional artists. So why didn't they use them?

  • by Smivs ( 1197859 ) <smivs@smivsonline.co.uk> on Wednesday April 07, 2010 @06:21PM (#31768392) Homepage Journal

    I bet it's that guy next door with 12 million computers!

  • Some guesstimate? (Score:5, Insightful)

    by w0mprat ( 1317953 ) on Wednesday April 07, 2010 @06:29PM (#31768492)

    I have many problems with these numbers, how many of those are dual boot systems with Windows? I have three machines like that. I'm not sure of any reliable way to differentiate dedicated stand alone desktops. Ubuntu is the kind of thing I muck around with alot, people such as myself drag up the stats if they are trying to work it out from downloads, respository use stats.

    On the upside the total number of machines that have at least one linux distro on them must be rather higher than typical market share stats suggest.

    • Re:Some guesstimate? (Score:5, Interesting)

      by interkin3tic ( 1469267 ) on Wednesday April 07, 2010 @06:39PM (#31768616)

      how many of those are dual boot systems with Windows?

      You're saying a dual boot system shouldn't count as a user?

      I own a wii. It's been unplugged for over a year and I play the 360 every day, but I am still a wii owner. Similarly, it seems to me if you have a dual boot system with ubuntu and windows, you're still an ubuntu user. Maybe there are ubuntu purists out there who would look down on you for that and would care to distinguish between the two, I don't know.

      I'd wonder more about the second part you hinted at:

      I have three machines like that. I'm not sure of any reliable way to differentiate dedicated stand alone desktops.

      Would you count as 3 users for this number? This article [internetnews.com] mentions that fedora counts unique IP addresses, if it said how the ubuntu number was found, I missed it.

      • Re: (Score:3, Interesting)

        "Maybe there are ubuntu purists out there who would look down on you for that and would care to distinguish between the two, I don't know."

        Actually, I kind of look down on dual booting. It's rather silly, now that we have several methods of running virtual machines. Especially since running a VM means that you need almost no AV and malware security software running. If I get a drive-by infection, I can shut down the VM and restore it to a snapshot - no need to jump through hoops for half a day to clean t

        • Re: (Score:3, Interesting)

          by Joe Tie. ( 567096 )
          Except almost everything I use windows for comes down to gaming. The one thing a VM fails at. It's not getting ultimate output in FPS like you mention, it's getting games playable at all. At least at the moment, my experience has been that a VM won't give anything even close to the reliability of wine when it comes to gaming. And wine itself is a bit of a gamble there.
          • Re: (Score:2, Interesting)

            by nhytefall ( 1415959 )
            I find this interesting, and wonder - how many of the users are the inverse of those above?

            Specifically, I run Ubuntu only in a VM - under Win7. The only thing I use it for is SSH tunnelling... and it works beautifully for it.
        • by Hadlock ( 143607 )

          Some people like to play games with 3D graphics. Until VMware releases software that allows you to play the latest games under emulation at full screen and similar framerates to what you see natively under windows, needing to dual boot is still a valid argument.

        • by Nadaka ( 224565 )

          VM's are not that good at accessing hardware accelerated graphics. I keep windows around mostly for playing the occasional video game, that does not tend to work all that well in a vm.

        • by krelian ( 525362 )

          I only have a core 2 6320 but trying to do intensive tasks in a VM (like a compiling a large code base) still feels very sluggish.

          • Try a better VM solution. My KVM guests on a bog standard consumer core2 machine are as snappy as the native OS, counting CPU, RAM, disk and net access. Network access is actually faster from the virtual machines for reasons yet unknown to me. That's with Linux as a host though.
      • by Hadlock ( 143607 )

        I run Ubuntu on my netbook, but I only use it maybe 3 hrs a week unless it's a holiday where I'm traveling. Otherwise I am firmly a windows user for games. Should I count as a user? Listing the total number of users doesn't make a whole lot of sense if the total is dramatically larger than the number of actual users.

        • Re: (Score:3, Interesting)

          by pipatron ( 966506 )
          Yes, you're an Ubuntu user. You're also a Windows user. If you go to the gym for 3 hours/week you go to the gym. If you watch TV 3 hours/week you're a TV watcher.
      • by h4rr4r ( 612664 )

        I would be willing to buy your wii. Care to sell it?

    • by grege1 ( 1065244 ) on Wednesday April 07, 2010 @07:09PM (#31769032)
      What about all the machine bought with Windows that gets wiped and Ubuntu installed. Microsoft still counts them.
    • by carlzum ( 832868 )
      The actual number is not that interesting, but if they used the same methodology 18 months ago, it suggests there has been significant (albeit relative) growth. I was hoping the article would offer some insight, but if the VP of OEM can't explain it, we can assume Ubuntu laptop and device sales aren't the reason :) The increase is consistent with my observations though. Ubuntu has developed a positive reputation among casual PC users. Former Mac and Windows users I know get very excited about the Software
    • Well, I dual boot too but I think I am bringing up the Windows stats.

      I log into windows maybe two to five times a year (usually accompanied by a lot of loud cursing).

      But of course I paid for my copy of windows so MS counts me as a loyal customer.

      Anyways, my point is that that dual booting argument can go both ways.

    • On the upside the total number of machines that have at least one linux distro on them must be rather higher than typical market share stats suggest.


      Net Applications, builds its stats from hits to the immensely popular - mainstream - brand-name - sites of its very big corporate and governmental clients.

      It isn't easy to make a convincing argument that the average Linux user isn't pointing a browser in their direction.

      Shopping Amazon. Poking about the videos on YouTube.

      If the Moz Foundation doesn't find

  • by ls671 ( 1122017 ) on Wednesday April 07, 2010 @06:31PM (#31768510) Homepage

    I run Slackware but I masquerade my OS string as Ubuntu ;-))

    I like to masquerade all ID strings, masquerading apache as IIS, sendmail as JavaMail etc. etc.

  • Total guesstimate (Score:4, Insightful)

    by nurb432 ( 527695 ) on Wednesday April 07, 2010 @06:32PM (#31768520) Homepage Journal

    So why are we even discussing it?

  • by Zocalo ( 252965 ) on Wednesday April 07, 2010 @06:34PM (#31768562) Homepage
    I suspect someone's guesstimate may be off as just about every "most popular distro" statistic I've seen has consistently put Ubuntu ahead of Fedora pretty much since Ubuntu first arrived on the scene, except for brief periods immediately after new Fedora releases. Reconciling a 2:1 advantage for Fedora with that is kind of hard, but not impossible; lots of big corporates and SMEs use Red Hat, so Fedora would be a logical choice for their techies' personal use or installs where paying the Red Hat license fee isn't an option for whatever reason, and chances are they'd only download each release once. I'd guess that I used to run at a 6:1 install:download ratio when I was doing this with Fedora, and the German office did something simmilar with Novell/SuSE, so maybe both numbers are actually in the ballpark.

    Either way, these are not too shabby figures for Linux market penetration. I wonder how many of those installs are on the Desktop though? ;)
  • Number of computer users worldwide = 1.2 billion (taken from various estimates)

    Linux market share = 1.12% [wikipedia.org] (composite of various sources)

    Ubuntu market share = 50% of Linux (source = same Wikipedia article)

    This gives us 1.2 billion * 0.0112 * 0.5 = 7 million Ubuntu users worldwide.

    • by Lehk228 ( 705449 )
      Ubuntu market share = 50% of Linux (source = same Wikipedia article)

      you are better off pulling a number out of your ass, it will be more accurate.
      • by Zocalo ( 252965 )
        It depends what counts as a "Linux" install too. Linux isn't just the downloadable distros that you can install yourself; it includes all the Linux based devices like netbooks and countless other types of appliances that use highly customised versions of it. If you include the whole cornucopia of systems that use a Linux kernel at their core then I suspect that the most popular Linux "distro" might well turn out to be something completely unexpected and off the wall. As an example I've seen repeated clai
    • by sznupi ( 719324 ) on Wednesday April 07, 2010 @07:08PM (#31769018) Homepage

      Number of computer users is certainly noticeably different from the number of computers in use.

    • So you start from 4 numbers, all of them are estimates...

      Then you arrive at a number which is somewhat far off the "official" Ubuntu number....

      And from that you learn.... what ?

      Absolutely nothing.

      • Re: (Score:3, Funny)

        by El_Oscuro ( 1022477 )
        Didn't you know that 87% of all statistics are completely made up?
      • by Unoti ( 731964 )
        That's a little harsh. The original article was vague about how they arrived at the numbers. He showed his work and provided some references. His number was a small fraction of ubuntu's estimate. Then that sparked some conversation about why it might be so radically different-- for example pointing out that the number of people isn't necessarily the same as the number of installs. It's called contemplation, or discussion. Contemplation and discussion may annoy you enough to just bag on all topics of d
    • by h4rr4r ( 612664 )

      50%? No fucking way.
      It just is not that popular in the server room.

    • You should read some Wikipedia articles on uncertainty and how to take account of uncertainty when doing arithmetic.

      To summarize it, you do math with numbers with high uncertainty which means your results will have even higher uncertainty, which means that the Ubuntu estimates are probably much more realistic than your results.

  • Analyzing server logs could determine unique IPs that request data from security.ubuntu.com and if you harvest that data you know how many different Ubuntu systems are live within a period of time.

    Update Manager seems to retrieve a list of update servers at least once and that data is a good indicator of the number of installations, that's a good starting number then add some statistics on corporate usage where updates have been centralized, support contracts and you are starting to get a figure.

    Downloads a

    • I would think even better than unique IP addresses would be something like requests for an update to a component guaranteed to be installed in the base system like say grub or the kernel. That way, if you have 3 computers behind a NAT you will, hypothetically, eventually get 3 counts.
      • And, if you have to reinstall four times in two days, that'd be counted four times. Yes, I've had cases where that's happened. Especially if you're new to Linux. Or working with some particularly dodgy hardware. Or if you just screw up the installation medium in a way where it appears to install fine but then you discover later that some critical package is fucked. I've had all of those happen to me at one point or another.

    • I have 4 real and dozens of virtual machines all running one flavor or another of Linux, with Ubuntu being the most common (I also have 3 CentOS boxes of various versions, one current Fedora box, one Mandriva, one SuSE, one Gentoo, one Debian). The real boxes all run or another Ubuntu version, one LTS server, and three desktop alternate installs (2 laptops with whole disk encryption and one "server" with a 5TB software RAID-5 array).

      These are all behind one static IP address.

      So, am I one user? 4 users? Does

  • I don't use Ubuntu, but i have ~5 Debian boxen, and 3 or 4 OpenBSD boxes... but I'm only one user. so... Yeah, not entirely sure what I'm getting at.

  • Tell Linuxcounter (Score:4, Interesting)

    by markdavis ( 642305 ) on Wednesday April 07, 2010 @06:47PM (#31768736)

    Well, you better tell linuxcounter.org then! They estimate a total of 29 million Linux users world-wide. If just two distros- Unbuntu and Fedora claim 12 + 24 million, that is already 36 million, and you haven't even started counting Mandriva, SuSE, Debian, Mint, RedHat, or the dozens of smaller distros! If you believe all that, then MY estimate (more like guesstimate) would be close to 60-100 million Linux users.

    • Bizzare website. And their estimate hasn't changed since 2005. I switched to Linux since then -- apparently someone else moved back to Windows!

  • 12 million? Bad. That is peanuts compared to the number of people who use my OS! I estimate that 50 millions machines run BestDamnLinuxDistroEver version 2.3!

  • I'm sorry, but there is no way Fedora is more widely installed on PERSONAL COMPUTERS (versus institutional/industrial settings) than Ubuntu.

    • who said anything about the count being personal computers only.

      As for Fedora, the link in TFA (yes, I had to go read TFA to check you were wrong, thanks for that) counts the number of Fedora installs [fedoraproject.org] by tracking IPs making yum requests, downloads, opt-in phone-home calls, (they recognise requests from behind NAT and proxies as being offset by users with dynamic IPs, though they say they think corporate users behind NAT make up a larger number than personal users)

      For example, they get 50,000 direct download

      • The map you linked to shows hits for EPEL 5, which is their add on package repository for Red Hat Enterprise 5 (and clones like CentOS). Since Red Hat tends to be more server-oriented, the popularity of 64-bit makes sense.

        It looks like nobody's done any back-end changes for the maps in a while, since they only have them for Fedora 7, 8, and 9, in addition to EPEL 4 and 5.

    • I suggest you read through the Fedora mailing lists or look deeper at the fedora forums.

      I don't think the numbers listed are completely correct but you still see a lot of FC (fedora core) systems running out there unsupported, which was out before canonical was even around. That said many also masquerade stuff for RHEL also . Rawhide etc.

  • That's the real question, isn't it?

    • Why? How do you pay for Ubuntu? I knew they were offering support for people that wanted, but the vast vast majority of computer users regardless of operating system never use services like that, so I hope that's not what you meant.
  • by decipher_saint ( 72686 ) on Wednesday April 07, 2010 @07:04PM (#31768968)

    Based on estimates, I have had meaningful relationships with many, many hot ladies.

    *Estimates based on downloads.

  • Until they jacked up some updates. I left RedHat Open Source product after using RedHat since 4.2. Ran into dependency hell with Fedora Core. I went Gentoo for a while (love the speed) but got lazy and tried Ubuntu. It has been my primary desktop and netbook distro ever since. (8.04) Solid as a rock. I even do the distro upgrades after being paranoid and backing everything up, but the upgrades have been flawless.

    Still use RedHat Enterprise products with no major issues except the occasional hardware s

    • Re: (Score:2, Informative)

      by rubies ( 962985 )

      +1 to dependency hell, largely because of the copyright issues over media playback and the completely broken way that Fedora tried to get around it.

      And not just dependency hell, but that "SELinux" stuff that secures your OS by the simple act of not allowing anything at all to run, ever.

      Ubuntu netbook remix is a winner.

    • Re: (Score:3, Informative)

      by gdek ( 202709 )

      Actually, Fedora claims 24 million *active* users between Fedora 7 and Fedora 12 -- a timeframe well after you would have run into "dependency hell" issues.

      We actually document our methodology, too. Right here: http://fedoraproject.org/wiki/Statistics

      So your usage of the past tense is incorrect.

  • by jc42 ( 318812 ) on Wednesday April 07, 2010 @09:14PM (#31770416) Homepage Journal

    I'd also wonder: What percent of those linux boxes were bought with MS Windows installed, and are thus also counted a satisfied customers by Microsoft?

    (And they must be satisfied, since they aren't calling Customer Support. ;-)

    Actually, one of my two linux boxes is running Ubuntu, but it actially came with Ubuntu installed. The other was a castoff Windows machine from my wife ("required for work"), and is running a rather old Debian. It works fine as a gateway/router/server machine, even if it does have less than a GB of memory. Some of us benefit from MS's upgrade process that encourages customers to buy new hardware so often. But it does sorta rankle that MS and their fanboys count our machines as Windows machines.

  • by Anonymous Coward on Wednesday April 07, 2010 @09:27PM (#31770528)

    Here is the hit count for operating system from Wikimedia:


    Linux has a 1.65% market share and of that share Ubuntu has 0.71%. Ubuntu has approximately 43% market share among Linux users, which by a very large margin makes it the most popular distro.

I've noticed several design suggestions in your code.