More Stats

  • This site:
    Country Vists Avg. Duration
    United States 70,885 00:03:03
    Canada 16,785 00:04:03
    United Kingdom 11,414 00:03:45
    Australia 9,346 00:08:34
    Germany 7,715 00:05:52
    Netherlands 4,400 00:01:29
    India 3,780 00:01:11
    France 3,066 00:00:42
    Finland 2,788 00:02:11
    Mexico 2,325 00:03:38
    …100 more…


    So, it seems to me the site is working and the effort is worthwhile no matter what the naysayers repeat ad nauseam.

  • Spain has rolled out Ubuntu GNU/Linux to 220K students. That should show up in the stats somewhere. Perhaps here:

    Nope. That’s just Clicky ignoring most GNU/Linux installations while noticing a new release of Ubuntu. The “Linux” category dropped the same way in many other countries without a huge roll-out of Ubuntu, just a new release.

  • Ubuntu does get a good share in many countries but Cuba is the most interesting:

  • Interestingly, Clicky shows Mac at less than Apple’s published shares… while NetApplications shows them greater… What’s with that?

  • Notice the sharp dips in Brazil’s stats from Clicky. Clearly they think most people have that other OS at home but work/school is GNU/Linux. We know there are a lot of people working/schooling, yet GNU/Linux share still shows tiny. What’s with that?

Overall, I think the stats again show that published web stats greatly underreport GNU/Linux. Even in Brazil where all students use GNU/Linux at school and most government employees use GNU/Linux and GNU/Linux is a best-seller at Walmart, Clicky still reports ~1%. The fluctuation between home and work/school are huge, like half the share. That should mean half the schools and government offices use GNU/Linux. Yet, they report a tiny share. What’s hiding behind the curtain? Inquiring minds want to know.

Here’s what I see of visits this month:

1.	 Windows	42,488	24.40%	
2.	 Linux		36,091	20.73%
3.	 Macintosh	6,386	3.67%
4.	 Android	2,386	1.37%
5.	 iOS		939	0.54%
6.	 iPad		936	0.54%
7.	 (not set)	856	0.49%
8.	 iPhone		740	0.43%
9.	 iPod		134	0.08%
10.	 BlackBerry	89	0.05%

To the argument that the site’s biased to GNU/Linux I would reply, “If so, why are so many users of that other OS visiting? They must be interested in GNU/Linux.” It’s all good. Clearly there are far more unique visitors than commentators on any side. I am reaching folks who want to know about how to do things in IT.

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.

48 Responses to More Stats

  1. ram says:

    You should be aware that Microsoft has few servers running bots that pretend to be browsing in order to make their market share look bigger.

    This is easy to check, just check who owns the IP address of your most frequent Microsoft browser users — tada! Microsoft Corporation!

    Might be throwing off your stats. For real laughs set up a page somewhere praising OS/2 and it will really get hit by Microsoft.

  2. oiaohm says:

    That Exploit Guy the diagram is not on the wikipedia its a video. It is interview with Microsoft own kernel coders.

  3. oiaohm says:

    That Exploit Guy the reality I just detailed a real flaw.

    That Exploit Guy
    –Hmph… “Look up tables”? “Cannot interface directly”? Did you look at the Wikipedia page, not understand what it was about (since you didn’t seem to know about what a “page table” was and instead called it a “PAE table”, which led to even more hilarity when you mistook the virtual memory address space for an individual process as a “box”), simply made up all this stuff and called it a day. Well, that’s kind of a nice try, but nothing more.–

    PAE table should have been PAE page table. Its a implementation detail. PAE page table how this works is nasty and effective. You know pages have rwx bits on them. Can you protect large blocks of memory quickly with PAE from being read/write.

    –virtual memory address space– is part of it. The change in table designs allow you to lock out big blocks at a time with PAE.
    http://en.wikipedia.org/w/index.php?title=File:X86_Paging_4K.svg&page=1
    http://en.wikipedia.org/w/index.php?title=File:X86_Paging_PAE_4K.svg&page=1

    PAE 4 pointers and you have changed the full memory map. You can now set 4 pointers to change between kernel mode and user mode.

    What is the new table PAE adds page-directory-pointer table(this have a bad habit of calling the PAE Table its the only table added). Can you quickly deference the user-space with PAE. Yep can you execute code if its first pointer is not in the page directory pointer table. No you cannot. Can you write to a page from userspace that the kernel has not provided you with the table to access no you cannot. Does PAE make it simple to have kernel and user-mode independent table sets yes it does. Once the userspace pointer is deference can you do SMAP or SMEP style attacks while the user-space is no longer referenced nop they are dead.

    You require PAE aware drivers to request memory to be put into userspace or got from userspace because you might have loaded the user-space pointer required up in different pointer also you are not required to keep that pointer around after the data has been transferred. To implement the security PAE can do by using many page tables.

    Those 4 pointers of PAE is what makes it practical to memory split userspace and kernel space completely. The fact those 4 pointers can be changed around is also why drivers have to be PAE aware to use it.

    PAE allows splitting user-space and kernel space pointers by the .

    Normal old style page tables on x86 you would have to slice threw slowly. page directory pointer table. Kernel mode is the only stuff that can alter that.

    Microsoft 32 bit allows old drivers still to use raw 32 bit pointers. Not having todo a request to find out what first 2 bits are required to access the wanted data and to get the page directory pointer table loaded.

    Windows 32 bit 4G limit is treating the first 4G as flat memory. It does not even allow for using page tables unique per application with what appears to be overlapping addresses.

  4. Wayne Borean wrote, “f you include Desktops, Notebooks, Netbooks, Servers, Tablets, Consoles, and Smartphones, Microsoft’s percentage of the OS market is about 30%”.

    That may well be correct but is irrelevant for some for two reasons:

    • the cloud/thin clients etc. make the desktop OS almost irrelevant, and
    • M$’s share of the decaying desktop/notebook market is declining very slowly. ie. They are still awash with cash.

    Both of those reasons are pretty weak because M$’s empire is not a castle but a house of cards. Pull out the right card and it crashes pretty quickly. Bye-bye cash cow. Some of the cards that can fall pretty quickly are:

    • the office suite – Google and others will take a huge bite as will LibreOffice/OpenOffice.org, and
    • retail shelf-sapce – Desktops and notebooks are not selling well to consumers because the other gadgets are sparkling and still quite pricey. Very little prevents retailers from offering GNU/Linux PCs to increase margin on low volumes. Netbooks with GNU/Linux still make sense and consumers and retailers will figure that out. Half of M$’s client market could disappear in a year or so.

    Not that we should mourn the loss of M$. They can last ~100 years just by firing their loyal employees and hiring Chinese. They will do whatever it takes to survive if the cash-cows dry up. Business is rapidly moving to web applications. They will be happy with cloud/FLOSS for desktop publishing/printing. So, the other half of the empire could go poof in a few years, too.

    The question of what happens to M$’s server software business is pretty clear. If people don’t have M$’s client OS they don’t have any use for inActiveDirectory and a lot of server components too. The empire has three legs and they are all weakening.

    “8” is their last shot at preserving the cash cows. “7” might do but they are not selling enough licences to hold back the tide. They could raise prices and some would pay but that would be the beginning of the end. M$ is terrified of waking up the users/suckers/slaves.

  5. Wayne Borean says:

    Simple. A lot of us hide our OS/Browser info. My Mac and my Linux boxes all claim to be Windows XP/IE7. At least they do in Firefox, in Seamonkey they report the truth.

    There’s still enough sites that tell you they can’t support Firefox, even though they can, that I keep it that way.

    Based on anecdotal evidence, Apple has about 25% of the notebook market in Canada. Linux is common on older Netbooks, but Microsoft did a pretty good job of killing it off, along with Netbooks themselves as a form factor. The iPad put the final nail in the coffin.

    If you include Desktops, Notebooks, Netbooks, Servers, Tablets, Consoles, and Smartphones, Microsoft’s percentage of the OS market is about 30% according to Tomi’s “Communities Dominate Brands” blog. I suspect he’s correct, but can’t prove it.

    What does that mean? Simple. Microsoft is one with the Walking Dead. The majority of Microsoft’s profits come from Office, and Office only runs on Windows and Macs. She total OS market share (Windows and Mac OS X combined) is shrinking (and Apple competes by supplying iWork as an Office competitor), the total number of Office licenses will shrink too.

    Yes, Microsoft still has a lot of cash, but much of it is overseas and can’t be brought back to the U.S.A. without paying corporate taxes (which is why they overpaid for Skype – purchase was done with overseas funds).

    Wayne

  6. That Exploit Guy says:

    And the hilarity continues…

    @oiaohm

    ‘When you push by WFP to kernel mode from a EXE resource. The problem is that EXE resource rw stats owns to the EXE. Basically WFP contains nice simple way to craft a rwx memory zone that you know exactly where it is since the application created it. This is a API/ABI design fault.’

    ‘rwx memory zone’? Just what on earth are you rambling about?

    You see how you are digging yourself into a hole here? You are trying to support an argument by pointing to things that don’t really exist. Seriously now, if you want plagiarise from Dan Rosenberg’s blog, at least try and understand what the guy says first before attempting to regurgitate it here.

    ‘The problem here that Microsoft Diagram does not in fact match reality as it first appears. Key hint you have the wrong diagram where is ntoskernel.exe its not listed on that diagram at all.’

    And where would it be in that super-duper secret Wikipedia diagram of yours? 😉

    ‘When you manage to get hands on full diagram you will find I am right on this.’

    Since when am I supposed to be the one doing all the leg work? You have made the claim, and you go and sustain it. If you can’t find anything to support your science fiction argument, that’s not my problem – it’s yours.

    ‘This is being a idiot who does not know the topic. Funny enough you have quote the PaX articles that explain it.’

    Oiaohm unduely called someone an “idiot”: 2

    Funny enough this has absolutely nothing to do with PAE and you are simply digging yourself into yet another hole with something you also have no understanding of.

    Two blunders for the price of one!

    ‘This is placing applications and kernel in different memory zones with different look up tables completely. You use PAE on 32 bit x86. So they cannot interface directly where they should not with pointers. So they have to make a PAE call to address across.’

    Hmph… “Look up tables”? “Cannot interface directly”? Did you look at the Wikipedia page, not understand what it was about (since you didn’t seem to know about what a “page table” was and instead called it a “PAE table”, which led to even more hilarity when you mistook the virtual memory address space for an individual process as a “box”), simply made up all this stuff and called it a day. Well, that’s kind of a nice try, but nothing more.

  7. oiaohm says:

    That Exploit Guy
    –As I have already explained, the user-space API is for user-mode components that reside in the user space. The user-mode portion of WFP simply works by make calls to the kernel-mode portion of WFP, not by somehow magically turning itself into kernel-space code. Do you know how a syscall works? Well, of course you don’t, but it’s the same principle regardless of you think it might be.–
    When you push by WFP to kernel mode from a EXE resource. The problem is that EXE resource rw stats owns to the EXE. Basically WFP contains nice simple way to craft a rwx memory zone that you know exactly where it is since the application created it. This is a API/ABI design fault.

    Basically allowing the usermode to load the kernel mode from inside the same file so linking the two memory spaces is a nasty big screw up. Some programs doing firewalls and otherwise have exploited this feature to send stats back to usermode without using WFP as well. You never want kernel mode memory given to user mode todo what it wants with it.

    That Exploit Guy
    –That’s very interesting. Mind if I ask from which part of this diagram you saw the following?–
    The problem here that Microsoft Diagram does not in fact match reality as it first appears. Key hint you have the wrong diagram where is ntoskernel.exe its not listed on that diagram at all. You have the internal architecture diagram not the physical. If you watch the Microsoft video on Windows NT internals you find this out. This gives you a color diagram of where everything really is.

    Graphical Digital Interface that appears to be in the Win32K.sys is not. In fact the Window manager under Win32K.sys is not in Win32K.sys either. Under vista on this is in dwm.exe.

    Each of the boxes in the Executive services was a .sys at one point so each does have a filename. Since then there have been merges. Most of the executive over time got merged into Ntoskernel. This is the problem most MS diagrams does not show the files the parts are in. In fact trick people into believing where stuff is not. Yes people thinking the GDI is in the win32k.sys is not uncommon yet wrong.

    When you manage to get hands on full diagram you will find I am right on this.

    That Exploit Guy
    –So you are telling me that limiting the total number of available addresses to 2^32 somehow magically creates a security hole that eludes all the greatest of minds in operating systems security. False, but imaginative.–
    This is being a idiot who does not know the topic. Funny enough you have quote the PaX articles that explain it.

    How do you implement on x86 smap and smep functionality without the cpu supporting those instructions.

    x86 segmentation is the answer. This is placing applications and kernel in different memory zones with different look up tables completely. You use PAE on 32 bit x86. So they cannot interface directly where they should not with pointers. So they have to make a PAE call to address across.

    X86 segmentation requires your drivers to PAE aware.

    So in effect 32 bit only limit disabled implementing smap and smep like functionality in Windows. Its a major security defect. PAE can be used for more than just allocating large memory. The fact PAE gives 4G blocks and can trap when something is crossing between is a huge security advantage.

    MS choice blocks means to implement other functionality based in PAE.

    In fact go back and read what the auther of PAX writes about how she implemented smap and smep functionality and others using PAE abilities to segment memory.

    How many segments does windows drivers allow under 32 bit. 1 segment. So you cannot segment so you cannot implement security features.

    That Exploit Guy you have liked quote her if you had read here you would know that the MS choice is a security flaw because it blocks implementing many different protections.

  8. That Exploit Guy says:

    @oiaohm

    This is getting really tiresome, don’t you think? You obviously can’t even see an API even if someone puts its entire set of documentation right in front of you. All you are doing here is sustaining a factually deficient argument with your faulty understanding of the most basic things in programming. That’s quite frankly not a respectful gesture to all the due effort I have put into replying your comments, is it?

    ‘Because a resource in your executable can remain writeable to your executable. So that kernel mode WFP you just loaded into kernelspace is now a large memory foot print you can overwrite from userspace.’

    As I have already explained, the user-space API is for user-mode components that reside in the user space. The user-mode portion of WFP simply works by make calls to the kernel-mode portion of WFP, not by somehow magically turning itself into kernel-space code. Do you know how a syscall works? Well, of course you don’t, but it’s the same principle regardless of you think it might be.

    ‘Look at the architecture picture’

    That’s very interesting. Mind if I ask from which part of this diagram you saw the following?

    ‘GDI call path is to subsystem what is Win32k.sys to Ntoskernl section wingdi.sys from there to video driver. Then hike all the way back. Because the executive is in the ntoskernel.’

    This could very well be the first instance of the ink blot test without the use of ink blots. Maybe I should get a degree in psychology. Hmph…

  9. oiaohm says:

    That Exploit Guy
    –You mean just like how WFP magically turns user space code into kernel space code?–
    Because a resource in your executable can remain writeable to your executable. So that kernel mode WFP you just loaded into kernelspace is now a large memory foot print you can overwrite from userspace.

    That Exploit Guy go speak to the reactos developers who did the disassemble over where the GDI truly is. GDI call path is to subsystem what is Win32k.sys to Ntoskernl section wingdi.sys from there to video driver. Then hike all the way back. Because the executive is in the ntoskernel.

    http://en.wikipedia.org/wiki/File:Windows_2000_architecture.svg Look at the architecture picture.

    That Exploit Guy OS/2 subsystem use to video out directly to Ntoskernel wingdi.

    That Exploit Guy you are trying to google answers and its shows there is a good video where the microsoft developers walk threw what is where. It also tells you the GDI is in Ntoskernel.

  10. That Exploit Guy says:

    @oiaohm

    ‘Win32k.sys is not the GDI that is the Win32 subsystem there is a difference. That Exploit Guy the GDI is bundled into ntoskrnl.exe along with HAL and a few other-things. That in early NT were separate dlls. There is zero separation from the GDI to the kernel.’

    No, you are wrong again.

    ‘And you are just like most people who don’t know what the Windows GDI kernel mode name is. WinGDI.sys is the name. You don’t see the file any more unless you disassemble windows.’

    Sure, sure! Did you use the same special gene that you use to jam wifi for disassembling Windows as well?

    ‘Also look at that diagram. DDI is the calls to display driver. The display driver is allowed to call the GDI using the same functions the user-space calls the GDI with.’

    You mean just like how WFP magically turns user space code into kernel space code?

    You have no idea what an “API” is or how it works, do you?

    ‘Microsoft support loading the kernel mode call back parts from a resource entry in the user-mode file. That Exploit Guy just as I said before about windows security you can query before doing to find out if you are currently allowed.’

    Nope, you apparently did not take anything to heart when I mentioned POSIX syscalls. That’s just typical oiaohm, isn’t it?

    ‘List of OS implementation of NX bit. Linux was first. Year 2000 for non mainline. 2003 for mainline. Windows XP SP2 is 2004.’

    Slow down there, oiaohm. Are you sure it’s safe to keep copying and pasting all these things apparently “most people don’t know” from Wikipedia to here? I don’t want to be abducted by a black helicopter, you know? 😉

    NX bit was retrofitted to PAE when AMD introduced the Athlon 64 series of processors. That happened in 2004. Are you telling me that Linux had already implemented a CPU feature even before it existed?

    ‘That Exploit Guy this artificial ceiling is a security hole.’

    So you are telling me that limiting the total number of available addresses to 2^32 somehow magically creates a security hole that eludes all the greatest of minds in operating systems security. False, but imaginative.

    ‘Because Microsoft only fires up one PAE table.’

    Ah, really? Mind if you show me how such magical paging mechanism operates in practice?

  11. oiaohm says:

    Win32k.sys is not the GDI that is the Win32 subsystem there is a difference. That Exploit Guy the GDI is bundled into ntoskrnl.exe along with HAL and a few other-things. That in early NT were separate dlls. There is zero separation from the GDI to the kernel.

    And you are just like most people who don’t know what the Windows GDI kernel mode name is. WinGDI.sys is the name. You don’t see the file any more unless you disassemble windows.

    –Thirdly, the function of the kernel space component of GDI to communicate with the graphics driver itself.–
    Its not just communication. Kernel mode GDI also contains software emulation for feature the video card might not support some of these feature video cards have never supported. Also look at that diagram. DDI is the calls to display driver. The display driver is allowed to call the GDI using the same functions the user-space calls the GDI with. Is it possible to never ending loop yes it is. It is possible for the GDI in kernel mode not to be able to tell the difference. So video driver advertises it suppors X DDI feature the GDI calls it and it calls back the GDI you have a never ending loop. Yes this has happened. That this can happen is poor kernel design.

    Most of the executive these days is not independent files.

    That Exploit Guy go back and read that WPF diagram you quoted. Notice the third party call backs and then notice the in kernel multi plexing WPF does. Then test and find out that you cannot tell the difference in userspace between a message coming from the call backs. So a message coming from a call back could be any one. Then look up how those call backs are loaded.

    The thing you need to check is how the kernel mode components were loading. Microsoft support loading the kernel mode call back parts from a resource entry in the user-mode file. That Exploit Guy just as I said before about windows security you can query before doing to find out if you are currently allowed. So its perfect-ally designed to virus infect kernel. I could not design a interface better for getting infection code into kernel space without user knowing.

    That Exploit Guy you need to dig deeper than reading the overviews. When I say if you dig. I really do mean dig. Lots and lots of reading. So one executable downloaded without placing extra files on the disc can insert something into kernel. This is not good.

    That Exploit Guy
    http://en.wikipedia.org/wiki/NX_bit
    List of OS implementation of NX bit. Linux was first. Year 2000 for non mainline. 2003 for mainline. Windows XP SP2 is 2004.
    Full PAE appears in Linux year 1999.

    –The 4GB limit is merely an artificial ceiling for maintaining compatibility with drivers that are not capable of handling an address space greater than 32 bits.–

    That Exploit Guy this artificial ceiling is a security hole. Because Microsoft only fires up one PAE table. Linux each application gets almost a 4gb of own addressable memory. This means unless memory is market shared on a Linux PAE system applications cannot look sideways with normal pointers. So a buffer flow in one application cannot go into another. Unless it crosses a shared point that is read write.

    So the Microsoft statement should read [Since our drivers don’t know how to support PAE we basically disable PAE and place everything in the one PAE box so you don’t get the security advantage of PAE.]

    Microsoft is slow to implement the kernel features for security and when they do they have a habit of basically disabling what they are design todo.

    That Exploit Guy this is the problem. Microsoft tells you that they have stuffed it. Problem is people like you are too either dumb or informed enough that you don’t straight up see that hey you should not have done that.

    PAE should be implemented properly. Stuff the drivers that don’t work get the driver makers to fix their drivers. This is a security issue. Microsoft will bend to what is simpler for driver makers then do what is good for OS security.

    Closed source driver makers for Linux hate Linux some of the time because Linux puts security first. So if there is a new security feature and it breaks your driver stiff fix it.

  12. eug wrote, “Is Linux market growth stagnated (at least)?”

    Only if you count faulty stats misinterpreted. Much of government, education, large businesses use firewalls which NAT PCs, so there is no way to tell the OS share behind the firewall. Do they count unique Useragent strings regardless of IP address? Some report page views. How do they report the page view of a thin client running GNU/Linux from a terminal server running that other OS or a thin client running that other OS from a terminal server running GNU/Linux? Then you have to explain why Google’s employees running GNU/Linux doubles GNU/Linux share for USA when only ~10K employees run the OS?

    For September 2012, NetApplications shows 1.93% GNU/Linux for “United States”. Without Sunnyvale, CA, they show 0.82%. We know there are many more users of GNU/Linux in USA than just Google…

  13. That Exploit Guy says:

    @oiaohm

    ‘Linux most of its drivers are mainline. This means as soon as a new cpu security feature is release Linux can implement it. Microsoft has to hurd the cats of driver developers that it is quite unable todo. Most public example in 32 bit windows only being able to address 4G of ram because windows drivers are not PAE aware like it took MS years to be able to turn NX on.’

    NX bit has been supported in Windows since as early as XP Service Pack 2. The boot flags “/PAE” and “/NOEXECUTE” enables the feature. The 4GB limit is merely an artificial ceiling for maintaining compatibility with drivers that are not capable of handling an address space greater than 32 bits. All these things are well-documented in MSDN as well as third-party sources. Look them up.

    ‘the permission systems of Linux with s LSM loads are stricter.’

    Again, you have completely misunderstood the purpose of SELinux and LSM in general.

    All that I need to explain have already been explained here. If you have anything to say, please reply to me in that blog entry.

    ‘Linux Security Models forbids applications directly looking at what is monitoring them without special permission.’

    Define “monitoring”.

    ‘Most people don’t know this. WinGDI.sys use to run in userspace along with everything else in the executive.’

    By ‘most people don’t know this’ do you mean ‘I looked up Wikipedia and saw this thing’? That’s quite a lousy way to impress people, isn’t it?

    Firstly, the kernel space component of GDI is provided by Win32k.sys. This is seperate from ntoskrnl.exe, which is the main kernel binary code itself.

    Secondly, most of library code for creating desktop objects are provided by gdi32.dll and user32.dll, which reside in user space.

    Thirdly, the function of the kernel space component of GDI to communicate with the graphics driver itself. See here.

    If what you worry about is the computer trying to show you something on the screen, you might as well just unplug it and bury it in concrete for good.

    ‘You don’t read the MSDN. And if you dig deeper the User-mode API can result in sections of the program being loaded in executive what is ring 0.’

    I am beginning to wonder if you actually understanding what you are reading.

    The user mode API does not result code being loaded into the exective. That’s preposterous. WFP merely provides two seperate APIs, one for user mode components, and one for kernel mode components. Here, read the diagram.

    ‘So NT is some horible mix between a Monolithic kernel and Mircokernel.’

    It’s called a “hybrid” kernel. Look it up.

    ‘Then the rest of your comments was insults That Exploit Guy.’

    I think it’s absolutely fair to say your comments are generally incomprehensible, lacks factual content and mostly irrelevant to what they are replying to. Given your frequent use of the word “idiot” to address those disagreeing with you, an honest comment on the quality of your posts ought not to be too much for you to take.

  14. oiaohm says:

    That Exploit Guy Let me remind you.
    –Thirdly, remember the Debian OpenSSL fiasco? Yeah, because having people not understanding the code to examine the code for you has always worked out so well for everybody.–
    This is complete wrong.

    That Exploit Guy I will give dougman comments are not complete right. But nothing he has said is completely wrong.

    dougman
    –The windows architecture is what makes it vulnerable, Linux is not as vulnerable because its architecture is designed right, the file system and permissions make it much stronger. The Linux philosophy regarding what may be executed and when, is much more robust.–
    This is 100 percent correct. Dougman most likely does not fully understand the details why.

    That Exploit Guy and I can the major ones.

    1 Linux most of its drivers are mainline. This means as soon as a new cpu security feature is release Linux can implement it. Microsoft has to hurd the cats of driver developers that it is quite unable todo. Most public example in 32 bit windows only being able to address 4G of ram because windows drivers are not PAE aware like it took MS years to be able to turn NX on.
    2 the permission systems of Linux with s LSM loads are stricter.
    3 Linux Security Models forbids applications directly looking at what is monitoring them without special permission.

    Basically the permissions start at the memory manager and cpu features. The fact Linux uses more of them than Windows.

    A high security OS has most of its drivers mainline.

    dougman
    –Windows is vulnerable primarily because they integrated the desktop and made networking an application, protected by third parties who have access to the kernel. So it will always remain flawed by design.–
    That Exploit Guy this from dougman is correct on this one. But you don’t know the correct information.

    Integrated the desktop refers to a particular time in the NT development. Early on in NT design it was a microkernel. Most people don’t know this. WinGDI.sys use to run in userspace along with everything else in the executive. Most of the functional equals WinGDI in Linux still do in user space same with a lot of other sections in the windows executive. Even after kernel mode setting and Wayland the GDI parts are still mostly userspace along with the complete windows manager. dwm in vista on moved the windows manager out of being ring 0. Lots and lots of bugs have happened in the GDI parts.

    http://en.wikipedia.org/wiki/File:Windows_2000_architecture.svg This is windows 2000. Please not even that the executive is displayed in a box. It runs at ring zero. Windows XP looks the same. The windows manager in XP is running in ring 0 with every else about the kernel.

    Also notice where your use security information is. Interesting enough not in ring 0 in the same ring as the user. Micro kernel location with the same old microkernel bug. You can in fact debug the windows security permissions up XP and change them on fly.

    So even running server core you have features in ring 0 that by NT design were not meant to be in ring 0. This is all due to what was done when the desktop was integrated.

    That Exploit Guy
    –“Protected by third parties who have access to the kernel”? I didn’t realise the Windows Filtering Platform was third-party and ran in user mode! Care to explain?–
    Even today out box Windows 7 and 8 don’t come with a fully functional firewall.
    http://msdn.microsoft.com/en-us/library/windows/desktop/aa366510%28v=vs.85%29.aspx
    –The WFP API consists of a user-mode API and a kernel-mode API.–
    You don’t read the MSDN. And if you dig deeper the User-mode API can result in sections of the program being loaded in executive what is ring 0. This is a hang over from NT being a Micro kernel with the executive in user-space. The file format of the executive and user-space is basically identical.

    The design documents of NT have the executive running as ring 1. This was dropped to increase performance. Higher performance lower security.

    So NT is some horible mix between a Monolithic kernel and Mircokernel. Please attempt to call hybrid. I call NT is botched Microkernel. Because lot of Microkernel stuff has been left behind.

    Linux and Minux(that is a micro kernel) have always keep system and user application binaries as different buggers. That will not load with the others loader. Minux system service is not the same as a Minux application. NT just to be scary they are basically identical.

    Then the rest of your comments was insults That Exploit Guy.

    Now I have read from the start to the end you are even more incompetent than I believed That Exploit Guy.

  15. That Exploit Guy says:

    @dougman

    ‘Pithy put-downs’

    “Chris Weig, the nagging Troll and arm-chair Geek squad guru speaks his mind!”

    ‘Name-calling and insults’

    “Exploited GUY”

    ‘Ad hominem attacks that try to negate an opinion by alleging negatives about the person supporting it’

    “Well, its rather obvious that you enjoy trolling for M$, hopefully they are paying you something.”

    ‘Impugning other’s motives’

    “He dreams that every prospective employer would soon show Linux users the door, as he needs to continue funding his MVP status while pushing M$ drugs.”

    ‘Emotional rants’

    “With M$ building all these data centers around the country, one would think that with all the money M$ has been milking ignorant users, that it could build a software repository, its obvious that they do not care, not one bit.”

    ‘Completely off-topic posts’

    http://mrpogson.com/2012/09/30/more-stats/#comment-98490

    ‘posting inaccurate “facts”’

    “The desktop can completely crash on Linux… and the system stays up.”

    Any more accusations you would like you make, my dear IT expert?

    By the way, have you considered answering my questions here yet?

  16. dougman says:

    Exploited GUY, your cooperation will be greatly appreciated if your buddy, Armchair Chris Wiggies, would cease and desist your trolling activities and return to the nearest cave.

    Trolls divert online discussions into non-productive, off-topic venues. They pose as part of a community only to disrupt it. Trolling is anti-social behavior. Some of the techniques trolls use to accomplish their objectives are:

    Pithy put-downs
    Name-calling and insults
    Ad hominem attacks that try to negate an opinion by alleging negatives about the person supporting it
    Impugning other’s motives
    Emotional rants
    Bullying and harassment
    Completely off-topic posts
    Posting inaccurate “facts”

    The traditional definition of trolling includes intent. That is, trolls purposely disrupt forums. This definition is too narrow. Whether someone intends to disrupt a thread or not, the results are the same if they do.

    The problem with trolling is that a small minority can destroy a web site’s usefulness for the majority of well-intentioned, well-behaved participants. Trolls won’t stop if you ask them. They hide behind anonymity. Most would not post the way they do if they were not anonymous. Thus mechanisms that undermine anonymity and enforce personal responsibility deter them.

    Amazon deters trolling through a qualification system. One has to qualify in order to post. Their system requires personal information, a verifiable email address, and a verifiable credit card. Other web sites qualify commenters through paid memberships, technical quizzes, or using real names in posts.

  17. That Exploit Guy says:

    @oiaohm

    ‘That Exploit Guy you have got facts wrong what I have had to be correcting.’

    In case you haven’t noticed, I have been trying to ignore you in this blog entry. You don’t even seem to have the courtesy to at least take note of what is being discussed but instead mindlessly drag a completely irrelevant discussion from another entry all the way to here. No, the OpenSSL fiasco has nothing to do with the kernel because the kernel is not what is being discussed here. Why not read the comments first before interjecting with your two-bobs worth?

    Your cooperation will be greatly appreciated!

  18. oiaohm says:

    That Exploit Guy what you have to play the person. You are meant to be an expert. Nothing about being an expert does it mean you have to go after the person like this.

    That Exploit Guy really dougman has never played the expert card. He has played the IT contractor card. There is a difference. IT contractors are not perfect we know this.

    That Exploit Guy you have got facts wrong what I have had to be correcting.

    You cannot expect dougman to believe you if you cannot basic description of what an exploit was right with why it happened.

  19. That Exploit Guy says:

    @dougman

    ‘What’s the matter, you don’t guarantee your work?’

    Yeah, because people are so unsure about their work that they need to resort to infomercial gimmicks in order to detach themselves from any responsibility 90 days thereafter.

    In all honestly, I have yet seen a serious IT contractor who doesn’t sell their solutions along with support services.

    Of course, this is not to mention the part about enticing potential customers with blatent misformation and hoax ‘surveys’.

    ‘Well, its rather obvious that you enjoy trolling for M$, hopefully they are paying you something.’

    I take this as your admission that you don’t know the answers to all the rest of my questions, then, my so-called IT expert.

  20. dougman says:

    Re: I certain don’t go around offering … a 90-day ‘guarantee’. That’s for sure.

    What’s the matter, you don’t guarantee your work? Oh wait what sort of work ‘DO’ you do? Most certainly nothing to do with IT that’s for sure.

    Re: you have mistaken me for someone here who has a cult-like obsession with what OS people use for their business.

    Well, its rather obvious that you enjoy trolling for M$, hopefully they are paying you something. if not then you have some vested interest in maintaining the status quo. Sorry buddy, the universe doesn’t play like that. The world is a dynamic place and is always ripe for change.

    The next big thing will be Android devices that roll into a Desktop. http://www.ubuntu.com/devices/android

  21. oiaohm says:

    –That Exploit Guy please don’t be a idiot and use examples you truly don’t understand.–

    Sorry should not have done that one. There is a implied context bracket in that line. People with general level english would not get it. So read it the wrong way in some cases.

    –That Exploit Guy please don’t be a idiot. In future use examples you truly do understand.– Is more simple english. Means exactly the same as what I typed before.

  22. oiaohm says:

    That Exploit Guy the debian OpenSSL fiasco contained no kernel breach code. It is a classic example why user-space is just as critical as kernel space.

    If you were running a OpenSSL like that on a high security OS kernel would it make any different to your means to steal data from it.

    That Exploit Guy
    –Yeah, because having people not understanding the code to examine the code for you has always worked out so well for everybody.–

    Technically the source code of OpenSSL was not defective. Its test-suite and configuration system did not check something.

    Openssl on debian is a example that the complier is just as import as the code. A+B=exploit case. So you examine A gcc what it is doing looks sane. You examine B what is Openssl looks sane. The problem is the exploit is C and only exists in C if A gcc was passed a particular option. Debian Openssl case you could have inspected the code bases of Openssl and gcc in isolation forever and not seen the problem.

    That Exploit Guy for you title you know jack about the Debian OpenSSL exploit. Because the Debian Openssl is not a failure of code inspection.

    Failure 1 to make a good enough test suite.
    Failure 2 not to check complier options at build time that they are in some form of known acceptable.
    Failure 3 the debian maintainer due to time skipped over running the test suite but this would have made no different to the outcome because the test suite of openssl did not check it function well enough to be able to detect the A+B event.

    Failure of test-suite and people failing to run test-suites get it right in future.

    Code inspection likely hood of finding the fault is basically zero without a testsuite. This makes no difference if openssl had been closed or open source.

    Remember closed source compliers like people use on windows you would not be able to look inside and see hey in this event the complier does something stupid to the code.

    The probability of A+B faults is higher on windows because a code inspect cannot read A and B and see what the resulting C is. There is a direct disadvantage to closed source. Without the source code to gcc working on what had gone wrong with OpenSSL could of taken longer. A closed source complier could have turned the feature on as a trade secret to run faster.

    The only way you know that you hit a fault like Debian OpenSSL on windows using Microsoft compliers is if the test-suite finds it.

    There are proper examples out there of having people not understand the code to examine the code failing todo their job. The Debian OpenSSL is not that.

    That Exploit Guy please don’t be a idiot and use examples you truly don’t understand. The reality here is the Debian OpenSSL bug should be a wake up call to everyone to make sure the test suites are quality.

  23. That Exploit Guy says:

    @dougman

    ‘CNET?? Tucows?? Both are known for malware, go ahead and Google it, oh wait make that BING. Google uses Linux and its evil.’

    First off, you have mistaken me for someone here who has a cult-like obsession with what OS people use for their business. I certain don’t go around offering to pave over their Windows system with Linux with a 90-day ‘guarantee’. That’s for sure.

    Secondly, if that’s about how much you manage to come back to me with after all those questions I have lobbed at you, you might as well stop trying to pretend you know anything.

    Thirdly, remember the Debian OpenSSL fiasco? Yeah, because having people not understanding the code to examine the code for you has always worked out so well for everybody.

  24. dougman says:

    If someone brought a Windows computer to you, you’d waltz over it with some Linux distribution and call the problem fixed….that shows our misinformed you really are.

    Re: No one but M$ can fix it. As long as they can force people to support M$’s extravagant lifestyle they won’t change.

    True, but I gesture to say that EVEN M$ cannot fix some things. I remember reading a story about Ballmer bringing in a malware infected desktop and told his engineers to fix it, they couldn’t without reinstalling the software.

    When you get infected its best to just reinstall the OS to be sure that nothing is lingering in the background, especially more so these days with all the zero-day exploits floating around.

  25. Chris Weig says:

    Oiaohm, I haven’t included you. I specifically meant dougman and Pogson. The recent ramblings of those two show their deep religious devotion to Linux and nothing else.

  26. oiaohm says:

    Chris Weig please don’t include me in that. I don’t call the problem fixed just by using Linux.

    Just Linux on the scale of OS’s for security is higher up the tree than Windows.

    Linux could be way stronger than what is now. So could Windows. If end users started demanding Microsoft and Linux world provide quality.

    I am not going to play that Windows is quality when I know its a lie. If a person wants to use windows and accepts its problems and its willing to pay to have those problems fixed I will fit windows for them.

    Same with a person who takes Linux. Its the people who think its all a 100 percent free lunch like because they caught a virus and its under warnatiy they don’t have to pay really annoy me.

  27. Chris Weig says:

    A good IT guy knows how to fix ALL IT. This includes Windows, OSX, Linux, iOS/Android smartphones, computer hardware and even thumb-drives chewed by one’s dog.

    That’s not you then. And it’s definitely not Pogson. You people hate anything non-Linux much too much. If someone brought a Windows computer to you, you’d waltz over it with some Linux distribution and call the problem fixed.

    Your skills as IT guys rate worse than that of butchers. (Apologies to the butchers.)

  28. oiaohm says:

    That Exploit Guy really you don’t seam to know much.

    Even kernel panic does not have to equal a Linux system dead. Ubuntu chooses that as the outcome.
    http://en.wikipedia.org/wiki/Kexec key framework to it. So you start up and load a reserve kernel sitting their waiting to take over in case if disaster.

    Yes it possible to do a full memory dump and recover what the user was working on.

    http://coldattic.info/shvedsky/pro/blogs/a-foo-walks-into-a-bar/posts/86

    Yes super stubborn Linux systems exist. That while there is a chance to run they will keep on attempting.

    Basically That Exploit Guy linux systems are no where near as strong as they could be by kernel design. Because the Linux User-space in most distributions is not activating those features.

    Can you get my point yet that OS resistance is also dependant on the user space setup correctly. Security and stability is a complex problem.

    That Exploit Guy yes shock reality right that the Linux kernel is design to cope with the case that the kernel itself fails. This is why there are options in selinux and others to trigger a kernel panic. Since if a attacker has not got deep enough to get at reserve kernel this can see every other alteration to the system flushed. The reserve kernel can be set to back and restore.

  29. dougman wrote, “A good IT guy knows how to fix ALL IT. This includes Windows”

    No one but M$ can fix it. As long as they can force people to support M$’s extravagant lifestyle they won’t change.

  30. dougman says:

    CNET?? Tucows?? Both are known for malware, go ahead and Google it, oh wait make that BING. Google uses Linux and its evil.

    With M$ building all these data centers around the country, one would think that with all the money M$ has been milking ignorant users, that it could build a software repository, its obvious that they do not care, not one bit.

    Package managers in linux have been around for years.

    The Windows store is expected to release Oct 26th, and M$ will call it the most innovative thing to date.

    I do not use Windows 7, have no need for it. Just because I do not use something or have a preference to another OS, doesn’t mean I do not know how to fix or use it.

    A good IT guy knows how to fix ALL IT. This includes Windows, OSX, Linux, iOS/Android smartphones, computer hardware and even thumb-drives chewed by one’s dog.

  31. That Exploit Guy says:

    @dougman

    ‘The windows architecture is what makes it vulnerable, Linux is not as vulnerable because its architecture is designed right, the file system and permissions make it much stronger. The Linux philosophy regarding what may be executed and when, is much more robust.’

    Just as I thought oiaohm was the only person who read a couple of product brouchers from Red Hat and considered himself a security expert, you suddenly came along and turned my perception completely upside down!

    ‘Windows is vulnerable primarily because they integrated the desktop and made networking an application, protected by third parties who have access to the kernel. So it will always remain flawed by design.’

    This seems all very interesting.

    “Integrated the desktop”? Where can I get this “desktop” on a Server Core install? You know, not just that slate of greenish colour on the screen. The desktop.

    “Made networking an application”? Sweet! Does it come in EXE or MSI?

    “Protected by third parties who have access to the kernel”? I didn’t realise the Windows Filtering Platform was third-party and ran in user mode! Care to explain?

    “The desktop can completely crash on Linux… and the system stays up”? Really?

    “Windows does not have this feature. Why NOT?” Well, ever heard of CNET? Tucows? They have been around since the 90s and everything does get checked before it gets put up there, ya know?

    “Actually, I don’t make any revenue by marking up software nor hardware, I make enough with support and services.” And that why it’s important to ensure your customers come back as frequently as possible for your “support and services”. *cough* razor-and-blade *cough*

    ‘and ESET NOD32.’

    Affiliate commission, here I come! *cha-ching*

  32. oldman says:

    ” So what if I make money on ignorant Window users, M$ does the very same and you applaud their efforts.”

    So tell me, do you actually Run windows 7 Doug?

  33. oiaohm says:

    The reality is that from a technical stand point of security Linux is not brilliant.

    Windows is a complete disaster zone. You pick up the old USA mil rainbow books that were deprecated because it was classed impossible to make a OS to that quality.

    Today we know its not impossible to make a OS to rainbow books quality there are 4 OS’s in existence that pass the all rainbow books and the newer certification. 3 are FOSS 1 is closed source. None is Linux or Windows. The 4 can run Linux applications contained.

    The issue is applications and hardware support.

    The new certification shows Windows and Linux about the same quality level. EAL4+. The bad part here is EAL4+ refers to development method.

    So I could make a OS that the only thing it did was display hello world and get the highest EAL7+ rating.

    Scary enough EAL numbers don’t demand any runtime protection from security flaws for the OS to contain.

    This is why I laugh at people who throw around EAL numbers.

    Selinux is based off the old rainbow books. So if you compare windows to linux by the old rainbow books that describe the security frameworks you have to have to get particular ratings. Windows and Linux are miles apart. The very best of the OS’s Linux is only minor-ally behind.

    So in the security world Linux development and test method is targeted for not being up to scratch. Windows is the same quality in this regard because if was not it would have a higher EAL.

    Pure security Windows vs Linux. Linux wins every time.

    Linux vs everything. Says Linux can do better. Some of the reason why it so hard for Linux to do better is the broad hardware support it has.

    Linux at this stage has not been fully hardened for desktop usage. Linux is taking on the core issues of desktop.

    Chris Weig the reality is the documentation to assess an OS’s security quality exist. Have existed for ages. Microsoft Trolls don’t want to admit that Windows is at the bottom of the security. There are FOSS things around that level. Like freedos.

    So yes it possible to write a tables with OS’s order based on security quality. From most secure in design to least secure in design and from most secure in development method to least secure in development method.

    One table MS and Linux is close to each other. One table they are miles apart. There are over 6 OS’s that are between Linux and Windows in security quality. Yes Windows is near the bottom end.

    That OS can be made to the rainbow book quality Microsoft has not announced when they will be delivering this or any progress to it.

  34. dougman says:

    Uber-geek Chris Weig, speaks with authority on a speech that I gave titled “No More Malware” but is ignorant of the facts of what I spoke about nor was he present.

    I always make it known that no operating system is foolproof, as there are no guarantees in life. Everyone in the IT industry knows that Linux isn’t malware-free, but it’s so much less susceptible to attack like .0001% less, than that of Windows.

    I even point out this link http://en.wikipedia.org/wiki/Linux_malware for people to read; whats pertinent is the question of management of the network and operating systems within the environment and system can be targeted and pwned.

    However, Linux is more secure than Windows, by orders of magnitude and it shall remain that way.

    The windows architecture is what makes it vulnerable, Linux is not as vulnerable because its architecture is designed right, the file system and permissions make it much stronger. The Linux philosophy regarding what may be executed and when, is much more robust.

    Windows is vulnerable primarily because they integrated the desktop and made networking an application, protected by third parties who have access to the kernel. So it will always remain flawed by design.

    Linux integrates the networking and does not give access of kernel to third parties and makes the desktop an app. On Linux the firewall IS the kernel… and it has the designed ability for user lock-down. The desktop can completely crash on Linux… and the system stays up. try that with Windows, awwww mannnn BSOD!!, I need to reboot!!

    When you are installing software through aptitude (or apt-get) then you’re installing from trusted repositories anyway, which is software that has been checked, tested and verified before it is accepted and re-distributed by your Linux vendor. Windows does not have this feature. Why NOT?

    Actually, I don’t make any revenue by marking up software nor hardware, I make enough with support and services. It would be unscrupulous to do so and something I totally call competitors out on. So what if I make money on ignorant Window users, M$ does the very same and you applaud their efforts.

    For example: The idiot down the street that tried to sell a *custom* computer to a businesswoman, for the low-cost price of $1000.00, in that was cheap in his own words. I had her purchase a new system with Windows 7 for $299.00 as that is what she wanted and got her all setup with Libreoffice, Chrome and ESET NOD32.

    Yep, I speak FUD and nothing but the FUD…………Fast Uncompromising Discussions regarding Linux.

  35. Chris Weig wrote, “You’re a poor FUD artist, dougman.”

    Takes one to know one.

  36. Chris Weg went on and on about stuff, “I know too many Linux users in my vicinity who pose as Linux “specialists”, but in most cases only possess dangerous half-knowledge.”

    I built a “whole” IT system at Easterville, not half. I designed and built it without anyone’s help except to string cables in the racks, to build a dozen custom-made PCs and to distribute and plug in thin clients. I detailed every part, every cost, built the servers, installed the software and made the schedule. I did make a couple of mistakes but that was because I had to do a month’s work in ten days because of other contractors not turning the building over on schedule. I forgot to enable LDAP until folks started changing passwords on different terminal servers… The system was usable on Day One under budget. Was the project dangerous? Nope. GNU/Linux is a well-tried OS known to be rugged and flexible.

  37. Chris Weig says:

    Well, dougman, let me reply to your well thought-out statement one by one.

    1. Your title “No more Malware” presupposes that Linux’s Desktop market share remains at 1 or 2 percent. You got that right. Aside from that your title “No more Malware” is as trivial as it gets because you’re only stating the obvious — that malware written for Windows doesn’t run on Linux. You pull wool over they eyes of your audience by deliberately misleading them to think that Linux can’t be attacked by malware per definitionem.

    2. Apparently you make good enough money selling your customers anti-malware and anti-virus software for Windows, good enough for not wanting to abandon this.

    3. You misrepresent what I wrote by suggesting that I’ve claimed that prospective employers would show Linux users the door. This is false. This statement was only directed at Mr. Pogson, as I sincerely believe that he couldn’t get hired by a real IT outfit (not: schools!). I know too many Linux users in my vicinity who pose as Linux “specialists”, but in most cases only possess dangerous half-knowledge.

    4. Your suggestion that I’ve claimed that Linux users don’t know how to do things in IT is also false. This statement, too, was solely directed at Mr. Pogson. Google etc. have nothing to do with this. Refer to the last sentence of 3., it also applies here.

    5. No, I think the type of Linux users crawling out of their holes on this very blog are obscure cultists. Another deliberate misrepresentation.

    You want to attack me, yet you can’t even quote correctly what I wrote. You’re a poor FUD artist, dougman.

  38. dougman says:

    Chris Weig, the nagging Troll and arm-chair Geek squad guru speaks his mind!

    He thinks Linux users are a bunch of obscure cultists, however he fails to realize that Linux servers generated $2.4 billion in revenue for vendors in the first quarter of 2012. That’s a 16.0 percent growth rate, which is better than the 1.3 percent revenue growth rate for Windows servers.

    He posits Linux users don’t know how to do things in IT, so when’s the last time you have used Google? How about Facebook? Made trades on the NYSE? Hint: They are all using Linux.

    He dreams that every prospective employer would soon show Linux users the door, as he needs to continue funding his MVP status while pushing M$ drugs. In an period of economic turmoil, he rather demand that business spend money needlessly.

    Sidenote: I gave a short talk recently about Linux to a group of people. The title was “No more Malware”, the amount of interest, questions and feedback I received was good. People do not care what OS they run on anymore. So long as they can browse online, listen to music, create documents/spreadsheets and email that’s all the majority needs.

  39. Ivan says:

    That and you are really funny, like an antisemitic Steven Colbert.

  40. Ivan says:

    If so, why are so many users of that other OS visiting?

    Well, we just want to learn how we can commit atrocities and then atone for them by using Debian, like IBM.

  41. In teaching, many teachers receive annually/semi-annual reviews. I usually score quite well with extra credit for IT. The job’s about teaching. The IT is a useful tool adding to the whole show. I was paid by Easterville extra to design and install the IT system. That’s how good my credentials are. They tendered the job and took the best offer, mine. The project was on time and under budget unlike many other IT disasters we read about with that other OS. The IT system there is the envy of all who see it. It’s ageing a bit but the thin clients are still good and the servers can be replaced by one or two new ones for just a few $thousand to make the system first-rate. It’s still competitive against the usual thick clients running that other OS because of the efficiency of GNU/Linux terminal servers (no network lag for file transfers from servers and file-caching).

  42. The schools where I worked are all over northern Manitoba, Saskatchewan, Alberta and the Arctic. Even with a private plane and shortest route, it would take you about 5000 miles of flying to reach them. I doubt any good would come of it. A few are on roads but that would still take thousands of miles driving over rough roads. Good luck.

  43. Adam King says:

    If he gives you that info I’m sure we’ll be hearing you crowing about how they gave him a negative review and then received a huge infusion of M$ bribe money.

  44. Chris Weig says:

    Uhhh… from 1999 until 2011, I was the “goto guy” in every school where I worked (11 in all) except one and I used GNU/Linux in every school where I worked, so your assertion is false.

    Give me the names of the schools. The next time I’m in Canada (should be around February 2013), I will visit them all.

  45. Chris Weig wrote, “You don’t know how to do things in IT. Every prospective employer would soon show you the door.”

    Uhhh… from 1999 until 2011, I was the “goto guy” in every school where I worked (11 in all) except one and I used GNU/Linux in every school where I worked, so your assertion is false. The fact that I could change jobs so easily reflects my value to employers. The system I set up in Easterville is still purring along at last report. How’s that for “knowing how to do things in IT”, setting up a system which is trouble-free with minimal support over six years (Ubuntu updates and one part-time (15 minutes/day) IT guy)? I rest on my record. The fact that many others schools and organizations with professional IT staff do the same as I do is further support for its validity.

    Of course Chris Weig is entitled to his opinion but it sure doesn’t seem based on any reason except NetApplications’ faulty web-stats and his hatred of GNU/Linux.

  46. Chris Weig says:

    If so, why are so many users of that other OS visiting? They must be interested in GNU/Linux.

    No, but I am interested in obscure cult movements.

    I am reaching folks who want to know about how to do things in IT.

    You don’t know how to do things in IT. Every prospective employer would soon show you the door. Perhaps you could work with Oiaohm in the Australian Bush. Or maybe ask Munich’s Mayor if he needs someone to damage LiMux real good.

  47. MK says:

    Perhaps people are like me, like a good nonsense story.

Leave a Reply