How M$’s Empire is Crumbling

I am not writing about money, which is all M$ really cares about, but share of “seats”. M$ can raise its prices for years to come to keep the money rolling in from the suckers but they cannot lock in more users it seems. Consider this. In October, Wikimedia records 91% of visits were “non-mobile”, mostly that other OS. 9% were “mobile”, mostly not that other OS. In October a year ago, the numbers were 95.2% and 4.8%. M$ is losing 5% share per annum on the move to mobile alone. In 2010, M$ had 84.29% share but now have 78.38% share, down about 6% per annum. Now that Android 4 is out and Android will make a big move in tablets, the slide for M$ will be faster simply because more mobile devices will be produced.

“8” won’t help M$. By the time “8” emerges, “7” will have locked in folks as strongly as XP has and no one, not even M$’s friends will want to invest in radical change to “8”. Meanwhile OEMs and consumers cannot wait to ship Android/Linux on smart thingies and soon, notebooks. No one who needs/wants a non-desktop GUI will need “8” if they already have seen Android/Linux in action. The general has sent troops east while the opposition outflanks him to the south and takes territory unopposed.

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.

91 Responses to How M$’s Empire is Crumbling

  1. Clarence Moon says:

    All that this seems to say is that there are a billion PCs in the world (Don’t you commonly claim that there are many more than that?) and they are trying to register more than 10 million so as to “prove” that there is more than 1% Linux in use on the desktop. So far they only have 184,786 of which 6100 are in Brazil.

    Since you are postulating some 30 million per quarter, 10 million a month, are being sold, their goal should be achieved pretty quickly. But they haven’t shown up yet.

  2. Facts not in evidence.

    Most of the usage may be in schools, business and government offices where the NAT routing puts all clients behind a single IP address with a single browser, so many machines count as one. Where I worked last year we had nearly 100 machines but the firewall passed them off as one.

    percentage based on the computers registered on the site


  3. Clarence Moon says:

    And yet the usage statistics for Brazil still show a Linux usage rate of about 1%. That is no better than in the United States or Canada.

  4. Clarence Moon wrote, “The statistics are percentages, comparing the frequency that one OS or another is used within each country of origin to access various sites. They show the same general sort of ratios in Brazil as they show in the United States for Windows vs Mac vs Linux.”

    Clarence continues to ignore the fact that governments of countries are actively promoting GNU/Linux.

    e.g. From 2007 to 2010, Brazil trained 812000 teachers (half of all their school teachers) in using computers in schools. The training included use of GNU/Linux. How could anyone believe the pattern of usage of GNU/Linux is not different from USA in Brazil? Brazil also rolled out 350K (of 500K planned) GNU/Linux PCs to schools. Brazil also produces its own PCs (thanks to trade barriers…) and 20% run GNU/Linux. Brazil produces low-cost PCs subsidized by the government and 20% of them are GNU/Linux.

    The effect of years of active promotion of GNU/Linux in Brazil is shown in this blog. Some businesses are rolling out 1000 GNU/Linux thin clients per day. One company ran 600 thin clients on a RedHat server but cut the roll-out to fewer to give a margin of safety. Brazil is seriously taking up GNU/Linux.

  5. Clarence Moon says:

    The statistics are percentages, comparing the frequency that one OS or another is used within each country of origin to access various sites. They show the same general sort of ratios in Brazil as they show in the United States for Windows vs Mac vs Linux.

    You are suggesting that a huge number of non-Windows computers are being consumed in these countries and yet there are no. statistics showing increased non-Windows usage for internet accesses. What other statistics do you speak of that are different from those that I cited?

  6. Notice that our most reliable web stats are heavily weighted by English sites. Many emerging markets use other languages, like Portuguese in Brazil.

    If we look for .br sites in Portuguese, we see dramatically different stats.

  7. Clarence Moon says:

    The one thing that seems to run counter to your theory is that the developing countries where you suggest the missing inventory is being consummed are themselves even more likely to be using Windows for internet accesses than the United States. It should be the other way around if there is such a high rate of uptake of Linux computers in these parts of the world since their installed base is small to begin with and the additional Linux computers would have a more immediate effect.

    Another phenomenon that seems worth of being mentioned once more is that there are thought to be more phones and tablets lready in use today than there are PCs. Yet the statistics show less than 10% of internet accesses occur via these devices when they have 50% or more of the total installed base.

    That strongly suggests to me that they are not going to replace conventional computers soon, if ever.

  8. Kozmcrae says:

    “Believing it is true doesn’t make it true.”

    Truer words were never said. Of course believing it doesn’t make it true, being true makes it true.

    Linux is on more servers than Microsoft. That’s all you need to know. (From my clipboard, about 30 entries down.)

  9. No, the installed base is higher than that. It takes the average PC 5 or 6 years to die. That’s why XP is fading so slowly. M$ did not really start to slip until a year ago when folks began to migrate seriously to “7” and discovered the high cost of changing both hardware and software during a recession. At that time surveys showed folks were seriously examining their options. M$’s response was to promote thin clients (which they derided a decade before) as long as they could be taxed. Thin clients were just a few percent of shipments but nearly 10% of installed base. No one needs M$’s stuff installed on a thin client so M$ has a CAL for that to “extract value”. Discovering that using thin clients with GNU/Linux on both client and server was illuminating. It caused many businesses to go to OS-independent web applications to avoid all taxes. It’s still a slow process to untangle a large organization from that other OS but emerging markets are not so encumbered. That’s where the change is happening, many new installations are using GNU/Linux and FLOSS from the beginning. The change in shipping rates reflects the fact that most of the growth is in BRIC countries and Africa. That growth could be increasing. Imagine 2012 with 100 million+ tablets shipped without that other OS.

  10. Clarence Moon says:

    So where is this 40% or so of non-Windows computers going? No one else is commenting on it and it is surely rather remarkable if true.

  11. In the “good old days” for Wintel, M$ could order OEMs to do silly things that harmed the OEMs business just because M$ wanted the monopoly to be stronger. That is no longer the case and OEMs are producing products with other OS in far greater share. At one point M$ had its hands on about 95% of new PCs. Now they are probably on less than 60%. That’s crumbling. The lockin may still be strong in several sectors of IT but at the factory it’s weakened.

    That’s not a belief but based on numbers from quite reasonable measurements of the market: M$’s own pronouncements and surveys by respected global firms.

  12. oldman says:

    Believing it is true doesn’t make it true.

  13. Kozmcrae says:

    Linux is on more servers than Microsoft. That’s all you need to know.

  14. oldman says:

    “Repeating it again and again doesn’t make it true. Being true makes it true.

    I am Elmer J. Fudd Millionare, I own a mansion and a yacht…


  15. Kozmcrae says:

    “Repeat again:

    I am Elmer J. Fudd Millionare, I own a mansion and a yacht…”

    Linux is on more servers than Microsoft. That’s all you need to know.

    Repeating it again and again doesn’t make it true. Being true makes it true.

  16. oldman, you have data. Are you allowed to share it with us? How many servers do you run wth what OS and what’s their workload? How many servers are just holding the hand of that other OS on clients (AD/file/print)? How many fewer servers could you run if you used GNU/Linux throughout?

  17. oldman says:

    “Forget Red Hat Clarence. Linux is on more servers than Microsoft. That’s all you need to know.”

    Repeat again:

    I am Elmer J. Fudd Millionare, I own a mansion and a yacht…


  18. Clarence Moon says:

    I think that IDC just mentions highlights of its reports that might entice someone in an allied business to buy the report to see the details. The company that I work for is in the software business and sells products mostly to IT administrators who need our stuff. I won’t say any more detail about specifics, but I will say that our management worries a lot about profits going up or down and occasionally puts a freeze on pay raises and/or new hiring or, worse, conducts a force reduction that puts everyone on their toes.

    From living that sort of life for a long time, I know that it is only useful to focus on the financial numbers. We think we are number one in our industry, based on studies that we pay for, but when the fat is in the fire, the only thing that matters is whether or not our numbers match what was predicted. If our sales and profits go up, everyone is happy regardless of what the other guys are selling. I am sure it is the same with them.

    Numbers up means bonus, raise, and beer in the breakroom on Friday. Numbers down means watch out for a whammy. Numbers were down a few years ago and we’re still missing some troops, even though the numbers are currently up nicely and we survivors are in fat city at the moment.

  19. I have to ask why IDC mentioned the share of shipments one time only. Was it a good quarter for M$ so M$ agreed that it be public? I don’t know, but I cannot see that M$ would discount its licensing fees that much. There’s nothing about the size of the server in the licensing information I have seen. You pay ~$1000 for the privilege of using your machine and $40 per client machine, a terribly high price. M$ is double-charging.

  20. Kozmcrae says:

    Forget Red Hat Clarence. Linux is on more servers than Microsoft. That’s all you need to know.

  21. Clarence Moon says:

    I don’t think that I claimed that the price for a server with Windows OS is less than the price for the same server with Linux, Mr. Pogson. I said that a greater proportion of the servers sold with Linux are the large, expensive servers previously associated with Unix. The implication here is that more of Microsoft’s revenues come from the smaller servers, such as blades, and so their unit count percentage is even higher than their revenue percentage in terms of their share of shipping servers.

    Your IDC cite is similar to others that I have seen, for example:

    These are summaries of what IDC uses to get interested parties to buy the whole report, but the one I cited from May 2011, while similar to yours, expands the revenue share note to say:

    “Microsoft Windows server demand also continued to show strong demand as Windows-based hardware revenue increased 10.1% year-over-year. Quarterly revenue of $5.8 billion for Windows servers represented 48.5% of overall quarterly factory revenue and 75.2% of all quarterly server shipments”

    So in May, Microsoft’s half of the revenue corresponded to three quarters of the shipments. If you then assume that Linux accounts for the great bulk of the remaining units, with Unix machines being expensive but few in number, the shipment rate for Windows servers in terms of units is some 3 times the volume for Linux servers.

    Microsoft’s server revenues do not consist solely of operating system licenses, I believe. They also include their SQL Server database and other types of servers such as Exchange. Also, Red Hat’s revenues of about a billion dollars a year is obtained from selling support contracts, which include an installation of their own version of Linux, and if Linux has 25% of the total server unit volume, that billion dollars translates into a cost applied to their 2 million units, ending up at $500 average. That is a lot less than what you suggest although the number of units for Red Hat is not likely to be that high since Novell and some others are also in that business.

  22. Kozmcrae says:

    By the way Clarence, Linux is on more servers than Microsoft.

  23. Explain how a server with that other OS costs less than the same server with GNU/Linux? By my calculations, GNU/Linux costs much less. You are overweighting the count of servers by price. Look for unit shipments.

    IDC claims about 2 million servers are shipped per quarter and M$ makes so many $billions.

    “x86 server revenues continued to improve in 3Q11, growing 6.0% in the quarter to $8.7 billion worldwide as unit shipments increased 4.0% to 2.0 million servers. “

    M$ is making revenue of $4.25billion per quarter, so the least they could be making per server is ~$2000, if the whole price of the server went to M$. They get nothing like that of course so they are taking much more per server and GNU/Linux takes nothing.

    HP took in $5.36 billion on servers/storage for Q2 of 2011, but only netted $699 million. Their unit share of servers is about 30%. So, the average price of an HP server is $5.36billion/(30% of 2 million) = $8000+ No doubt HP can make servers for far less, say $4000 average so M$ is getting more than $4000 per server and they get their hands on less than half the servers. A modest server includes a tax of $1000 plus $40 per CAL so a medium to large business can easily throw that much money at M$.

  24. Clarence Moon says:

    You repeat your mantra over and again, Koz, but what does it mean? By what means are you counting “servers”? Many Linux fans don’t seem to be able to distinguish web servers, of which there are many more Linux sites than Windows sites, from the totality of servers in service for purposes other than web page serving. If you are unable to see that distinction yourself, you might want to ask a friend who might be able to tell you the story.

    Alternately, if you mean all types of servers, then you would have to explain how it is that annual unit shipments of Windows servers far outnumber annual unit shipments of Linux servers in commercial channels and how it has somehow resulted in Linux servers outnumbering Windows ones.

    I would guess that your only recourse would be to postulate a stealthy flow of uncounted units, much like Oiaohm seems to be doing for desktop machines. Perhaps both secret servers and secret desktops are winding up in the same secret place, awaiting the day when they will be unveiled and suddenly applied to the world’s computing needs.

  25. Kozmcrae says:

    Clarence, of course it’s no a popularity contest unless you are talking about the desktop, then it is a popularity contest. Linux is on more servers than Microsoft.

    You can wave your hands around all you want, but it won’t change a thing. Linux is on more servers than Microsoft.

    You can throw a bunch of words on the monitor but the fact is, Linux is on more servers than Microsoft.

    You seem to have a problem with Linux being on more servers than Microsoft.

    It just plain arithmetic, Linux is on more servers than Microsoft.

    Do you understand now that Linux is on more servers than Microsoft.

    The economy of servers is one thing, the number of servers is another. That’s simply a side effect of the open source paradigm. That’s the missing element in your logic. You refuse to admit that Linux is on more servers than Microsoft.

    By the way Clarence, Linux is on more servers than Microsoft.

  26. Clarence Moon says:

    Mr. Oiaohm, fuss with your figures as much as you wish, but until you find those elusive missing Linux computers somewhere, you are full of beans.

    Mr. Koz, PC software, including server software, is a business, not a popularity contest. Microsoft is making a lot of money there and that can be measured and plotted month to month, quarter to quarter, and year to year. They are incredibly successful and have been for a long time and will continue to be in the foreseeable future. Even as a popularity contest it would appear that far more units are sold as servers with Windows server OS installed than are sold with Linux installed, roughtly 3 times as many.

    You may think that Microsoft do not deserve their riches, and there are others who feel the same way. I myself think that they deserve what their efforts produce. I don’t work for Microsoft nor do I own any stock other than what my IRA and 401K funds might own. I never read the tomes that they occasionally send me, just the bottom line as to what they are currently worth. But I do believe in free enterprise and think that success will come to those who deserve it by providing products that people see as better than others.

    It is OK for you to posture and prance and claim that Linux is better than Windows and Oiaohm and Pogson may claim that there is a huge secret society of Linux users who somehow have evaded detection by the business media. No one is harmed by those thoughts, but then nothing in the real world is affected either.

    Red Hat is the premier provider of Linux server software to the commercial world. Period. Novell was second, but they could not make enough money to pay for their overall operations and theie assets have been sold piecemeal to various companies including Microsoft.

  27. Kozmcrae says:

    Clarence, you are kicking a dead horse. Linux is bigger in the server market than Microsoft. End of story. Does Microsoft make more money than Red Hat. How the Hell do I know. What the Hell difference does it make. Red Hat is just one company and one of the few, comparatively to everything else, who is charging for their product and services.

    You are doing it again. Comparing ONE LINUX DETAIL, in this case Red Hat, AGAINST THE WHOLE OF MICROSOFT’S SERVER !!!***SALES***!!!. If you really want to compare, then compare all the Linux servers in the World against all the Microsoft servers in the World. That would be a comparison of Linux servers against Microsoft servers. The asinine way, that would be your way, would be to single out one Linux company and compare it to Microsoft. That is not a fair comparison.

  28. oiaohm says:

    Clarence Moon who unit volume. IDC discredit one that they have deleted from there reports and site. IDC is now revenue only for OS split.

    Gardner does not do unit of Windows shipped either.

    So what firm are you using Clarence. You said IDC before ie not possible. Don’t mix unit numbers with revenue numbers. They are very different.

    w3tech is the only two that does direct unit numbers of servers that publishes for free. w3tech have it as 63.90% Unix related vs 36.10% Windows. secuirty space who only cares about Linux numbers has it as 60 percent Linux. Serous-ally with secuirty space anything non Linux goes into other BSD Unix Windows are all pooled into other that they don’t care about. Secuirty space is a Linux secuirty support company so they are interested in how much Linux is out there nothing else.

    Remember the biggest installs of Linux don’t use Redhat. 1 billion dollars is only the tip of the Linux iceberg.

    Clarence Moon
    “For one thing, there is quite a bit of industry news published regarding Linux server sales versus Windows servers and Unix Servers. Linux is currently last, incidentally, behind Windows, Unix, and IBM mainframe sales in that order.”
    Clarence Moon so the most common form of death is car crash? because its most reported in the news.

    Most reported does not make something correct. Some of the biggest lies have been most reported like the earth being flat. When you lose the means to check what you are being told you become sheep who will believe all kinds of non realities.

    Pull them up you will find repeatedly they are revenue numbers not unit numbers.

    IBM openly admits that they sell more units of Linux Mainframe servers than they sell of AIX IBM mainframes. So just by IBM Mainframes Linux should out number AIX IBM mainframes. Wait that is right. IDC the f wit counted IBM Mainframe sales as Unix yes its in there report when you read it. Even that IBM sells more IBM Mainframe suppers running Linux not AIX. There is a 30 70 split in IBM mainframes. AIX ~30 percent ~70 percent Linux. So how can Linux be behind IBM mainframes unless someone cannot do there numbers right. Linux has to be behind Linux this is logically flawed IBM numbers tell you this.

    IDC working out market share they are hopeless. Who numbers do you see PR everywhere. IDC’s because they are released for free. Gardner last free release of numbers was 2007.

    Also something you have to become very aware of Clarence Moon the Microsoft licensing more often than not is worth more than the hardware for server.

    This is why revenue counts don’t work and appear like you have to divide windows in half times Linux by at least 2 and unix prity much stays normal to get something that looks near to right. Of course having to perform that level of extrema maths makes the complete stuff junk.

    Remember I have a redhat license. My server dies. I don’t need a new license. Next Windows volume license for server is the same. Its one of the last remaining windows license that is not a upgrade on volume licensing. There should be quite a large figure in shipping of servers that goes out without a OS that we don’t know what in heck they are due to the Windows and Redhat licenses and other Linuxs.

    $16 billion include items like sharepoint and cals. does not really tell you how many units are deployed in a quarter. Please note redhat does not charge extra for something like sharepoint.

    True fact of the matter redhat charges a min 1/10 of what MS does for the same provide. 1 billion redhat could be the same number of real world units as MS 16 billion.

    Idea that MS making 16 billion is doing more business than redhat is flawed. Redhat could be doing more business at a lower profit mark up and most likely is.

    Clarence Moon you are simply a idiot who thinks revenue tells you what was going on.

    If I apply this logic back in 1995 MS did not sell much that year compared to Unix. Unix makers had revune numbers that put the sale of Windows 95 to shame. Unix did not move many units. Unit are marketshare. Revenue is not. 1995 tells you this clearly.

  29. Clarence Moon says:

    I am just stating the facts, Koz. You, on the other hand, seem to be living in a dream world. Microsoft might very well forego making server software if they were crazy, but that group produced some $16 billion dollars in sales over the past year and is increasing quarter after quarter. Red Hat, the largest Linux supplier by far, sold less that a billion dollars for the whole year. So every 3-4 weeks, Microsoft does more business than Red Hat dooes in 12 months.

    I don’t think you are very well informed.

  30. Kozmcrae says:

    Clarence, why do you bother to trying to place Windows over Linux in servers. Most Microsoft boosters have long since given up on that notion. They readily acknowledge Linux’ ascendancy in the server market. You should too. That would let you help Microsoft maintain their death grip on their last and fading monopoly, the desktop.

  31. Clarence Moon wrote, “at least in terms of hardware prices, that price is proportional to capacity and that server capacity is more important overall than unit volume?”

    No. Prices for hardware are rather arbitrary. Suppliers charge what the market will bear in relation to competitors’ prices. Most servers are idling which justifies server consolidatio by virtualization. What is a small organization going to virtualize if they have only one or two servers? The vast majority of users of IT are in smaller organizations so throughput of individual servers is not widely important. For instance, I have been in several schools who did not even have a server let alone use one that was on the LAN. Availability is the most important feature of a server. I have been in several schools running 2003 and all they did with it was manage logins… What a waste of thousands of dollars.

  32. Clarence Moon says:

    Do you not agree that, at least in terms of hardware prices, that price is proportional to capacity and that server capacity is more important overall than unit volume? Available statistics for unit volume in the server markets show that Windows have the highest unit volume overall, followed by Linux, and then Unix and mainframes in the distance.

    Virtually all the server business for Windows is on x86 platforms that tend to be small such as blades. Linux has much more “big iron” sales proportionately than Windows and essentially all Unix and mainframe sales are large machines.

    There is certainly a “stealth” server market wherein servers are constructed by adding Linux to recycled desktop machines and that does not register on the industry statistics since there is usually no money involved. That may give Linux a numerical edge in units although from a business point of view I don’t think most commercial companies really care since admins who create this kind of server are not very likely candidates for a multi-thousand dollar backup product like Symantec or a commercial database such as Oracle. They represent a sort of “unsold” market where the needs of the users are met without any correspondingly high investment.

  33. Clarence Moon wrote of servers, “Linux is currently last”.

    Sales dollars, perhaps, not units. Every GNU/Linux machine with a network connection can be a server so the category is not well defined. It is also irrelevant. Many servers running that other OS end up being installed of GNU/Linux because it is the cheapest way to repurpose them. I have made many servers out of client PCs. Many servers are bought with no OS as well. So, the measure Clarence follows is more about the high prices of that other OS on servers than anything else.

    Server software prices vary but a simple server with that other OS costs $1000+ more with that other OS and a few CALs. GNU/Linux is $free costing just for installation.

  34. Clarence Moon says:

    For one thing, there is quite a bit of industry news published regarding Linux server sales versus Windows servers and Unix Servers. Linux is currently last, incidentally, behind Windows, Unix, and IBM mainframe sales in that order.

    Keep on believing in magic, Mr. Oiaohm. It makes you look cute. I will continue to believe that the published metrics for Windows, Macintosh, and Linux desktop share based on financial returns and usage measures are correct.

    Maybe one of these days if your detective work can uncover a stash of hundreds of millions of Linux PCs you will be vindicated. Maybe they are hidden in the Austrailian deserts or the Canadian wilds. Look there.

  35. oiaohm says:

    Clarence Moon
    “You are only extrapolating figures from various sources that have no way of being audited and that are not correlated with one another, Mr. Oiaohm.”

    They all collate to financial year figures from many sources.

    Simple point go find a number for the number of desktop Units shipped for the 2010-2011 timeframe you trust.

    Basically I want you todo some leg work Clarence Moon.

    You will find it is 360~ million 364 points to a particular source. Microsoft Annual reports tell you how much Windows 7 has shipped. That gives you the 200 million for the time frame.

    Go to Apple stock report find it under 20 million.

    You have problems these numbers are audited by the stock exchange.

    Only figure that is possible off is number of desktop units shipped for 2010-2011. Windows licenses sold is dead on. Apple units shipped is dead on.

    Clarence Moon
    “In one post you say IDC overcounts unit shipments”
    They have been caught doing errors in the past.

    But IDC numbers on overall unit shipments to do align with everyone else. Over counting to trying to divide revenue to produce unit numbers because companies don’t give them unit numbers on numbers of Windows/Linux… shiped. Invalid method is to attempt to convert revenue to unit. But published units is a different matter. Everyone is using the same published unit numbers for overall total generation.

    Gartner IDC that you guys like agree on the 360~ million desktop figure. All other number collectors agree as well.

    Total unit figures are universally agreed. Since big OEM makers do give this to everyone.

    Basically its a universally agreed number Clarence Moon. I would like to see you find somewhere that says shipments were someone in the 220 to 260 million range as require for IDC OS percentages from revenue to add up.

    I have told you clearly Clarence Moon go away pull your own numbers. You want to audit mine and find fault so get self out of problem. Really I have audited mine I just want you to dig yourself into a hole with figs you have sourced yourself or prove me wrong with figures you find.

    Problem is I did not acquire my numbers from a single source. I aquired them from many and only using figures that can be confirmed by many independent sources. This is a requirement of data collection to prevent error.

    Remember no XP was sold 2010-2011. No vista was sold either. Nice clean time of no cross over of Windows versions on sale for desktop.

    Pick your time windows. This is a perfect time window to double check if IDC percentages are valid and work with reality.

    Yes 2008 number of desktop machines shipped with current shipping rate of Windows 7 does not add up either. 260 million was about 2008. This is why you are so stuffed. To keep up with market growth MS should be shipping closer to 300 Million but they are not.

    I will make it very simple Clarence Moon if my extrapolating figures are wrong you should be able to produce a workable set proving me wrong. Its put up or shut up time. Do your own figures confirm your sources and prove me wrong if you can.

    This is science method. Job is each person does there own experiment and they should agree. If they don’t the theory has to be changed.

    Basically I have made it clear from the start all I am after is a set that work properly.

    Clarence Moon
    –Only the headline would be “Shipments of Linux PCs reach 30-40 million units”.–
    We never have seen a head line like this for Linux in the server world. Why should the desktop be different. Linux desktop and Server world don’t have a focused PR department or brand image.

    Basically know what you are hunting. Linux is very stealth in nature. PR normally will not appear. MS and Apple will yell this kind of stuff in a press.

    Android is one of the rare accept-ions to this stealth nature. Remember Ubuntu only makes up less than 10 percent of Linux desktop users. There is no major PR group. And with each of there PR groups wanting to talk 30 to 40 million unit most likely appears at best to 3-4 million under one form not worth reporting.

    You are asking for something here you should not expect to exist.

    I am looking for how you normally detect Linux spread in markets. Look at the unit numbers where silent running don’t work.

  36. Kozmcrae says:

    “I did say that the first time…”

    Maybe so but you said with fewer words this time and in a way that made much more sense. If you could write like that more often we could save a lot of time around here.

  37. Clarence Moon says:

    I did say that the first time, Koz. Waldo did not seem to understand, though, so I just re-phrased it to help his thinking on the matter. No bother.

  38. Kozmcrae says:

    “Rather they are focused on convincing existing customers to re-buy Windows either as an update or as a replacement computer system. That is where the great bulk of their market lies.”

    Why didn’t you just say that in the first place? Do you enjoy hearing yourself type?

  39. Clarence Moon says:

    No, I mean something like this:

    Only the headline would be “Shipments of Linux PCs reach 30-40 million units”.

    Also, such an article would be dated much more recently than 2007. Articles would not focus on how Microsoft is going to manage a repeat of yet another record year, rather they would discuss the future of the Linux PC as a viable product market, just like they discuss Microsoft and Apple today. The articles that you cite are focused totally on Microsoft and what Microsoft might or might not do to improve their already huge revenue stream.

    I might also point out that these articles fail to mention any huge market for Linux desktop PCs impacting Microsoft’s businesses and rather discuss ebb and flow of the PC business in general suggesting that Microsoft’s fate is tied to the overall market and not on competition within for OS dollars.

  40. Clarence Moon says:

    I might ask, Waldo, just what do you think Microsoft could actually do to capture the Apple “share” of the market. Remember, now, Microsoft makes a software OS product, namely Windows, and Apple sells computers, the Macintosh line, at the only point of competition.

    To capture the Apple market share, either Microsoft would have to convince Apple to use Windows instead of OS X or Apple would have to go out of the computer business entirely. There is no strategy for Microsoft to accomplish either of those ends.

    So Microsoft’s “market” is limited as a practical matter to the OEMs who compete with Apple. Outside of a handfull of computers sold by companies that Mr. Pogson referenced somewhere else in this blog a short while ago, all of the world’s computer OEMs offer Windows essentially exclusively.

    Imagine that Microsoft is “struggling” if that brings you some comfort, but the reality is that Microsoft is not focused on capturing new customers from other OS users or even on preventing existing customers from switching to another OS. Rather they are focused on convincing existing customers to re-buy Windows either as an update or as a replacement computer system. That is where the great bulk of their market lies.

  41. Clarence Moon says:

    You are only extrapolating figures from various sources that have no way of being audited and that are not correlated with one another, Mr. Oiaohm. You do not even cite the sources for the numbers that you are using. In one post you say IDC overcounts unit shipments and in another you rely on the IDC derived numbers to make a point. Perhaps you do not see the problem with doing that.

    But ignore all that and focus on the fact that there is no one in the whole world that is writing about this “black hole” as a business opportunity. Surely the fate of 40% of the world’s PC production would warrant an article or two in the trade press.

    Of course you can suggest that there is a world-wide conspiracy crafted by Microsoft and designed to hide this flow from the public who would surely run out and buy one themselves if only there were a place other than Hong Kong where they were on the store shelves.

  42. Waldo Frankenhammer III says:

    Your first paragraph Clarence, declares Microsoft as nearly owning the PC desktop market. You brush aside any portion that Apple may own by saying, “there is nothing that Microsoft could do, short of buying Apple stock, to obtain any of Apple’s share of the overall PC market.”.

    That is pure fantasy. If Microsoft could own the desktop 100%, they would. As it is, they are struggling to keep what they’ve got in all of the markets they compete in.

    That is the reality today. That wasn’t always the case so maybe you are confusing what was in the past with what is now. Their stock has remained flat while others like Google and Apple have gained.

    Microsoft was and is a great company but they are not as great as they once were. That is a small but important difference. If people do not face up to that fact then things will get worse for Microsoft. I’m not talking about you, it doesn’t matter what you think. I’m talking about the people who make Microsoft go.

  43. oiaohm says:

    Clarence Moon at 60 million copies is still 100 million out at 70 you are still over 60 million out. The black hole is huge. 10 million here or there does not make it disappear. 20 million here or there does not make it disappear 30 million also does not make it disappear.

    This last fig is deception on wikipedia.
    “On July 12, 2011, the sales figure was refined to over 400 million end-user licenses and business installations.”
    Yes that 400 million the business installations equal company volume license installs as well and business installations are not stand alone so each one requires a windows license from somewhere.

    Yes business installations is the name MS gives to there volume licensing systems. This 400 million is total since release.

    “As of July 22, 2010, Windows 7 had sold 175 million copies” 225 for the year 2010-2011
    2010-2011 364 million desktop and laptop machines sold. And the first number 175 million does not include business volume licenses or was not meant to be. That still leaves you missing over 100 million. Remember 364 million does not include white box production. There need to be left over. Lots of left over because MS is selling to white box production.

    None of the 364 million was shipped out OS less. Every one was shipped with something even if it was freedos. The big OEM is impossible to by a machine OS less.

    If you do a graph of reported totals of windows 7 sales you find that it a straight line growth or very very close to. MS PR has been saying growing sales and this is also not backed by MS own numbers. If growth is somehow selling at the same rate yes its growth. Yes MS Spin at work.

    Clarence Moon
    “Also, these figures do not mention the revenues that Microsoft receives from its various site license schemes that it offers to big business purchasers, giving them a blanket right to install Windows on as many computers as they may wish with no accountability on a unit basis.”

    Sorry no such Licenses exist any more. Have not for over 8 years. Remember I acquire Microsoft Licenses for big companies I know exactly what licensing MS offers. Last OS with licenses like that was windows 2000 and MS bought all those back. Windows 2000 don’t work too well on modern hardware. You cannot buy a site license to clean install with. All current site licenses of Windows are upgrades the machines must have pre existing licenses. Also the volume licenses are reported in the 400 million Windows 7. This also adds serous question about business uptake of Windows 7.

    This is the problem there is no more of magical extra units somewhere.

    The black-hole is showing something criminal or something different.

    Clarence Moon
    “The IDC figures are for a later quarter yet and presumably the shipping rate for Windows 7 could have been higher than in previous quarters.”
    Grown faster than the shipping rate of Windows 7. This is why the black hole is getting larger. We can math out that quarter as well.

    Waldo Frankenhammer III mine is a error in method both IDC and Gartner. Like Linux support contracts are normally aquired independently to the hardware. So Linux installed hardware licensing is cheaper by a large margin. So don’t show up in revenue. Also the desktop revenue figures are not broken down by OS in side the Major OEMs. A desktop sale is a desktop sale.

    Its not that IDC and Gartner are bias to Microsoft its that there method is bias to the difference between how Linux and everything else does business. Unix sells the support contract with the hardware.

    Clarence Moon
    “That leaves the ISVs and other after-market types who want to understand demographics and trends and vitally need accurate information in order to allocate their development funds optimally.”
    This is why it critical for the numbers to be correctly worked out. Why you should not be able to find black holes unexplained.

    Existence of a black hole you cannot close shows a error in the numbers if IDC and Gartner for the market percentage.

    Clarence Moon I said do the figures make a set that work. Trying to find errors in mine you may problem is none of my errors are large enough to make the black hole disappear. With the errors I have 10-20 million missing per quarter with MS sales the numbers still should have closed up.

    Understanding how IDC and Gartner makers there numbers makes you very aware there numbers are not reality and should not be used to attempt to say reality of market share. Who is making the most profit is all those numbers are good for.

  44. Clarence Moon says:

    I do not think that you have the right idea here, Waldo. Microsoft has an overwhelming lead in their market and, if you really think about it, have essentially 100% of the revenues that are actually available to them. Don’t look at Apple’s share as something Microsoft does not have; there is nothing that Microsoft could do, short of buying Apple stock, to obtain any of Apple’s share of the overall PC market.

    Companies pay for market research in order to help them formulate strategies on how to effectively address target markets. Microsoft knows that it has all there is to have and is only concerned with how fast that market might grow and they are their own best source of information since they can simply ask their customers about what the future might hold. Their salesmen have been living with the situation for decades and are much better suited as well as more strategically placed for getting the unvarnished truth from the industry. They are not going to pay some outfit like IDC for those numbers.

    They are not going to pay anyone for numbers that are cooked up and intended only to flatter their egos. The people on the hot seat have to produce and failure to do so results in a quick trip out the gate to other pastures.

    Buyers do not look at these numbers either. They buy want their perceptions of need and satisfaction tell them to buy. Whether Microsoft Windows has 80% or the market or 90% of the market or 100% of the market is the same to them, namely Microsoft is the mainstream choice and only the few depart from that.

    That leaves the ISVs and other after-market types who want to understand demographics and trends and vitally need accurate information in order to allocate their development funds optimally. They are not going to put up with fantasies created to please or scare Microsoft. They want and need the truth and will not pay for any BS.

    Of course IDC, Gartner, and many other, smaller firms do not have a perfect view of the market either. They are just making an effort to discover facts and operate on them with conventional marketing wisdom to arrive at some unique conclusion that some people will be eager to pay to hear and understand how it was constructed.

  45. Waldo Frankenhammer III says:

    Let’s assume for the moment that IDC and Gartner know their business (gathering data in certain markets and producing accurate reports on that data). Microsoft is without a doubt one of their biggest customers. It is in Microsoft’s best interest for the reports coming out of both IDC and Gartner to put Microsoft in a good light. This is mostly the case but not always.

    I have wondered why, with Microsoft’s money and influence, IDC and Gartner don’t always produce glowing reports. If IDC and Gartner are as good as we suspect they should be, they could be sitting on some data that would expose Microsoft in a very bad way. After all, it would be unusual for a company as large as Microsoft to not have some skeletons in its closets. As a reminder of the data bomb they might be sitting on, IDC and Gartner will release a report once and a while that puts a little tarnish on Microsoft’s image.

    I believe they may be blackmailing Microsoft. It would not be inconceivable. We’ve seen worse in the corporate world. Far worse.

    This idea of blackmail is easy to dismiss. I know it’s a far fetched notion but it does fit the facts as we see them.

  46. Clarence Moon says:

    It is not very productive to exhaustively analyze this situation, I think, since it is plain that no such stealth market exists for non-Windows/Mac computers, regardless of what numbers you present. However, I did take a cursory look at Wikipedia article statements at:

    Here it shows the 2011 sales claimed by Microsoft to be over 60 million copies per quarter at the middle of the year with an accelerating rate. Also, these figures do not mention the revenues that Microsoft receives from its various site license schemes that it offers to big business purchasers, giving them a blanket right to install Windows on as many computers as they may wish with no accountability on a unit basis. Finally there is the piracy factor that you mentioned previously.

    The IDC figures are for a later quarter yet and presumably the shipping rate for Windows 7 could have been higher than in previous quarters. Taken overall and coupled with the notion that DC and Gartner over-count the unit volume (a theory that is used to dismiss the market share of Windows servers by Mr. Oiaohm), the totals seem to be roughly in line.

    If there were a genuine difference that is accounted for by sales of Linux PCs, then there should be some evidence somewhere of the strength and health of that business segment discussed in the PC business media. What has been considered so far are non-synchronized press releases or overviews that are unaudited and conflict with one another. When the business press talks of vast numbers of unit sales of Linux computers and starts predicting future volume, you will have a real market share. Until then all you have is a mistaken calculation.

  47. oiaohm says:

    Yes I forgot that IDC and Gartner when they unit numbers don’t match up to MS unit numbers either.

    Basically we have a black hole we cannot explain away. Why don’t the unit numbers add up. They should.

  48. oiaohm says:

    Clarence Moon

    There is a problem. “You, on the other hand, pull numbers out of thin air and complain that they do not properly tally.”

    I am not pulling these numbers out of thin air. the 200 million a year are MS reported numbers for the year of desktop license sold rounded up. 364 are the agree number of desktop and laptop units produce by large OEM rounded down before being totaled. Yearly figures in Units.

    And the 20 million apple is apples own agreed number rounded up for the same year.

    Only thing I have done to the numbers is performed some basic maths on them to get to 39 percent not accounted for.

    The numbers don’t add up that is so simple.

    “These companies have been around for decades selling their marketing reports to companies whose businesses depend on having accurtate market informtion and who a) stay in business and b) continue to commission the reports.”
    Does not matter how many years they have been around selling reports.

    Both IDC and Gartner over the decades producing reports have been caught doing wild guesses. Yes they both got caught doing units by dividing revenue by made up number even that the IBM reports had units for that section. Ie the IBM numbers in the IDC and Gartner was not matching the real IBM numbers for the same timeframe. After that I started building my own numbers.

    The is the problem businesses are depending on IDC and Gartner who numbers have been repeatedly proven questionable.

    HP, IBM, Dell don’t report Ubuntu/Linux/Freedos machines sold on desktops independent to Windows machines. That those companies don’t report this then IDC and Gartner have to guess. They do report total units for the year also they don’t split these either. Its we sold X number of desktops not what were they. HP, IBM, Dell do report a server split. Yes it quite straight forward when you know for sure Dell has sold Ubuntu laptops and Desktops yet there is nothing in their reports showing the fact. So error will be in IDC and Gartner numbers because of this.

    IDC and Gartner are based off stock exchange reports to a large percentage. Basically I am using a difference section of the same source documents they use. A section they should have been calculating to prevent the possibility of error.

    Clarence Moon
    “The obvious solution to the mystery is that your numbers contain cumultive errors, either overcounts of production or undercounts of OS use or both.”
    Undercounts of OS is not possible unless MS reported figures for sold OS’s is wrong. The 200 is directly from MS reported sales figures of MS Windows.

    Overcounts of production are also not possible for 364. This does not include unsold items in storage. If error is any way is under by large margin. It is important to remember 364 is a low number only counting big OEMs production only this is not all OEM globally. I tried my best to make the numbers work basically Clarence Moon. 364 is a known undercount by a huge margin.

    So by these numbers sources I should have windows licenses coming out my ears. I should have left over that don’t own to any machine. A black hole should be impossible even if some OS holds a 15 to 20 percent share that we don’t know about. Then numbers should close up to no gap at a min.

    Every time I do these numbers I am trying to make the work. I though when I was having 12 coming up I had solved it closer. I was highly annoyed when I found out it was a redo maths error. When I was redoing the maths I was thinking something was wrong. When I got home and compared to my notes I was like O my god. Yes I required the numbers for my first post. Other than the bad subtract the total number of machines of the year matched all my other prior runs.

    I have not done the maths once. I have had other people do the maths as well. We are all getting the same result over and over again.

    Unit maths and IDC and Gartner should be close unless miss counting by IDC and Gartner is going on.

    You will see I know where the miss count IDC and Gartner is doing. OEM’s have not split there Linux Desktop sales away from the Desktop windows sales. Also OEM most likely have not split any freebies.

    Of course I might have a error. The question is: Are there any versions of Windows 7 that MS does not count? This means Microsoft is under reporting to share holders number of Units shipped. This is the only possible cause of undercount on MS Windows licenses large enough to explain the problem.

    I do have a suspect for part of error if there are more Windows machines. Maybe Windows 7 Starter is not counted? Again this would be false reporting to share holders by Microsoft yes a crime.

    There is no way that its a over count of production my methods was very careful to keep that major-ally under-counted. I should be doing corrections guesses to remove this undercount what I have not done. Heck it does not even proper work if I subtract 100 million from the 364 million and just use 264 million computer sold for the year. No matter what subtracting 100 million should make it add up.

    Basically there is a error showing in the Unit numbers Clarence Moon. I know there is a error I want an answer what it is. The error is that large that there is no way to explain it other than something we cannot see and are not told about.

    There is also a error that is direct up shown in the revenue numbers as well by the way OEM split them. So you can say that IDC and Gartner market share numbers of desktop are guess work at best. At worst pure deception.

    Clarence Moon IDC and Gartner numbers for long time have been quested if suspect by method.

    IDC and Gartner numbers do tell you one thing correctly what OEM is making the most profits.

    Clarence Moon please do the numbers. I am wanting this solved. If you can somehow pull the unit numbers and some how make a picture that works that when I check it is right I will be a very happy person.

  49. Explain how M$’s figures of 50 million licenses of “7” sold per quarter square up with IDC’s numbers of 90 million PCs sold per quarter. IDC counts desktops, netbooks and notebooks only.
    PCs include Desktops, Portables, Mini Notebooks and do not include handhelds, x86 Servers and Media Tablets (i.e. iPad and Android-based Tablets). Data for all vendors are reported for calendar periods.

    see IDC

    If anyone is pullling numbers out of the air it is M$ or IDC, not anyone here.

  50. Clarence Moon says:

    The answer is simple, although you will not be pleased with it. Your figures, taken from various sources, suggest that some hundred million or more PCs are somehow missed and therefore they must represent a huge gain for Linux. Somehow this is overlooked by the major sources of PC market statistics.

    These companies have been around for decades selling their marketing reports to companies whose businesses depend on having accurtate market informtion and who a) stay in business and b) continue to commission the reports.

    You, on the other hand, pull numbers out of thin air and complain that they do not properly tally.

    The obvious solution to the mystery is that your numbers contain cumultive errors, either overcounts of production or undercounts of OS use or both.

    If any commerce existed at the scale you suggest, it would be noted far and wide, yet there is no such news. The world is not conspiring to denigrate Linux. Your figures are in error.

  51. oiaohm says:

    The case here is that the missing 39 percent are from big OEMs. So the machines shipped with something they don’t ship blanks. It either was Linux or Freedos or maybe MS is giving OEMS a huge number of free copies of windows they don’t need to pay license on. But it has to be something.

    39 percent Freedos would kinda be insane. Then a percentage would have been converted to Windows just like a percentage of Windows machines are converted to Linux.

    The illegal bit does not explain what they shipped with. This is the black box of the IT world. If it never cracked open one day we could be looking at PC sales going we know X number of computers sell each year but we have no clue what is on 90 percent of them. Great your a software developer and you no longer have a clue what is going on.

  52. I think Ballmer claims illegal copies account for 15%, perhaps half the missing, but it’s still a mystery.

  53. oiaohm says:

    Basically I would like someone to explain what the 144 million unknown is I find when I compare shipped unit numbers to Windows license sales.

    Other thing I forgot to mention is yearly that black hole as been growing. At current growth rates of the blackhole. 50 to 60 percent of the known unit numbers produced a year of desktop and laptops will have a unknown OS on it by 2014.

    At some point someone has to go into that blackhole in numbers and find what is really in there because it will be the dominate factor in computer world. Anyone producing software should want to know what the 39 percent is.

    IDC and Gartner numbers are not showing this coming. Purely because they are revenue extracted numbers. All I can presume is that they have mixed unknown revenue numbers in with Microsoft revenue numbers somewhere. Or we have piracy on a scale that I cannot dream of. BIG OEM’s producing machines and not paying for them. That I cannot dream to be the case if it is there will be a very interesting anti-trust case at a min.

  54. Kozmcrae says:

    Your observations about what I’m “on” about Clarence, are mostly correct.

    “Knowing how to take a walk in a jungle alone is not a sign of success.”

    It is if you make it out alive. You have no idea what success truly is. “…having a safe and secure infrastructure is the real sign of success.”

    Do you mean, like, slash and burn the jungle? Clear it out and plant a mega mall in its place? Is that your idea of “success”?

    “Zeno of Elea obviously failed high school calculus.”

    Who knows, but he was one of the giants whose shoulders Newton and Leibniz stood upon to create calculus.

    “It was suggested by others that Linux has a 40% or higher unit volume “share” of the earth’s PC production.”

    If you say so, I’ll believe you.

    The rest of the babble? Of course it makes no sense to you. I basically said my words would pass through your brain like neutrinos pass through the Earth.

    You select a detail about GNU/Linux and weight it against and entire proprietary ecosystem.

  55. oiaohm says:

    Opps I maths errored

    It not 44 million not accounted for. is 144 million not accounted for. Yes I can get my maths wrong. 39 percent directly missing not 12 percent.

    Basically my numbers are way under where the reality are.

  56. oiaohm says:

    Clarence Moon. My position is that IDC and Gartner number cannot be trusted they fail second source check what is unit numbers directly from the makers. Yes they exist but extremely vague on what is going on.

    Both contain web number both set of web number do not match to unit counts from OEM’s.

    The unit counts are extractable from the big OEM reports to the stock exchanges around the world to get some clue what is going on the desktop.

    This shows a clear problem with method IDC and Gartner use. IBM and HP both have called IDC and Gartner on the Unix vs Linux numbers for the simple fact the both make more revenue from Unix but ship way less units. Yet IDC in particular had IBM shipping more units in Unix than Linux. Ie the divide error IDC divided Linux and Unix revenue numbers by the same figure opps IBM stock report contained the exact units of each shipped and they did not match. This kind of stuff should not be going on any company caught doing this no longer is a trust-able source.

    Unit numbers for web servers are very well know will very good breakdowns.

    Clarence Moon I would love some correctly collected data. Problem is it don’t exist for small business or desktop in detail. So I have to generate it from not approx sources not bad approx sources but sources that cannot give me all the details I would like.

    We do have some unit numbers of PC shipped for the desktop and laptop market. We do have unit numbers of copies of MS windows sold and Unit numbers of number Apple desktop machines sold. There is a gap of something. Note IDC and Gartner don’t use the unit numbers provided by Microsoft and Apple on copies of OS they have shipped.

    The fact we know MS ships about 200 million copies of windows per year they report this to there share holders. Apple ships about 20 million copies per year of desktop also report this to there share holders.

    Total new machines for 2011 364 million units not count all those made by white box production. Ie person buys case motherboard power supply and assembled themselves or small companies who don’t report back the numbers of machines they make.

    So you have a 44 million machine black hole in the desktop market at a min with no reported OS. ie 364 known production – number of copies of Windows – number of copies of apple= 44 Million.

    Remember not all the 200 million copies of Windows are in fact installed on new machines.

    44/364 What is at a min 12 percent not accounted for they have to be running something people from the big OEM cannot buy a computer without a OS provide even if its crappy freedos.

    364 million units is the big OEM production for 2010-2011. When you take into account 364 millions units don’t include all the white box produced machines. You are talk at least another 50 to 150 million in white box production. I will go low. 50 so unaccounted grows to 90million/414 or at least 21 percent unaccounted. When you remove about the 1/3 from windows what are basically company upgrades of old hardware either old XP machines or Vista machines to business versions of 7 so that is up another 60 to unaccounted. 150/414 or 36 percent unaccounted we know about and this could still be low.

    So somewhere between 12 percent to 36 percent+ allowing for best case outcomes on new desktop and laptop machines we don’t know what is on them. Again that don’t show in the IDC numbers.

    Worst case we are talking 60 percent of the new desktop market sales unaccounted for.

    The unit numbers start telling very interesting stories. Problem is it only telling a part story.

    Yes its 12 percent+ unknown. We 100 percent know this. Its not lower than 12 percent.

    Of course Phenom will try to say piracy. Problem again we have no number. Also piracy is not common in major makers where the number of PC made comes from. White box pirated is way more common. So yes there is still 12 percent that give people like Phenom a very big problem because it don’t match up with what is going on that is all piracy. 1 percent Linux would be a very reasonable presume out of the missing 12 percent. Microsoft own says own reports more likely same out of the missing 12 is more like 5 to 6 percent Linux.

    IDC and Gartner do not include the black hole unit numbers clearly shows existing. This is a problem IDC and Gartner numbers are not matching to unit numbers. So yes I know they are bogus. Problem I don’t have anything that 100 percent work to replace them with.

    Basically I have done what you should do Clarence Moon compare what you are being told to many sources.

    IDC and Gartner are not telling the truth on sales figures we know this. If they were they would be something explain the black hole. What was the 12 percent there report do not answer.

    Really I prefer to work from Unit numbers they are more correct count to what is going on than revenue. You should be able to clearly see even with my best efforts I still don’t have clue what the 100 percent true story of what is going on. I have clues but that is really not much of a help.

    Simple point unit numbers are not imagination Clarence Moon failure of alignment with them shows big problems in collection methods with IDC and Gartner.

    Now I am going to bring some real reality to this.

    China 18.5 million new PC a year
    USA 17.7 million new PC a year.
    Remember 364 million are what ship per year. Where is that 300+ million going. Basically USA and China are big but compared to the market still a drop in the ocean.

    Also here is the shocker. MS Windows sales 200 million divided Big oems sales ie 364 million equals only a 55 percent market share in new machines. Remember that is high because 200 include upgrades.

    With apple at 6 percent market share. Basically unit numbers look nothing like the IDC or Gartner numbers. Microsoft might in fact be less than 50 percent of all sales of new machines at current time.

    This is why 30 percent Linux is possible unit numbers tell us that is possible. Problem is we cannot prove it and no one can disprove it because no data exists.

    Software makers don’t care about pirates why they will not buy there software anyhow.

    So yes China and USA are less than 5 percent each of number globally. And when someone accounts for less than 5 percent of the market it can see real warping away from reality.

    Also none of my number take into account duel boots or Linux in bios or any of those other ways Linux could be there.

  57. Pingdom monitors what sites? Oh, a bunch of English-speaking partners of M$ who pay $40/month to be counted? How is that representative of OS-usage?

  58. Clarence Moon says:

    So your position is that both IDC and Gartner are totally mistaken in regard to sales mixture of Linux, Windows, and Macintosh? I can see where YOU might believe that. It is consistent with the rest of your imagination.

    I don’t believe that 1% of sales are Linux on the desktop. I believe that the figure is very close to zero. Linux PCs start life as Windows PCs and someone converts them to Linux.

  59. oiaohm says:

    Clarence Moon and Dr Loser. IDC and Gartner do webserver numbers as well. The problem is due to the fact they are revenue that don’t match and the look the same. Ie 60 windows 30 Unix stuff. We now this is backwards. So we know that IDC and Gartner numbers and other revenue numbers are 100 incorrect for what the real market shares are.

    Little bit of homework you find out that you have no usable numbers for the total server market. Remember I said IDC overlaped. I should have been more clear that the IDC report contains a web server section and its out of alignment W3Techs so proving it bogus status.

    Gartner and IDC both got caught using generated Unit numbers not real as well. ie take revenue and divide by some magical fight to get units and hope to get right number. Of course this came clear when you compared there web numbers to real unit surveys they were way off.

    Rule of statistic if you can find one major error the numbers are worthless.

    This is the problem Dr Loser and Clarence Moon both the numbers you are trying to use are known bogus.

    Yep subscription to IDC or Gartner both tell you nothing anywhere near correct about market share in any segment.

    Really you need to read this carefully.

    Last unit servey on anything close to desktop was net-books in 2009. That had Linux running at 30 percent market share.

    Tablet computers also funny enough MS is not existing.

    Dr Loser
    “I mean, really. Believe it or not, I am personally interested (and in fact unbiased) in the numbers for OS servers — I make my money out of programming the damn things.”

    Same here does not help that people keep on quoting bogus numbers like from IDC or Gartner. We need more proper Unit counts. Only numbers you can validly use are Unit counts what for small business we don’t have a clue.

    But the known unit counts give you some idea at particular times how many units were injected into the market at that time. Android and ipads are a particularly good metrics. We know exactly how many units in phones and tables were injected into the market place and when. This gives something to kinda calibrate web numbers against. Of course this still is a major-ally rough approx of what is going on.

    From Linux movements in webnumbers a 30 percent of new desktop sales is possible. Problem is due to the approx error factor it could be between 5 percent and 65 percent. That is the true best you can do.

    This is my problem there are no solid numbers at all that cover the markets of most interest. Small business servers and desktops. Sad part is no company is doing proper numbers in either market.

    Yes Kozmcrae Zeno of Elea is 100 percent correct example. We look at web numbers we have to remember Zeno of Elea example. Small slip backwards in fact require a huge number of units to be moving.

    Lets just say Clarence Moon is true about 1 percent of sales being Linux. Result would be zero in fact negative movement on Linux. Why because Linux is in fact higher than 1 percent.

    Look at both set of numbers more closely from 2010 to 2011 Linux desktops and android both grew.

    Due to the Zeno of Elea effect of head start the growth in fact require quite large numbers entering the market basically you have to displace the sales over the last 10 years at least. Also note windows 98 machines are still turning up. So 10 years of displacement is kinda optimistic. Yes the greeks had worked out most of the rules of market share statics and how to understand what you were seeing.

    Most modern people are fairly much idiots at seeing the true pictures in the numbers.

    Zeno of Elea is why webnumbers on anything other than webbrowsers is next to worthless. Reason people don’t have to go out and buy hardware for webbrowsers or have any major skill to install web browers. Hardware acts as a huge movement slower. Failure to take account for this give you as false impression what is going on.

    Also sales studies have repeatedly shown the golden point to sell person software for a computer is the first 3 years of owner ship. After that the person is not as likely to buy software.

    So anyone producing software the number of units sold each year of a particular OS is what they really need. This tells them how big the most viable market is. Webserveys are basically worthless for people thinking about producing software.

    Tablet and Mobile phone markets do answer what software makers need to know.

    All the top 20 you are point to Clarence Moon show very large change really over the last 12 months. This is the problem you need two points in time not one to try to build what is going on with new sales in that area. It the change that is the value of interest from web number. But the change figure from web numbers give you a very foggy view of reality.

  60. Clarence Moon says:

    It is hard to tell just what you are on about, Koz. At the risk of encouraging you to feed your ego any more of your baloney, I would make the following observations regarding your post above.

    Knowing how to take a walk in a jungle alone is not a sign of success. Rather it is a sign of desperation and not having to ever take that walk due to having a safe and secure infrastructure is the real sign of success.

    Zeno of Elea obviously failed high school calculus.

    It was suggested by others that Linux has a 40% or higher unit volume “share” of the earth’s PC production. I have merely pointed to the few public references available for metrics related to that claim that suggest the idea is preposterous.

    The rest of your babble makes no sense at all and that is all that I can say about it.

  61. Kozmcrae says:

    Clarence, I have accused no one of being a mercenary, quite the opposite. Read my post a little more slowly, again. It is not a “fact” that rich countries have any more wisdom than developing countries. Take a walk in any jungle alone and see how long you survive. You have no wisdom there. “…a resounding “No”.” You are taking the whole of the Linux experience as a detail in a speck of time. You are talking like an idiot. Are you an idiot?

    The proof you are suggesting is reductio ad absurdum. Zeno of Elea (circa 450 b.c.) proved that a Tortoise could beat Achilles in a race if the Tortoise was given a head start. The premise is that by the time Achilles had covered half the distance between him and the Tortoise, the Tortoise would have traveled some more distance. Each time Achilles covered half the distance, the Tortoise would have covered a little more and so, Achilles would never catch up.

    By narrowing your focus on Linux in a narrow space and time you ignore the incredible distance it has already covered. You are inferring that Linux will never catch up to what the proprietary interests have accomplished.

    If I could give you some facts that would prove with out any doubt that Linux has handily beaten some, you would not change your mind about where Linux and the proprietary interests stand. If I didn’t give you any facts but pointed out the fallacy of your logic, you still wouldn’t change your mind. If I insulted you, that wouldn’t work either. So what the Hell are you doing here? …?

    I’ve already been down the road you’re on. It’s a dead end. When you finally reach the end, you’ll have to dig your own self out. Good luck with that.

    By the way, if you choose to respond to this comment, please read it carefully so I don’t have to correct you on your reading skills.

  62. Dr Loser says:

    Ask Amazon? What on earth for?

    Since the original proposition was that Web Servers (70% Linux) are a small proportion of Business Servers (God alone knows what. As a matter of fact, I have no idea why those figures are so hard to come by unless you subscribe to Gartner et al) …

    … the fact that one large-scale Web entity (surprisingly enough, not from the BRICs) uses Linux for all its Web Servers is not germane.

    I mean, really. Believe it or not, I am personally interested (and in fact unbiased) in the numbers for OS servers — I make my money out of programming the damn things.

    But this flailing around with random statistics isn’t helping much.

  63. Clarence Moone wrote, “Linux use in the BRIC region mimics the 1% found in the rest of world.”

    The world is not a puppet of the USA. Many countries diverge dramatically from the USA in use of IT. GNU/Linux has spectacular growth in many parts of the world.

    Ask Amazon whether or not web servers are critical components of the business.

  64. Clarence Moon says:

    Koz, you confuse recognition of facts with arrogance. That is just as inaccurate as confusing it with cultism, worship, and being mercenary. You have quite the repertoire!

    As to wealth and trendsetting in BRIC, it is not an issue as to whether or not OEMs will become involved in those countries, most already have. Rather it is a question of whether they will attempt to sell Linux on a large scale and the answer, based on available metrics, is a resounding “No”. Usage of Windows and Macintosh in these countries seems to closely match such usage in the United States and Western Europe. It is only in the very poorest nations that circumstances seem to cause people to adopt Linux in excess of the world averages and then only by a small amount where they have 2% or better versus the less than 1% norm.

  65. Kozmcrae says:

    “And, again, why would a common Chineese install Linux, when Windows costs 0?”

    How the Hell would you know Phenom, you sure as Hell don’t have any common sense or humility for that matter.

    Oh, I forgot, you’re a member of the Master Race. One of the chosen few to be born in a rich country. Well, why don’t you tell those common Chinese what they can and can’t do. I’m sure they would greatly appreciate it, especially coming from a Westerner…. Wait, I think you just did.

  66. Kozmcrae says:

    “Poor countries cannot set examples to the rich ones, unless negative ones.”

    OUCH! Someone from a rich country just set a negative example of arrogance and ignorance.

    Phenom, you just made first position on my shit list, permanently.

    Hey Microsoft, these idiots are going to kill any good will you are trying to establish in the non-industrialized poor countries. Pay them to shut up or pay for the damage they cause.

  67. Phenom says:

    Now, you know what PPP is, glad to hear. It is also so kind of you to quote the source I did for you a few posts ago. 🙂

    Anyway, compare the current values with these of EU (esp. old members) and USA, and make the conclusions. Esp. in Russia, the current value and growth rate are based on the high prices of gas and oil, not on a working econonomy. These countries are 3-12 times poorer than USA.

    GDP growth can set trages only when GRP PPP starts to catch up. This is definitely not in the foreseeable future, esp. considering the hude problems China seem to be facing.

    And, again, why would a common Chineese install Linux, when Windows costs 0?

  68. Phenom wrote, “The notion that BRIC would be trendsetter is ridiculous indeed.”

    Hmmm… Huge rates of growth of GDP. Huge populations. That sets trends. OEMs will go where the money is. These countries are not poor.

    GDP per capita:
    Brazil – $10800 GDP growth rate 7.5%
    Russia – $15900 GDP growth rate 4%
    India – $3500 GDP growth rate 10.4%
    China – $7600 GDP growth rate 10.3%

    While India may have lots of poor people they add up to a powerhouse, fifth largest GDP in the world. see CIA Factbook

  69. Phenom says:

    The notion that BRIC would be trendsetter is ridiculous indeed. Poor countries cannot set examples to the rich ones, unless negative ones. That is just the way a successful person would treat a loser.

    What I fail to understand is why certain people in this forum, lead by Pogs, think that BRIC would adopt Linux to start with. In these countries software piracy is blooming, and Windows costs 0, Office costs 0, computer games cost 0. Why whould someone bother with Linux, and lose access to MS Office rich features and integration, lose access to games, when one pays nothing anyway? In BRIC, the price advantage of Linux is not there.

    Posgon a few times asserted that BRIC governments would somehow force linux adoption into the administration. Lets assume that is true. Then, however, administration is a fraction of the total computer users. There is no way that would lead to massive linux adoption. It won’t serve as an example of successful e-governments, either. All the administration of BRIC is so corrupt, that any American or European would only frown in disgust. No good example to follow here, too.

    P.S. High time you leave Russia out of B(R)IC. BIC have at least secondary industry. Russia has none of it.

  70. Clarence Moon says:

    I guess you do not bother to read what you cite, Mr. Oiaohm, or else you are dumbfounded by the words and cannot understand their meaning. The server market is very much larger than the web server market for which you cite a statistic. The conventional belief is that Linux accounts for 70% of the web servers, due to its low price for this relatively non-critical function, so your 64% is surprisingly low for this factor.

    As to unit volume for servers, it is fair to point out that almost all of the Microsoft sales are for the x86 blades and other small units whereas Linux is often used on much larger and more expensive servers. Hence a revenue advantage for Windows servers translates into a much larger share of volume, percentage-wise.

    If you think about it for just a short time and abandon your obvious prejudices, you can only conclude that your postulating a sort of stealth market of huge proportions for non-Windows/Macintosh desktops, some 40% as claimed here, is a pure fantasy and that there is no support for your thesis even among sites highly favorable to the Linux cause.

    You may be very competent when it comes to twisting the knobs and pushing the buttons while administering computer installtions on a large scale, but you are a fool when it comes to understanding and arguing in regard to the financial and legal surroundings of the PC business.

    For third area where you seem to lack any expertise, I would suggest that focusing on the tastes of backward countries or those newly emerging on the technical scene is not particularly beneficial either. BRIC countries are followers of trends, not trendsetters. What they might do today is not likely to be followed by the rest of the world and they will soon fall into line with the rest of the world. From the statistics that I cited regarding Linux usage in these and other world regions, it would seem that they have already done so and Linux use in the BRIC region mimics the 1% found in the rest of world.

  71. oiaohm says:

    Clarence Moon really you need to check your numbers. Revenue numbers has MS large than Linux in servers. Units shipped every one of those has Linux as larger than MS. Yes by all unit shipped numbers compare Linux leads in servers.

    Yes the price bias is quite major.

    Basically IDC numbers don’t show true picture Clarence Moon. W3Techs is units. Note that W3Techs and IDC number collection overlap. So IDC and W3Techs should be in alignment if there was not a price of OS bias factor.

    Ie 63.90 in units somehow does not correlate to a 38.7 percent in revenue. So shows that revenue is not showing true market share.

    Basically we have the numbers to prove that IDC numbers are basically wrong for number of machines on ground.

    Your reports about Linux being 1/3 the size of Microsoft are wrong we have the numbers that prove it. Anything doing units shows it the other way over. Linux is the dominate beast in the server world. MS is over priced.

    The big point I was pointing out is current unit sales don’t show up quickly. Also you have to now see that there is a major cost difference between the two solutions.

    So large in fact that revenue numbers are basically worthless.

    Simple fact here you have attended one too many Microsoft conferences and be brain destroyed by there PR. Clarence Moon

    Even with Unix. Unix makes more revenue than Linux even that its shipping less units and has less market share in reality.

    The issue I described is fact. That the market share of new devices selling is not displaying full in wikimedia or anything else. And the amount of movement in wikimedia does support the possibility of 30 percent of all sales being Linux. Again we really don’t have the numbers because too many parties just use IDC numbers and not go after getting unit numbers instead of revenue.

    Displacing the current deployed systems takes a while. Cuba Linux growth is about 2 percent a year. Venezuela is about 3 percent per year. Venezuela also pulled of a 24 hour migration from Windows to Linux in all government departments. Yes staff went home next day all there systems were Linux. One of the most impressive Linux migrations to date. Of course MS growth is negative in those areas.

    BRIC is harder because nightmare of nightmare they don’t have good internet systems for working out who is who. So collecting numbers there is hard going.

  72. Brazil is producing for domestic consumption about 20% GNU/Linux PCs. Dell, alone, is shipping millions annually. They have 200 stores pushing Ubuntu PCs in China.

    The argument that the world must be pushing 30% GNU/Linux because 0% are shipping in USA is bogus. There are several businesses whose major operation is producing GNU/Linux PCs.

    Lenovo will sell GNU/Linux PCs to anyone in quantity.

    Web stats are quite unreliable. Wikimedia stats, which I think are OS-neutral clearly show several times higher share than NetApplications, for instance. They can’t both be right and I know where Wikimedia gets its numbers.

  73. Clarence Moon says:

    You are nothing if not a die-hard, Mr. Oiaohm! One has to smile at your dedication and is reminded of Monty Python’s Black Knight sketch. But your explanation is clumsy and full of obvious holes, the first of which would be that, if 45% or so of the world’s production of desktops with Linux installed were simply misclassed as “servers” as you suggest, then Linux would surely lead in the server market rather than being, by all reports, about one-third the size of the Windows server market.

    With Cuba’s 6 percent as the world leader, the notion that BRIC is an imposing consumer of Linux desktops falls flat and that is not the local party line I am afraid.

    Even at that your reply is embarassingly shy of any factual citations that might lend credence to your claims. No cigar for you, I am afraid, and you are not even getting close.

  74. oiaohm says:

    Clarence Moon There is the fun one Linux Workstations sold are not counted in desktop OS revenue numbers. So there is not zero percent in the USA.

    They are counted as servers. Yes what the. Because Linux is a server OS. Now a Windows 7 machine selling on the same hardware is counted as Desktop machine.

    So low Linux desktop revenue numbers are partly this what the.

    From Top 20 countries that run Linux the top two Cuba and Venezuela both governments run exclusively Linux no windows machines at all. They also have a internet to home policies. USA has a greater percentage of the population with computers without a internet connection.

    When you look to BRIC countries they don’t have what Cuba and Venezuala has.

    Also you have to remember the in-place market share. Clarence Moon 30 pecent of all new machines being Linux still will take many years to grow by 1 percent. This can be clearly shown by how fast ipad and android is growning in the market. Remember combined they are selling more units than PC’s.

    MS is very well dug in. The PC turn over rate is less than 10 percent per year. So new sales barely show. So yes 30 percent in new sales at best 3 percent.

    Most of windows 7 growth is in upgrades. Companies with vista upgrading to 7 and other upgrades.

    Even when a company has upgraded there machines to 7 by replacing the machines they normally sell off the machines so they remain in circulation. This here basically halves the number again to 1.5 percent. Once you addin android and ipads you half that again. Get to a nasty max growth rate is over all market share of .75 per year from a 30 percent new PC hardware sales.

    Am I expecting Linux market share to grow fast no I am not. Even if Linux had 100 percent of all new devices and PC for a year the numbers would barely move at wikipedia.

    Clarence Moon you don’t understand statistics.

    To be correct the biggest question the wikimedia traffic numbers.

    Windows 2010 3,552,391 2011 3,272,901 . This is not growth. 200,000,000+ visits from windows machines disappeared in the cycle compared to the last cycle.

    Next months cycle will be interesting to see if this is a on going trend.

  75. Waldo Frankenhammer III says:

    If some Microsoft boosters from a decade ago were to fall asleep and woke up just yesterday, they would have been shocked right back into a coma at the status of GNU/Linux and Microsoft.

    Each hurdle GNU/Linux overcomes, each barrier it crashes through, the Microsoft boosters just take in detail as a minor victory soon to be overshadowed by some loss. In some cases, as with Android for one, Linux is not even considered as part of the success.

    And so it is, year after year GNU/Linux marches on. Yet to the Microsoft boosters it’s still back where it was a decade ago. Not so with the Rip Van Winkles.

  76. Clarence Moon says:

    I concede that no one in my local orbit, save my one brother, has recently been to Brazil, Russia, India, or China. And my brother didn’t seem to pay much attention to this possibility or at least has never mentioned it.

    But, other than your own calculations, can you find a credible source that makes a distinct claim that Linux has captured some 30% or more of PC shipments? If they are essentially 0% in such huge markets as the United States, then they must be much larger than even 30% in India or China or such. That is a fairly significant fact that must be echoed in some more formal setting.

    Some figures available from conventional sources do not suggest any such thing. For example:

    BRIC markets are not even listed in the forefront of Linux users by Linux sources, namely:

    It seems to me that it is unlikely that everyone seems to have missed such a revolution.

  77. Clarence Moon wrote, “Surely, there would be some evidence of that in retail stores, public media, or chatter among one’s friends, family, and acquaintenances.”

    Surely, your acquaintances do not report conditions to you in Brazil, Russia, India or China, all countries where the government promotes FLOSS and uses it and requires suppliers to supply FLOSS. There there are shelves full of FLOSS. Here’s a picture of a Dell outlet in China. GNU/Linux is on millions of PCs sold annually around the world and growth rate is huge compared to M$’s flat growth.

    There is a high rate of illegal copies of that other OS in some regions, too, but globally it is only about 15%, that still leaves 30% of new PCs unsullied by M$. 5% of PCs get MacOS and there is some illegal copying, too, but a lot are getting GNU/Linux.

  78. Clarence Moon says:

    It seems preposterous to suggest that 45% of the PCs being sold are non-Windows versions. Surely, there would be some evidence of that in retail stores, public media, or chatter among one’s friends, family, and acquaintenances. If you go into Target, Best Buy, Office Depot, or similar, there are nothing but Windows computers and the occasional Mac. That is hardly 55% in my opinion. Whatever calculations you are using to arrive at such a figure are patently in error.

    Even on-line ads for sources in Asia and other countries that actually advertise Linux computers also advertise Windows computers at a ten to one ratio. That figure just isn’t real no matter how you present it in some geographical niche.

    There is no question that Bing is in a very inferior position relative to Google and it is a continuing wonder as to why Microsoft would bother with continuing. Google has every bit of the monopoly on web search advertising as Microsoft may have on PC OS and, just as I think Microsoft will dominate on the PC forever, it seems obvious that Google will dominate as well in their arena. Google got to their lofty perch by people using Windows to access the internet and OS share is not important in maintaining that position.

  79. Thanks oiaohm. It was late… I was tired. Fixed it.

  80. oiaohm says:

    Clarence Moon you cannot develop better if you don’t know the mistakes of history. If you go through that compare of windows kernel vs Linux kernel part way through there are a list of features Other OS’s do better than the both of them.

    Basically its about knowing were we stand. Linux still has features to implement to have all the features of every unix that ever existed. Mostly solarias features are left.

    Clarence Moon many vendors are reporting as mobile device sales go up traditional windows pc sales are not expanding as expected after the down turns.

    Now either there forecasts were badly wrong or there is something starting to change. I really suspect something starting to change.

  81. Clarence Moon wrote, “The statistics quoted do not really represent the sort of “seat” quantity asserted in this blog article,”.

    Tell that to Bing which is losing ~$1billion per quarter. M$ said it is all-in for the cloud. Losing share of web access is a huge barrier to success. Meanwhile, Google continues to do well on all fronts. Opening up a new class of devices and rushing through the gap builds share for Google, kills Bill Gates’ dream of one licence per hard drive and promises to reduce seats if “8” is not a breakthrough. By the time “8” emerges, iOS or Android will be everyone’s favourite and resistance to change will make “8” a radical choice. “8” is years late to the party.

  82. Clarence Moon wrote, “Nothing in the history of Windows seems to have caused people to shun Windows for a future purchase, including the epic tales of poor performance by Windows Millenium and Windows Vista versions.”

    Not so. Most people get their OS with their PC and for the longest time nothing was offered but that other OS. Wherever there is choice we see GNU/Linux selling like hotcakes: eeePC first edition, BRIC countries, etc.

    Even businesses who are seriously locked in are doing their utmost to stick with XP to avoid Vista/”7″. M$ is selling 50 million licences for 90million PCs per quarter being shipped. That’s 55%, far below the peak of nearly 90% in the bad old days of monopoly. The monopoly is weakening rapidly. There is nothing before “8” to change that and changing to “8” from XP is too radical for many people. GNU/Linux is a more comfortable desktop environment than changing to “8”. Do the maths.

  83. Clarence Moon says:

    Such a tiresome comparison for and old Linux design to an old Windows design seems like a waste of time to me. Even then there are no apparent conclusions drawn from the litany of items covered. What is the purpose of such a thing?

    People who buy and use these OS products do so based on fitness for purpose and affordability. People have good and bad experiences with products and that history affects their future procurement activities. Nothing in the history of Windows seems to have caused people to shun Windows for a future purchase, including the epic tales of poor performance by Windows Millenium and Windows Vista versions.

  84. Clarence Moon says:

    The statistics quoted do not really represent the sort of “seat” quantity asserted in this blog article, I think. People using a mobile device to access the internet are not necessarily abandoning the use of concentional PCs. I admit that I am a little backward in terms of being on the front edge of adoption of these devices, for example I still use “feature phones” and have not bothered to obtain a data plan that would let me use them to access the internet, and I still have a couple of PCs, laptop and desktop. On the other hand, I have a lot of acquaintances, family, and friends who do use these things almost continually and they all have regular computers in addition to the mobile devices as well.

    Revenues and profits for Microsoft Windows can be dismissed out of hand if you want, but they still show that the financial impact of mobile device use is not yet a disaster in the making for Microsoft or other PC software suppliers for that matter. Symantec, Intuit, and the like are also doing well, given the general economic climate of the world.

    It is not in dispute that mobile devices are rapidly increasing in popularity and use, but there is no direct evidence to show that this represents an end to the existing PC software and hardware markets. I do not think that the two things are actually replacements for one another. Rather everyone I know has both.

  85. oiaohm says:

    Compatibility is one of those interesting myths about Linux.

    Linux kernel syscalls have a fairly strict rule about being compatible with prior forms. You break you better restore it. Most Linux compatibility issues trace to user-space that can be split by chroot support.

    Kernel space of Linux not supporting binary drivers have caused some issues from time to time. is another place to visit to see Linux internals

    There is a lot simlarity between way Linux kernel is broken up and NT design.

  86. Conzo says:

    Oiaohm, thanks for the link! Allow me some time to read through it 🙂

    I’ll get back to you on it, and I’d definitely be interested in reading an extensive document. Let me stress that I’m simply quite interested in the topic, because although I know a reasonable amount about it, I’m far from an expert in kernel design, so I find it very educational. Friendly teasing aside, waging war over OSes is quite far off my radar (crusading for Miles Davis’ Electric Period, now that’s a whole other ballgame 😉 ).

    From what I know about it, given the constraints (not in the least compatibility constraints which, although they don’t make the 2008 code base more elegant or optimized, in my experience (and use cases) compatibility is rarely the issue like it is in Linux) ) I lots of 2008 kernel design makes sense to me over the whole range of applications and use cases it has to cover.

    Otoh, I’d not call myself an expert (and have more things that I want to study in depth, than I have time for anyway) so, like I said, I’m quite interested (and wouldn’t want anyone to feel ‘called their bluff’).

    Will read the link you sent though, and thanks again!

  87. oiaohm says:

    Conzo I wish I had the IRC log from the Reactos channel on freenode.

    When I went head to head with one of the lead developers. He truly does know NT inside and out. I truly do know Linux not 100 percent but it was enough to give him major headaches as I hit design issue after design issue in NT. Funny enough mostly not caused by NT base design that is perfectly good but by the way MS implemented things. Like posix support not stand meaning posix permissions don’t work right on NT.

    Remember he had the idea that he would convert me to seeing Reactos as something other than a sub OS for compatibility reasons.

    There are some key differences

    Good place to start. He maintains a fair unbiased compare of the two cores.

    One of the big differences is due to the difference in design starting a process on Linux is very cheep compared to NT even that Linux is in fact starting a process with more secuirty controls on it.

    Basically I could write a 20000 word document on the disadvantages of 2008 code base compared to Linux.

    Do you really want it? Conzo or will you just settle for a fairly unbiased compare.

  88. Conzo says:

    You are extrapolating from, what, a whopping TWO data points within a system that’s highly non-linear (not to mention unpredictable) by nature ?

    Wow, gutsy!

    As an exercise, can you write me an in-detail essay, comparing current Linux kernel architecture and Vista/2008 codebase kernel, and explain why Linux “is” so “vastly” “superior”, and the 2008 codebase “is” such “clearly” “inferior” design?

  89. oiaohm says:

    You goofed this octobers link. Yes correct for oct link is

    The Linux Other figures are interesting. It clearly shows that Ubuntu accounts well less than half of the Linux desktop world. Now what are they using.

Leave a Reply