R
Rex Ballard
On Jul 9, 1:33 am, "Moshe Goldfarb." <brick_n_st...@gmail.com> wrote:
> Linux has been around for 15+ years.
> Linux is free.
> Linux applications are free.
>
> So, the question is why is Linux still sitting at below 1 percent of the
> desktop market?
>
> Something that is free is used by so little of the population?
> People would rather spend big bucks for Vista or even more expensive switch
> to Mac, like a lot are doing, rather than run free Linux?
>
> Something is seriously wrong with Linux....
>
> Here is yet ANOTHER article, offering the same, or similar numbers that
> other sites have been reporting, including the BBC....
>
> http://blogs.zdnet.com/hardware/?p=2179#comments
This benchmark is well-known and notorious for it's unreliability.
Every aspect of that survey is skewed to ignore as many Linux browsers
as possible, and count as many IP addresses as Windows as possible.
I did find the sharp uptick in the curve very interesting. It would
appear that even with all of the attempts to try and suppress Linux
counts as much as possible, the counts are going up.
Flaw number 1 - the survey only counts IP addresses, each IP address
could be a single DHCP address that is assigned for only a few hours,
or a corporate NAT address that represents 450,000 users.
Flaw number 2 - only one OS per IP Address - survey only counts one IP
address, and if a DHCP address assigned to a Linux PC is later
reassigned to a Windows PC, the Windows PC overrides the Linux entry.
Flaw number 3 - Browser Signature Parsing - Survey only looks at the
first entry of the platform section of the signature. On Windows, it
will be Windows and the version number, on Mac, it will be Mac and the
OS version number. On Linux it will usually be X11, the Linux entry
is typically the 3rd or 4th entry in the signature.
Flaw number 4 - Sites need to use IIS - content is skewed in favor of
Microsoft.
Flaw number 5 - MSN, AOL, AT&T, and Verizon are "linux hostile" DHCP
services which have huge IP pools. Go to http://ipchicken.com/ and
look at your IP address, then look at ipconfig, if the IP addresses
match, you have a public DHCP address, and you could dial-in 3-4 times/
day and be counted 90-100 times per month. If there isn't a match,
you have a NAT address, which you could be sharing with up to 16
million other users (if your ipconfig address starts with 10) or just
65,000 users (if you ipconfig address starts with 192.168).
Linux users have started using USB "cellular modems" which actually
work quite well, however, they have the ability to keep it from
"hanging up" automatically. When the same modem is used with Windows,
it will time out after being idle for about 10 minutes. When using
the device with Windows, I went through 12 IP addresses in 2 hours.
When using the same device with Linux, I only went through 2 (I moved
to a different location, so I had to unplug the USB device).
Here's a few other surveys
This one shows Linux at 5%
http://www.ews.uiuc.edu/bstats/latest.html#BrowserPlatforms
Browser Platforms %
--------------------------------------------
1. Windows 2K (NT5) 57.4
2. Windows NT 8.7
3. X11 5.1
4. Windows 2.9
5. Macintosh 2.8
6. other 23.1
Keep in mind that the "other" catagory is 90% Linux as well.
Many Linux users use squid to cache content for users, so squid gets
counted instead of the real browser.
Linux at 3% (lots of NAT issues in this one).
http://www.w3schools.com/browsers/browsers_os.asp
Here's a good article on the problems with Browser Stats
http://www.upsdell.com/BrowserNews/stat.htm
There are much better surveys, which are kept and published by
companies like Google and Yahoo, that keep trick of registered users
and count all platforms used by that user. But those surveys (which
show Linux with MUCH higher numbers) are expensive and protected by
nondisclosure agreements, which is why you won't see them quoted
here. IDC has given numbers for Linux as high as 17% before they
stopped publishing those numbers after 2002. Other metrics have shown
Linux client share growth doubling every 12-18 months. If you do the
math, that would mean that Linux users would have grown to about
30-35% of the market (assuming that Windows growth was 20%/year and
Mac growth doubling each year since the release of OS/X and slightly
higher since the release of Vista.
> Moshe Goldfarb
> Collector of soaps from around the globe.
Excellennt quilification to be a WinTroll.
> Please visit The Hall of Linux Idiots:http://linuxidiots.blogspot.com/
Any on that list should be honored, since only the best advocates are
listed :-D
Rex Ballard
> Linux has been around for 15+ years.
> Linux is free.
> Linux applications are free.
>
> So, the question is why is Linux still sitting at below 1 percent of the
> desktop market?
>
> Something that is free is used by so little of the population?
> People would rather spend big bucks for Vista or even more expensive switch
> to Mac, like a lot are doing, rather than run free Linux?
>
> Something is seriously wrong with Linux....
>
> Here is yet ANOTHER article, offering the same, or similar numbers that
> other sites have been reporting, including the BBC....
>
> http://blogs.zdnet.com/hardware/?p=2179#comments
This benchmark is well-known and notorious for it's unreliability.
Every aspect of that survey is skewed to ignore as many Linux browsers
as possible, and count as many IP addresses as Windows as possible.
I did find the sharp uptick in the curve very interesting. It would
appear that even with all of the attempts to try and suppress Linux
counts as much as possible, the counts are going up.
Flaw number 1 - the survey only counts IP addresses, each IP address
could be a single DHCP address that is assigned for only a few hours,
or a corporate NAT address that represents 450,000 users.
Flaw number 2 - only one OS per IP Address - survey only counts one IP
address, and if a DHCP address assigned to a Linux PC is later
reassigned to a Windows PC, the Windows PC overrides the Linux entry.
Flaw number 3 - Browser Signature Parsing - Survey only looks at the
first entry of the platform section of the signature. On Windows, it
will be Windows and the version number, on Mac, it will be Mac and the
OS version number. On Linux it will usually be X11, the Linux entry
is typically the 3rd or 4th entry in the signature.
Flaw number 4 - Sites need to use IIS - content is skewed in favor of
Microsoft.
Flaw number 5 - MSN, AOL, AT&T, and Verizon are "linux hostile" DHCP
services which have huge IP pools. Go to http://ipchicken.com/ and
look at your IP address, then look at ipconfig, if the IP addresses
match, you have a public DHCP address, and you could dial-in 3-4 times/
day and be counted 90-100 times per month. If there isn't a match,
you have a NAT address, which you could be sharing with up to 16
million other users (if your ipconfig address starts with 10) or just
65,000 users (if you ipconfig address starts with 192.168).
Linux users have started using USB "cellular modems" which actually
work quite well, however, they have the ability to keep it from
"hanging up" automatically. When the same modem is used with Windows,
it will time out after being idle for about 10 minutes. When using
the device with Windows, I went through 12 IP addresses in 2 hours.
When using the same device with Linux, I only went through 2 (I moved
to a different location, so I had to unplug the USB device).
Here's a few other surveys
This one shows Linux at 5%
http://www.ews.uiuc.edu/bstats/latest.html#BrowserPlatforms
Browser Platforms %
--------------------------------------------
1. Windows 2K (NT5) 57.4
2. Windows NT 8.7
3. X11 5.1
4. Windows 2.9
5. Macintosh 2.8
6. other 23.1
Keep in mind that the "other" catagory is 90% Linux as well.
Many Linux users use squid to cache content for users, so squid gets
counted instead of the real browser.
Linux at 3% (lots of NAT issues in this one).
http://www.w3schools.com/browsers/browsers_os.asp
Here's a good article on the problems with Browser Stats
http://www.upsdell.com/BrowserNews/stat.htm
There are much better surveys, which are kept and published by
companies like Google and Yahoo, that keep trick of registered users
and count all platforms used by that user. But those surveys (which
show Linux with MUCH higher numbers) are expensive and protected by
nondisclosure agreements, which is why you won't see them quoted
here. IDC has given numbers for Linux as high as 17% before they
stopped publishing those numbers after 2002. Other metrics have shown
Linux client share growth doubling every 12-18 months. If you do the
math, that would mean that Linux users would have grown to about
30-35% of the market (assuming that Windows growth was 20%/year and
Mac growth doubling each year since the release of OS/X and slightly
higher since the release of Vista.
> Moshe Goldfarb
> Collector of soaps from around the globe.
Excellennt quilification to be a WinTroll.
> Please visit The Hall of Linux Idiots:http://linuxidiots.blogspot.com/
Any on that list should be honored, since only the best advocates are
listed :-D
Rex Ballard