On-line Poll: 33% Use LibreOffice in Finland

Tietokone IT magazing did a poll which shows that LibreOffice and OpenOffice.org combined have a bigger share than M$’s office suite and LibreOffice has 33% share. Not too shabby… Did they poll readers about GNU/Linux? Nope, but they did ask about FLOSS in government. 40% replied that government should prefer FLOSS. Asked if they use security software, 19% replied “no, because I use GNU/Linux”. It’s not scientific being self-selected but it is an indicator that LibreOffice and GNU/Linux are very healthy in Finland.

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 and tagged , , , . Bookmark the permalink.

57 Responses to On-line Poll: 33% Use LibreOffice in Finland

  1. oldman says:

    “So the real question now is are you willing to put your money where your mouth is. Chris Weig. Even if I pasted the numbers here a person like you would not believe me. So you have to see for yourself and the only way that is possible is pay.”

    Thats YOUR job Hamster, Not Mr. Weig’s. Unless you can share with us the actual numbers, then we only have your “word” that this exists. And given your past problems with facts, You word is worth NOTHING!

  2. oiaohm says:

    Chris Weig really there is a problem here. I am not the one who came here throwing stats without technical merit.

    http://www.idc.com/about/methodology.jsp?t=1348663708131#forecasting_metho

    Note I said surveys. Not survey. Every year IDC does many surveys at many physical locations face to face with people.

    Response rate really that is a good one. Some of the IDC surveys are mandatory.

    There are many thousand different ground surveys forms IDC do a year. The have Analyst for particular demographics that create surveys for that demographic specially. As well as doing research to size that demographic.

    Lets just that describing the IDC ground serveys into collated result in a blog post is impossible with full detail. There are many, many checks and balances. The even include how there selections could have been bias.

    The one thing about IDC ground surveys you know a individual has not been counted multi times by mistake. You have real head counts.

    Statistical population is true for IDC ground survey numbers that is not true for the web stats you have been using.

    Response rate this does alter based on the ground surveys you merge to build a complete demographic picture. Each demographic used comes from tax and census information. So you have a solid count of what each survey is trying to represent.

    All this fuses into a over all year report from IDC. In fact you can expand the range of there projected bias numbers.

    IDC design if you don’t add a demographic you will have a section marked unknown/error.

    Chris Weig
    –Linux is simply not happening on the desktop. That’s the only reality anyone should care about.–

    Yet you have nothing to back that. The IDC ground surveys are due to having access to their source data.

    Has anything else happened recently to undermine what you are talking about. Redhat and SUSE investing in major Linux desktop development. That is out of charter.

    Basically IDC ground surveys is true proper multi sampled in the physical world. Something web surveys should be trying to compare against to see how badly screwed up they are.

    Chris Weig I cannot in fact link to it since you need a IDC login to access it. Not like I will give you my username and password. This is a case you have to pay or person know someone who pays.

    So the real question now is are you willing to put your money where your mouth is. Chris Weig. Even if I pasted the numbers here a person like you would not believe me. So you have to see for yourself and the only way that is possible is pay.

  3. Chris Weig wrote, “And when Linux is at 1 percent, or at most at 2 percent, then it is not happening. Despite the fact that I, too, have seen real people use Linux”

    The North of Canada where I mostly taught is sparsely populated. Typically communities are about 50 miles apart, mostly close to bodies of water for transportation, and modern transportation is by air. From 2000 until 2011 when I last taught, I worked in a different community almost every year and I used GNU/Linux everywhere I went. Until 2003 I did not meet anyone who knew about GNU/Linux before I arrived in community. Since then everywhere I have gone I have found GNU/Linux used by some in the community. That’s a very crude poll. While IDC reported 3% and Google 1%, my crude poll shows even the remotest places in Canada had a few in 1000 (a few tenths of 1%) exposed to GNU/Linux before I arrived and roughly 25% after I left. So, GNU/Linux in education has probably been a major vector for transmission of knowledge of GNU/Linux. At first, I met adults who had used GNU/Linux in some college course or field of research but later I met students who had used GNU/Linux in other schools in the South.

    An example: Easterville, Manitoba is a town of a little over 1K people but the school had 500 students. Almost all of the students used GNU/Linux daily. The students from Grade 5 and up visited the GNU/Linux computer lab and the younger students visited the legacy OS. All classrooms had a few seats of GNU/Linux for students and teachers. So, more than one third of the population used GNU/Linux. There were only a handful of people there who had used GNU/Linux before the school was built. The whole town now knows about GNU/Linux. So, GNU/Linux is happening like this in every corner of the globe. One way or another GNU/Linux is making it on the desktop and servers of the world.

  4. oiaohm wrote, “Total cost is about $15 000 is what you need to spend to get access to all the good numbers.”

    I am thinking about doing a poll by telephone myself. For 100 calls I could probably tell whether 1% is nonsense in my local calling area. For 1000 calls I could have a digit or two correct. There are laws about such polls. I would have to check into it but it could give me something else to blog about this winter. It would take a bit of research to form polling questions and to generate a list of numbers to call. I expect I can identify numbers for Winnipeg because they recently filled up and added an new area code. I would not want to disrupt businesses. I will concede that business is locked-in to M$’s office suite here but that could be a sub-category or another poll. Perhaps I will generate a list of questions and post about it to receive comments on modifications.

  5. Chris Weig says:

    Mr. Pogson, I implore you: enlighten us with your intimate knowledge from the survey instead of paying only attention to the least important point of my comment.

    And when Linux is at 1 percent, or at most at 2 percent, then it is not happening. Despite the fact that I, too, have seen real people use Linux.

  6. Chris Weig denying that real people use GNU/Linux all over the world wrote, “Linux is simply not happening on the desktop. That’s the only reality anyone should care about.”

    I have seen GNU/Linux happening. People love the speed and efficiency of it. An anecdote: At Easterville where I equipped a whole new school with GNU/Linux, mostly thin clients, the fire suppression system went crazy and activated all the sprinklers. Students removed their jackets and flung them over the PCs preferring to get wet themselves rather than the GNU/Linux PCs being damaged. They might have done the same for that other OS but at least this showed they valued the system a lot. They cared about GNU/Linux on the desktop.

    GNU/Linux is definitely happening in education, government, science and technology, and, of course, on the majority of web servers where the world depends more on web applications than on local applications. It is happening on some but not all retail shelves and OEMs are shipping GNU/Linux in volume transactions.

    So, GNU/Linux on the desktop is happening. Could more happen? Yes. Canonical predicts 5% of PCs will ship with Ubuntu GNU/Linux next year. They claim 8 million shipped last year, about 3% of PCs shipped in 2011. That’s a big happening.

  7. Chris Weig says:

    Link to it or it didn’t happen. I don’t care if you have to pay for it or not. Apparently you and Mr. Pogson have intimate knowledge of the survey. So do give us the facts:

    – Statistical population?
    – Response rate?
    – Survey design?

    Linux is simply not happening on the desktop. That’s the only reality anyone should care about.

  8. oiaohm says:

    Chris Weig IDC survey from 2003. Ground surveys from IDC are paid for access. They are done yearly.

    Chris Weig the reality is that you have not spent the money for access to the numbers. Just like comscore you have to pay to get the good numbers.

    Chris Weig I have not disregarded the flaws in the mag poll. Interesting enough the people who respond are also the more out spoken marketers. Its the bias why its useful for future forecasts.

    Total cost is about $15 000 is what you need to spend to get access to all the good numbers.

  9. Chris Weig says:

    Quit running in circles. We have killed that myth years ago. IDC by survey determined GNU/Linux caught up to MacOS about 2003 and that was before the huge migrations and OEMs cranking PCs with GNU/Linux out by the millions.

    Link to the survey, please. The only IDC survey from 2003 which I remember concerned Linux on servers in Australia and New Zealand. Perhaps you mean this:

    http://www.macobserver.com/tmo/article/IDC_Says_Linux_Will_Pass_Mac_OS_Market_Share_By_2005_Possibly_End_Of_2003/

    A statement from ONE analyst at IDC, apparently based on guesswork, not on hard data. So… a lot like you.

    Where are these Linux computers anyway? I don’t see them around me. If Linux had something like 5 percent back in 2003, it easily should have 10 percent by now. That’s not backed up by anything. It’s not backed up by making observations in real life.

    I have presented unbiases webstats several times larger than 1%.

    Your “unbiased webstats” are not unbiased. You merely seek out webstats which prove your point, ignoring all their flaws. Happened just recently with your mentioning the Office poll in the Finnish IT magazine.

    What have you to support your opinion? Nothing but stuff known to be terribly biased like NetApplications that only counts GNU/Linux if Google runs it.

    You can scream “bias!” all you want. The fact is that three webstats — Wikipedia, Netmarketshare, StatCounter — by and large agree on Linux being down in the dumps. That’s a lot more than you have to offer.

  10. oiaohm says:

    Yep the IDC ground survey trolls stay well clear off that suggests Linux is between 5 to 8 percent. Same size as OS X.

    Reality Chris Weig the numbers are not on your side.

    comScores numbers that MS used in the best by presentation a while back to not sell Linux also suggests 5 to 8 percent desktop Linux again the same size as OS X.

    ComScores graphic is kinda unique between all stats companies. Its a donut Yes a circle with a hole in the centre style pie graphic. Its not a normal line or pie chart like any of the other stats use. So you can spot at times when ComScores data is being used.

    The reality the more reliable webstat don’t agree with the 1 percent either. Yes the ground stats and the pay money webstat basically agree. So method error in the others anyone?? They don’t take long to find.

    Microsoft avoid using ComScores numbers because it does not say what they want it to say. But they avoid using the other webstats more. Since somewhere like IDC ground surveys or ComScore could make a laughing stock out of them.

    Chris Weig I have told you what you have to get if you want to attempt to keep up the 1 percent arguement and when you do get it you will not be able to.

  11. Chris Weig wrote, “you can only say that with any conviction because you wish it to be true. But apart from flaky guesswork you have nothing to support your claim.”

    Quit running in circles. We have killed that myth years ago. IDC by survey determined GNU/Linux caught up to MacOS about 2003 and that was before the huge migrations and OEMs cranking PCs with GNU/Linux out by the millions. I have presented unbiases webstats several times larger than 1%. What have you to support your opinion? Nothing but stuff known to be terribly biased like NetApplications that only counts GNU/Linux if Google runs it.

  12. That Exploit Guy says:

    @Robert Pogson

    I was introduced to Linux back in the 90s via a friend of mine. If my memory is correct, he burned me a copy of Red Hat 4.2 (or was it 5.2?) and asked me to try it. I did. The software itself wasn’t really all that impressive to me in anyway (since I wasn’t familiar with Unix-like systems at the time), but given I was young and impressionable, I did the same thing you did and started preaching to everyone about Linux, FOSS etc. and used virtually every single catchphrase you have in your blog entries – “Wintel” (yes, it was a perjorative created in a time when the only ISA in a low enough price range and a reasonable built-up of software for personal computing was X86, but I am sure, with your “40 years of observation” and what not, know exactly why that’s the case), “M$”, “Internet Exploder” and so on. Of course, as time went by and I started learning more about Linux and Unix-like systems in gerneral, I eventually grew out of the advocacy. You know why? I started as a Windows user, took up being a Linux user, and, at the end, I understood why people preferred one OS to another and that it is not my decision as to what’s best for everyone. I don’t know what people think is best for them, and I certainly don’t have the rights to make that call. Reading your blog is like looking back at my younger self – an egotistical youngster who thinks he knows everything and has all the solutions to every problem.

    When you notice that people sharing the same rhetoric as yours are for some strange reason mostly at about 1/4 to 1/3 of your age, Mr. Pogson, you know it’s the time to go to the mirror and have a good look at yourself.

  13. Chris Weig says:

    The ~1% number is clearly too low.

    And you can only say that with any conviction because you wish it to be true. But apart from flaky guesswork you have nothing to support your claim.

    I will have to do a survey…

    I guess we don’t have to hold our breath then. Pogson methodology doesn’t work for anything.

  14. oiaohm wrote, “Wikimedia is useful if you are after the home user kind of numbers.”

    A lot of people do research in business and Wikipedia will be used. A lot of homes now have a wireless/wired router so the count of IP addresses is quite variable. Some ISPs give out two IP addresses by default but because of the routers only one is used. The average home in USA has more than 2 PCs. My home has 14 all behind a router with NAT turned on.

    All we can say is that web stats are utterly unreliable but it’s the best we’ve got. The ~1% number is clearly too low. There are probably that many GNU/Linux PCs just in thin clients. I will have to do a survey…

  15. oiaohm says:

    Robert Pogson with individual sites you still have to be-careful about collection method.

    Wikimedia bad if you want to know what the business usage of a OS is. Counting 1 per IP is bad that way.

    Wikimedia is useful if you are after the home user kind of numbers.

    The bias fact to each number is useful to know how Linux is going in each segment.

    The Trolls using these numbers don’t understand Bias that it might not be an complete Linux Desktop usage overview but its still saying something if you know the bias that is effecting the number from research.

    Wikimedia suggests home usage is about 1.55 percent Linux. That might be valid claim based on method and its bias.

    Problem is the 1.55 percent home usage does not tell you how big the business usage is. Guess want one is important for future direction of Linux.

    Effect of collection method alters who you see in the numbers and who is basically vaporised into statistical error. Interesting that everything wikipedia is quoting is bias to home user by method and sites.

  16. The web stats from individual trusted sites are much more reliable than these stupid “global” samples given by NetApplications etc. Of course they only reflect their region or subject area but when they are far above ~1% we have a very good indication that NetApplications is wrong. e.g. NetApplications shows Mac=6.7% when Apple itself only claims 4.5%. NetApplications shows Google = 900% of GNU/Linux use in California… NetApplications show not a blip in countries whose governments use and actively promote FLOSS.

  17. oiaohm says:

    Chris Weig you will find if you track my handle I have spoken against trusting those stats else where. This is not a unique case for me.

    –It’s not that you don’t trust statistics, you just don’t trust statistics which are unfavorable for you.–

    Reality I trust statistics that are favourable to me less. Because I am always looking for method bias. The more it goes the way I am suspecting the more I will look for method bias.

    Read what I said about the mag poll its only worth anything because its looking at a very particular demographic. Its what you use for long term forecasts that can be out wildly.

    Does that means I trust that survey. Not alone. Have there been other surveys done by other mags of the same class around the world showing the same trend yes. Is the a population overview no.

    I do I trust that stat to be 100 percent right no. Its only a guide for a long term forecast. That you will only find out if its right or wrong in 5 to 10 years time.

    –Tietokone IT magazine doing a cheap internet poll in Finland is methodologically sound all the way. We all know it. This great poll is to be taken at face value.–

    I have not taken that pole a face value and if you read Robert Pogson post neither has he.

    Chris Weig the reality here is you who come in claiming a face value we should trust stats that are not trust-able. You don’t appear in locations like the link you provided pointing out that.

    Robert Pogson I point part of this out to him a while back.

    Really the TM Repository who is meant to be against fud should have landed on this fact that web stats used by Microsoft Trolls, Linux People are in fact bogus. So been pulling those people up for using bogus numbers.

    They are so busy not doing there homework they fail to do what they about says they should be doing.

    Chris Weig really you have no right to laugh because you were doing exactly the same thing before I have dropped on you like a ton of bricks. You like many others have been trusting numbers at face value without doing any homework to work out what they can be trusted for.

    So I should be laugh at you for being such a twit Chris Weig. Since I am the one who as not made the error I have the right to laugh over it. You need to be taking a serous look at one self to make sure you don’t do it again.

  18. Chris Weig says:

    Sorry, correction necessary for clarification:

    As a matter of fact, let me remind your highness that you yourself conceded that few countries could influence the overall statistics, even if among those countries the USA is not:

    Becomes:

    As a matter of fact, let me remind your highness that you yourself conceded that a few countries could can, in fact, already influence the overall statistics, even if among those countries the USA is not:

  19. Chris Weig says:

    Gasp, my little friends, gasp. But you won’t escape the inevitable truth, no matter how much you bullshit around.

    It’s not that you don’t trust statistics, you just don’t trust statistics which are unfavorable for you.

    It’s funny though, that StatCounter could pick up on the usage of Linux rising in Venezuela, which — for once — can be assumed to coincide with facts, as FLOSS has been prioritized by presidential decree there.

    So it’s a typical case of hypocrisy.

    Tietokone IT magazine doing a cheap internet poll in Finland is methodologically sound all the way. We all know it. This great poll is to be taken at face value.

    Mr. Pogson spouts ridiculous things about Wikimedia stats being biased in favor of the oh-so-locked-in USA. Real pity, because page views originating from the USA only account for a bit less than 25 percent of all page views. So plenty of other countries can influence Wikimedia’s numbers for Linux.

    As a matter of fact, let me remind your highness that you yourself conceded that few countries could influence the overall statistics, even if among those countries the USA is not:

    http://www.linuxinsider.com/story/74186.html

    Wow, how you all jumped at the fantastic increase to 1.4 percent.

    So you distrust any unfavorable statistics, but peruse the very same statistics to “prove” a point.

    I can’t stop laughing.

  20. oiaohm says:

    Robert Pogson up until 2004 we had a very serous number to work from.

    I will believe that Linux was 1 percent in 2004. The number comes from a big enough collector that it has to be pretty down right close.

    Robert Pogson google stopped generating the collated data and went to per site data in 2004. This is about privacy. To get good global stats you have to place a unique tracking cookie on each visitor and collate all that data into one.

    Google could start generating the data again but they would have to over turn there privacy policy call on it. Because you don’t want to count 1 user 60 times because that 1 user visited 60 different sites.

    Google could produce a per site average. Would not be perfect but I would trust it way more than stats counter.

    Google is not a NDA issue they are just not generating the stats. What was the point did not do Google any good to produce correct numbers. Other parties were generating the stats that were crap and people were comparing the Google stats to them and claiming google had it be bias because others did not agree with Googles numbers.

    So basically Trolls nuked our source of good numbers on what the heck is going on. Same trolls normally who are using the bogus sites.

    Net Market Share is very careful not to allow you to roll back to 2004 where you can put the google numbers next to theirs and show they were under counting Linux.

    Microsoft people were pushing the Net Market Share stat over using the Google one in 2004 because Net Market Share was lower.

    Chris Weig I use to have good stats to work with but due to people like you quote crap stats I don’t have good stats to work with any more.

    Now the only way the companies with good stats are going to release them is if people stop going to the firms who really cannot deliver.

    Google Analytics and Quantcast Measurement are the only two who have formal certification by many countries measuring standard bodies.

    Quantcast Measurement due to possible error level are not willing to put their name to global stats. The will give you how many visits to your site you have or sites with high standard of quality.

    You would not use a ruler that is someone magically dreamed up measuring system that conforms to no standard. This is what you are doing Chris Weig.

    The other 3 in top 5 have certification from at least one countries measuring standard body.

    The rest after the top 5 that are not google(Double click and other sub Google products that appear lower in list) has not had their method approved or checked by a standard body for measurements. That is another reason why if you want to quote a number it has to come from the top 5.

    http://trends.builtwith.com/analytics

    The reality is unless you can get 1 of the 5 to publish you really don’t have global stats.

    Chris Weig that is the problem if you want proper numbers you have bugger all choice. comScore is it. If you are not quoting comScore global numbers as global numbers you are 100 percent sure quoting bogus. The reality of they being only one choice at the moment generating global stats by a method that is somewhere near sane. They have the volume of input as well to back it up. Not as good as google. I will be willing to accept comScore numbers.

    Now of course you could put pressure on Facebook and Google to start generating numbers again.

    This is the problem about me I have done my homework trying to get proper stats to work with. It really pisses me off that we don’t have proper numbers to work with.

  21. oiaohm wrote, “no one is paying for the report”.

    Not true. That the reports are produced means someone pays for them. It’s the folks rolling in dough like M$ and “partners” that can afford that. Google has it as well but why give it away when they can sell it? I bet the reports have an NDA or such to prevent public release. Still you would expect “according to Google …” reporting but I have not seen much. Google used to provide stats as a free service in the old days.
    see Nov 2001

    and
    Nov 2003
    and the last one I see is June 2004.

    I guess they got tired of publishing “1%”…

  22. oiaohm says:

    Chris Weig comparing statistically garbage numbers to each other. Don’t magically make them correct.

    If Linux is really 1 percent. I would like to see this from a system that can kinda tell us. That would be someone pay comscore to publish it. They would have the data. Just no one is paying for the report.

  23. oiaohm says:

    Chris Weig

    –The difference between Wikipedia, StatCounter and Netmarketshare is negligible for our purposes. All three statistics agree on the same thing: that the Linux desktop market share is flat and not higher than at most 2 percent.–

    Problem is we know Wikipedia does not count business users correctly. Neither does Netmarketshare or StatCounter.

    If you want to say at most 2 percent in the home desktop market you would get away with that. Its kinda what we would be expecting.

    Chris Weig wrote, “They even want to make us believe that Wikipedia is biased against Linux!”

    Excuse me. I am not saying Wikipedia is bias against Linux. I am saying Wikipedia is bias against a particular demographics. Wikipedia is bias against Business/School/Government Desktop demographics. So Wikipedia will show the home desktop demographic ahead everything else. The home desktop is after an al-rounder. This is a very hard egg to crack. Business 80/20 splits are possible by assign Linux to users that are doing roles that don’t require everything.

    What is the demographics that has the money to fund Linux Desktop Development. Business/Government Desktop. Its not like the home desktop is going to fuel Linux Desktop Development. High successful FOSS projects are that way because Business or Governments depend on them.

    Chris Weig the stats you are using are not telling me the number I need to know to know what way Linux is going. As one of the Microsoft people said. Developers Developers Developers. Do you think home users go out and pay for Developers directly Chris Weig. You are look for directly employed developers for desktop internally you are looking at business and government.

    Chris Weig FOSS every year sees more full time developers join it from new never seen a developer from before companies. In the last 5 years most of these have been going to desktop related projects. The 5 years before that most went to server related products. We know there has been a shift in the business and government desktops. Problem is we don’t have any proper stats to see how much.

  24. Chris Weig wrote, “They even want to make us believe that Wikipedia is biased against Linux!”

    That’s nonsense. The USA is the largest part of Wikipedia’s stats. The USA is hugely locked in to GNU/Linux in business and retail shelves are also locked so it’s a captive market for M$. Wikipedia is reflecting that not, the prevalence of GNU/Linux in the world but in one dirty little corner of it. Wikipedia tells us nothing about prevalence in Brazil or Russia or China or Malaysia. Wikipedia is not against GNU/Linux. It runs on it. Wikipedia can only report the data it sees from those who connect. Most of the world does not speak English and does not connect. If they broke down stats by Language or region, we would have something worth discussing.

  25. Chris Weig says:

    The difference between Wikipedia, StatCounter and Netmarketshare is negligible for our purposes. All three statistics agree on the same thing: that the Linux desktop market share is flat and not higher than at most 2 percent. And rounding up to 2 percent means giving Linux the benefit of the doubt already.

    Mr. Pogson and oiaohm are like fish out of the water. And now they gasp for air. Because they simply can’t accept that three web statistics generally agree: Desktop Linux is not happening, and it’s certainly not growing. They even want to make us believe that Wikipedia is biased against Linux! Is there a special Linux Wikipedia version that is only accessible from Linux computers? Show me where it is.

    Robert Pogson accepts a methodologically not sound pro-LibreOffice/OpenOffice survey without second thoughts. But three big web statistics which aren’t just guesswork, and which all basically say the same thing, are not enough.

    Face the music, lads, instead of searching for ever more pitiful excuses.

  26. iLia is now banned. He/she/it dropped links to racial hate sites demanding to be banned.

  27. oiaohm says:

    iLia what are you trying to make a joke out yourself.

    Dec 13, 2005 is the last update on http://www.twospots.com. Its not current data.

    I cannot read Russian but after that goof I could not exactly trust you quoted sites as showing current data.

    iLia really you want some kinda valid numbers pay comscore. If not just accept you don’t have them.

  28. oiaohm says:

    Lets to the wikipedia OS usage stats.
    http://trends.builtwith.com/analytics
    Net Market Share.
    Too small 40 000 sites does not even show in the top million sites.
    W3Counter
    These guys are not game to say how many sites they have builtwith only detect 4000 in the top 1 million active sites. So way too small most likely smaller than Net Market Share this is why they are most likely not game to say.
    Wikimedia
    Too small to register in number of locations but due to being fairy demographic netural it might work. Why it cannot work collection error in method of reducing to IP so 10 computers from 1 IP appear to be one.

    StatCounter Global Stats
    Registers but has a collection error in method. 1.5 percent you need to be targeted and to know the difference between a person visiting a site 10 times and 10 people visiting the site.

    So all the data here
    http://en.wikipedia.org/wiki/Usage_share_of_operating_systems
    This basically worthless to tell you anything about OS numbers on the desktop that are out there. They are either too small or collection method buggy.

    Reality web stats are so bogus its not funny little homework shows it.

  29. iLia says:

    They are seriously flawed, inconsistent and biased.

    Ok, I admit that

    Wikimedia 1.55%
    Statcounter 0.84%
    w3counter 1.73%

    and http://www.twospots.com

    Linux 0%

    are nothing but Microsoft shills, so are hotlog.ru:

    Linux 0.89%
    Windows 8 — 0.28 %

    Wow, Linux in Russia is 3 times more popular then “the biggest Microsoft mistake ever”, what a achievement.

  30. kozmcrae says:

    Who is desperate here? In just a short time 24 comments (as of this post) appear. The Cult of Microsoft really don’t want to see posts like this one. In their world it must be knocked down immediately and with prejudice. I see TM Repository reverted to his Thorsten Rahn personality.

  31. oiaohm says:

    Really I am sick of playing nice with you Chris Weig so I will now point out why you are being such a idiot.
    http://trends.builtwith.com/analytics
    Of course you would not be aware of builtwith that tells you what analytics solutions in use where.

    Anyone who is trusting analytics should have checked out how much real market share it has. You have not Chris Weig.

    You know how stats counter mentions in use on 3 million sites. As I said that is only 1.5 percent bugger all.

    61.97% – Google Analytics
    21.16% – Quantcast Measurement
    16.32% – Facebook Domain Insights
    13.35% – comScore
    12.03% – Omniture SiteCatalyst

    Now those would be the only ones a sane person would look at. Now out of that 5 only one does global stats and the bugger charges for them. comScore. Why does it charge for them it knows it has the best numbers you can get. Those 5 also have sites in the top 10 000.

    The top two would be preferred but they don’t do up global stats. Quantcast Measurement in fact states they are not willing to publish them because they cannot trust the quality due to the error factors they know. The numbers are far from perfect and they are truthful about it.

    WordPresses stat counter is in use on 56,069,943 sites. Many times larger right Statscounter in fact in the top 1 million sites Statscounter and WordPresses are about equal. Problem is this again does not attempt to produce global common numbers its not really large enough.

    Yes http://piwik.org. the open source solution is listed before Net Market Share in fact Net Market Share is that small it don’t register at all. Yet people quote it like its meant to mean something.

    I have already told you the wikimedia method error. Its bias to home users. Locations with more than 1 machine to a IP only register as one on the wikimedia stats.

  32. Phenom says:

    TM Repository is now banned. Get lost. Good riddance.

    You are a coward, sad old man.

    Please don’t bother banning me. I will not visit this site again anyway.

  33. According to Netcraft, Google’s netblock is 64K addresses.

    That’s nowhere near enough to amount to 8% of share in California with 37million people and probably more PCs than people.

  34. TM Repository is now banned. Get lost. Good riddance.

  35. iLia wrote, “data from different sites and countries correlates quite well shows that this sites use quite accurate methods of gathering information.”

    ROFL!!!

    Wikimedia 1.55%
    Statcounter 0.84%
    w3counter 1.73%

    With the total number of sites counted the accuracy should be of the order of +/- 0.001% but they are all over the map. I have showed that NetApplications seriously overcounts Google so that with and without Sunnyvale numbers swing from 9% to 1%. So the methods are not “quite accurate”. They are seriously flawed, inconsistent and biased. I have shown webstats from individual sites without obvious bias and they are far above those numbers. Dell is selling GNU/Linux in ~1000 stores in China and India, countries still deficient in PCs yet the stats from those sources show little or no GNU/Linux.

    Numbers from surveys back in 2003 showed ~3% and GNU/Linux was not nearly as popular then as it is now so the real number should be much higher than 3%.

    Example:http://news.softpedia.com/news/An-Entire-Californian-School-District-Is-Saying-039-Good-Bye-Windows-Hello-Linux-039-48523.shtml In 2007 Windsor Unified School District switched to GNU/Linux thin clients. Are they counted? Nope. It’s just a few terminal serving machines to the web stats. That was 5K machines.

    Santa Rosa did the same in 2009.

    Dysart Unified School District switched 3K PCs to GNU/Linux in 2009.

    In 2008, San Diego Unified School District switched to buying GNU/Linux Lenovo ThinkPad R61. It’s the second largest school division with over 100K students. Later in fall 2009 they switched to netbooks with either GNU/Linux or XP but they still have a large number of GNU/Linux machines in use. Not even a blip shows. Net Market Share for San Diego, 2009, Linux 0.43%, less than they show today.

    So, how do all these GNU/Linux PCs in California Schools count way less than a few thousand at Google?

  36. Chris Weig says:

    StatCounter is a service which has to be installed by the users themselves on their websites. So unless there is some very universal bias against using it among webmasters whose sites are predominantly visited by Linux users, StatCounter’s methodology — despite the obvious flaws inherent in nearly any web statistics — offers much better results than the pure guesswork of Mr. Pogson & Co. which hasn’t any methodological basis to begin with. Also, StatCounter delivers quite distinct per-country patterns for Linux. For example, in Venezuela Linux’s market share has nearly doubled during the last 12 months according to StatCounter. What’s more likely: that this is real or that for no discernible reason websites visited almost exclusively by Linux users suddenly installed StatCounter during the last 12 months? But if Venezuela’s results are real, then the other results are equally real.
    And by and large these results show that in almost all countries Linux is flat, hovering at around one percent. And this is consistent with Wikimedia’s statistics which show Linux at around 1.5 percent. Wikipedia is unarguably a website with a phenomenal outreach, not least since for many, many searches on Google Wikipedia results appear in the top spot. So if you distrust their statistics — do Linux users use Wikipedia significantly less often? I don’t think so. — you should simply call it quits and trust NO statistics at all. Then you’re back at Pogson’s random guesswork.

  37. iLia says:

    What about the rest of there numbers again nothing to support them.

    Why you are not telling the same things to our dear Mr.Pogson, when he brings forward some Linux-friendly statistics?

    And if you don’t like the data I gave, I can give you some more:

    http://en.wikipedia.org/wiki/Usage_share_of_operating_systems (Wikipedia)

    Operating System Market Share (Netmarketshare)

    And the fact that the data from different sites and countries correlates quite well shows that this sites use quite accurate methods of gathering information.

    Or that there is some kind of global conspiracy concerning Linux.

  38. oiaohm says:

    iLia go read the statcounter.com faq as per comment 7 I put here.

    Do they tell you what sites they are monitoring in each country. In fact they don’t.

    Basically go back and read comment 7. There is a collection error iLia.

    For a rough count of bug to number causes in each country http://www.proxynova.com/proxy-server-list/

    iLia the exactly what proof do you have that statcounter has the right sites in Malaysia to even get correct value of usage. That is right nothing.

    What about the rest of there numbers again nothing to support them.

    iLia reality those so called world wide numbers have no scientific method to them and lack the data to confirm how true or false they are.

    The russia one I need the method used in english or at least somewhere a person can conform a google translate or something else is not altering meaning of the method to see how good of quality its numbers are.

  39. iLia says:

    Linux doesn’t even appear in its worldwide OS statistics.

    Mr.Pogson you once said that Linux is popular in Russia, sorry but you war wrong — in Russia Linux has only — %1.02

    And Russian site openstat.ru sais that Linux is used by 4.08%, BUT, 2.88% are actually Android users, so 4.08% – 2.88% = %1.2

    Ubuntu 0.31% (77’735)
    Mint 0.01% (1’704)
    Fedora 0.01% (1’264)

    And some others, but I am not sorry to say but no Debian here.

    And my favorite:

    Os/2 — 0.0% (77)

    There are 77 people who use OS/2 in Russia! I like it!

    And what about Malaysia? Sorry no Linux here, only 1.76% of ‘other’.

    Mexico — 1.14% of ‘other’

    Argentina — 0.73%

    Paraguay — 1.17%

    But you can be happy in Spain 1.94% use Linux.

    In Finland 2.2%. (it seams to me that the fins use LibreOffice/OpenOffice.org on Windows)

    You can find more statistics here

  40. oiaohm says:

    TM Repository
    http://googleiprange.hapara.com/
    That is not all the google owned IP blocks.

    Yes the issue here is goggle has a huge number of external facing IP addresses. Some of those blocks are for development staff testing out new versions of the spiders and other items from their machines.

    Its not all Google staff get straight up external IP addresses. Its there development teams. For some users giving them external IPs so they can straight up test stuff improves productivity. But compared to most businesses google have external IP addresses many times over. Also more staff have access to a true external IP. When you think what they are developing it makes perfect sense.

    The reality what I am saying here is not a lie.

    TM Repository internal survey of google shows the desktops are Linux and the laptops are mostly Apple.

    http://en.wikipedia.org/wiki/Goobuntu So its basically 50 percent Goobuntu with 49 percent Apple with MS getting a little 1 percent inside google.

    Yes and the split is warped. Laptops vs Desktop split is quite extreme at google.

    TM Repository goggle does internal OS usage serveys if you did around you will find them. This you will find what you were seeing is only half the story in goggle. I guess you were in a room with laptops no google desktops to be seen.

    Yes google warpness that laptops and desktops are basically completely different OS’s.

  41. Oiaohm, is a complete and utter liar! His posts are based on fantasy and he attempts to bury readers in a mountain of text. Provide evidence or shut the hell up. You ridiculous, contemptuous LIAR.

    You’re a mockery of anyone who cares about open source, you’re a mockery of anyone who cares about Linux. You mock Pogson though he doesn’t even realize it. You are a horrible, lying, panderous asshole who will misrepresent and misinform no matter what the cost. You’re like a reverse troll; You harm that which you claim to love much better than any real troll could!

  42. “There is a reason. So if there is a error with a google site. The employees see the site from external ip just like there customers. So Google has a IP block they just use for their staff. Most business cannot afford to go out and buy a full IP block just for them. Let alone just so there staff can have external IP addresses.”

    What a utter LIE! Google does NOT BUY IP BLOCKS FOR THEIR EMPLOYEES! Please, oh please, show me some indication that this is true. I would love if you, oiaohm, prove even one of your fantasies to us with some EVIDENCE!

    Do you hear me? EVIDENCE! Show us you LIAR! Prove that you aren’t just imagining scenarios!

  43. “How does anyone explain a few thousand Google employees running GNU/Linux having such a big effect in the great state of California?”

    When I was at the Googleplex for DjangoCon, most Google employees were running macs. For example, the inventor of Python, Guido van Rossum. Funny enough, I was one of maybe two PCs at the conference.

  44. “TM Repository hey cut some slack here.

    –It’s not scientific being self-selected but it is an indicator that LibreOffice and GNU/Linux are very healthy in Finland.–”

    No, it’s an indication that it’s healthy within the readers of this tech magazine.

    I could pull equally bullshit stats by claiming that 90% of finland plays Halo 2 because a Finish gaming magazine did a reader poll asking what their favourite game was!

    Equally, if you read a magazine like MacLife, would you expect their reader poll to show positive stats for Linux OR Windows?

  45. oiaohm says:

    Robert Pogson
    –Does Google give its employees unique IP addresses?–
    There is a reason. So if there is a error with a google site. The employees see the site from external ip just like there customers. So Google has a IP block they just use for their staff. Most business cannot afford to go out and buy a full IP block just for them. Let alone just so there staff can have external IP addresses.

    Google is a true abnormality. Runs like no other business in more ways than 1.

    Other thing is the google browser in the Google linux automatically disposes of cookies. So anything using cookies to determine unique visitors is screwed.

    The thing is Google showed how normal business are not being counted. Because bigger migrations in areas caused no changes in stats.

    StatCounter would count number of google users close to same as home users but but under counting business.

    netmarketshare would go off deep end and over count.

    This is thing once you start reading method you start understanding why these massive changes appear to happen. Also you start to understand how these web stats are failing to count business.

  46. oiaohm says:

    Chris Weig
    –That’s in direct opposition then to what Mr. Pogson claims.–

    This is because I am not guessing I am looking at the method. Then I look at how home and business networks are commonly setup. Then factoring how that method will error. Not if error will error. Googles is not a common setup. Mr Pogson based a theory from a Google move where it shows up as a large number appear. This is due to how many IP address google has and how many of there users get direct Internet no proxy. This is a google abnormality.

    –Chris Weig the bias is mostly to home not business. Because when you are at work you only visit a limited number of sites because you are more bussy doing you job.–
    This is a natural bias statement. People visit more web sites at home than at work. This is a known bias.

  47. How does anyone explain a few thousand Google employees running GNU/Linux having such a big effect in the great state of California?

    Does Google give its employees unique IP addresses? Do they run multiple virtual machines all day to count more? I did a test today and the GNU/Linux score is 9% with Sunnyvale and 1% without it, in a state of 37million people. There are hundreds of thousands if not millions of GNU/Linux users in California.

  48. oiaohm says:

    http://www.netmarketshare.com/faq.aspx

    Yes netmarketshare methodology is even worse.

    40 000 sites.
    76% participate in pay per click programs to drive traffic to their sites.
    43% are commerce sites
    18% are corporate sites
    10% are content sites
    29% classify themselves as other (includes gov, org, search engine marketers etc..)

    So there is every chance no FOSS sites are include in netmarketshare.

    40 000 sites basically is does not register to give quality numbers yet you quoted them. Chris Weig.

    Nothing you are quote is large enough or sourced from a non bias enough source to be sure you are getting a proper cross section.

  49. Chris Weig says:

    Chris Weig the bias is mostly to home not business. Because when you are at work you only visit a limited number of sites because you are more bussy doing you job.

    That’s in direct opposition then to what Mr. Pogson claims.

  50. oiaohm says:

    Chris Weig that does not change anything. We don’t have the demographics of statcounters sites and they do not have 10 percent + of sites in existence providing them with numbers.

    Chris Weig the bias is mostly to home not business. Because when you are at work you only visit a limited number of sites because you are more bussy doing you job.

    ==We measure internet usage trends. To accurately measure usage, we have to base our stats on page views (and not unique visitors). Let’s look at an example:==
    From the FAQ. They use nothing to solve to unique. This method also means a people in a company behind a proxy record as less web hits.

    Chris Weig take netcraft number of sites and work out the possible error factor to statcounter its still worthless.

    http://gs.statcounter.com/faq#methodology
    –Our tracking code is installed on more than 3 million sites globally.– Not enough sites

    –In August 2012, our global sample consisted of 17.1 billion page views (US: 4.2 billion); 2.5 billion of these were search engine referrals (US: 524 million); 262 million of these were social media referrals (US: 121 million). —

    Break down does not show intentional selection to avoid bias. Search engine referrals to the sites they are on. Search engine referrals to sites they are not on don’t show up at all.

    Huge stack of social media that targets more the home user than business user.

    –I don’t think that business employees suddenly would change their behavior just because they’re using a Linux desktop.–

    They are not seen effective either way by StatCounter numbers if you had read there FAQ.

    Chris Weig thinking lot of the recorded Linux migrations are business and there bias to home I would expect StatCounter to show nothing.

    If you people would read the FAQ’s about the stats you were using you would know they are bogus and stop using them.

  51. Chris Weig says:

    The reality is we are not getting the numbers we need to know what is going on.

    We don’t need to know. If the statistics cited above indeed have a bias towards business, then Linux numbers should rise. If, as Mr. Pogson always claims, there’s indeed a big migration towards Linux going on in the business world. I don’t think that business employees suddenly would change their behavior just because they’re using a Linux desktop.

    So the ugly truth is probably that widespread Linux migration in the business world is nothing but hot air.

    Oh, and since StatCounter has a lot more sites tracked than Netmarketshare, its data is probably nearer to the truth. So much so that Linux doesn’t even appear in its worldwide OS statistics. That’s about right then. Linux is meaningless on a global level.

  52. oiaohm says:

    Chris Weig really StatCounter and Netmarketshare don’t provide any long term growth information because you don’t have the demographics of the people those numbers are from. For how large there possible effect on the user base is.

    Chris Weig you would expect to see as per all Linux Migrations that have been successful and uptake of OpenOffice or LibreOffice to proceed it.

    Again your numbers are lacking the numbers we need to make any form of forecast. Chris Weig.

    Other surveys like it have been showing the same kind of upswing globally.

    The reality is we are not getting the numbers we need to know what is going on.

  53. Chris Weig says:

    Mr. Pogson fails at basic statistics again.

    Said magazine according to the publisher has a readership of 108,000.

    http://www.sanomamagazines.fi/en/mediaguide/target-groups-and-magazines/magazines/tietokone-en.html

    Therefore the response rate — under the premature assumption that all respondents are indeed readers of the magazine, which is not clear either way — is about 11 percent. Much too low to draw any conclusions.

    As for Linux, let’s see, Finland fares a little bit better:

    Netmarketshare

    StatCounter

    The important thing to take away from this statistics is: Linux is flat. There’s a bit of fluctuation, but Linux has hit the ceiling. Not poised for growth.

    This is good. This is good.

  54. TM wrote, “This isn’t a national poll, it’s a READER POLL from a tech magazine! The readers are all tech enthusiasts. What are most Linux users? That’s right, enthusiasts.”

    It’s a respected magazine read by professionals. It has been around since 1996 and is much broader than just FLOSS or enthusiasts. The PDF of http://www.tietokone.fi/content/download/520591/8121782/file/tk02_2011.pdf is 41.8 MB. It’s not a lightweight. They are published by Sanoma Magazines which have 40 magazines read by 3 million Finns. There are about 5.4million Finns in the country.

    I skimmed through the magazine. It’s 70 pages mostly about a wide variety of hardware and software products. “Linux” did not appear until page 34…and only occurred 9 times. That other OS occurred more than 100 times. 12 monthly issues subscription it 91.7 €.

    A minority of GNU/Linux users are tech enthusiasts as in every grouping of humanity.

  55. oiaohm says:

    TM Repository hey cut some slack here.

    –It’s not scientific being self-selected but it is an indicator that LibreOffice and GNU/Linux are very healthy in Finland.–

    He has stated clearly that its not scientific so not 100 percent trust-able.

    TM Repository
    –Those readers are enthusiasts and don’t represent a broad-enough cross section of the general public!–
    Unfortunately in this case this is slightly wrong.

    The reader ship would include those giving advice to the general public.

    He is valid to claim its an indicator of health because if you don’t have the support people on side who are mostly the enthusiasts the product does not stand a chance no matter how good it is. Its not a indicator of market share.

    To see what is fully going on we need more scientific numbers.

    TM Repository I don’t see MS Trolls being up front that the 1 percent webstats they throw around are not scientific and possibly wildly incorrect by a large margin. As I documented they are not scientific.

    I would rip into him that is over claiming. You are reaching with the enthusiasts claim heading to bogus mostly because you don’t know enough.

    Who do people go to for computer advice TM Repository. The enthusiasts. So knowing what the enthusiasts are using is important. Microsoft, OS X and Linux enthusiasts those who are giving advice.

    TM Repository numbers don’t agree with your world view they have to be invalid. Not that there meaning is limited.

    If you go back a few years for other surveys like it MS Office was around 60 to 80 percent in the enthusiasts. This is showing quite a major drop in enthusiasts support. The enthusiasts are the word of month sales people so losing them will cost MS sales. So this number change is not healthy for Microsoft.

    This is basic demographics 1O1. Particular groups in a demographic control the actions of other groups in the demographic. Those groups that control the actions of the other groups is where you look to see what the future forecast possibly looks like. So the survey Robert is looking at is a possible long term forecast. Just like all long term forecasts can be wildly wrong.

    This case the computer enthusiasts partly controls the action of general public. This is back-able by surveys asking people why they choose particular computer products or phones.

    Government software and standards policies also controls the actions of the general public.

    The change in the enthusiasts may not show in the general statistics for 5 to 10 years. So Microsoft does not need to straight up panic but it would pay to pay attention to this shift.

    TM Repository they also found in that survey 19 percent were running Linux. So future trend looks good and healthy showing reasonable growth over the next 5 to 10 years based on this survey and others like it.

    It would be nice to have properly collected numbers to base line current market share against so we can monitor the changes and start creating correction formulas for future forecasts to make them less wild guesses.

  56. Another desperate statistic; This isn’t a national poll, it’s a READER POLL from a tech magazine! The readers are all tech enthusiasts. What are most Linux users? That’s right, enthusiasts.

    I don’t assume that “90% of Canada plays Halo 2” because I know that’s a reader poll for a gaming site, NOT a national statistic! Those readers are enthusiasts and don’t represent a broad-enough cross section of the general public!

    Besides, I’m a little surprised you’re not pushing Google Docs more, seeing as you keep claiming “all apps are moving to the web”. And you realize that LibreOffice runs on OSX and Windows, right?

Leave a Reply