Originally Posted by
Simian Man
Except that the 1% number comes from a reasonably well-balanced study and the 4.8% comes from a site that is geared specifically to developers. Our internal server is accessed by Linux machines 100% of the time; does that mean that everybody uses Linux?
I guarantee you that desktop Linux is well south of 2%.
No the 1% figure is from a source that is very windows centric, it is far from representative of a well balanced cross section. If you know anything about statistical analysis, you would know that the data set needs to be data set needs to be a representative sample of the the whole population. if I took a survey of support for gay marriage in San Francisco, California, it would look much different than if I took it in Salt Lake City, Utah.
The best source for a neutral data set would be the usage logs from google, but they are not available. However, Wikipedia would be a reasonably representative sample of the entire internet, and their data is available.
Their data is as follows:
Code:
Windows 3,233,130 86.53%
Mac OS X 254,100 6.80%
Linux 67,070 1.80%
iPhone 59,828 1.60%
BlackBerry 15,267 0.41%
Symbian 6,974 0.19%
iPad 3,921 0.10%
DoCoMo 498 0.01%
SunOS 375 0.01%
FreeBSD 309 0.01%
OpenBSD 33 0.00%
Total 3,736,217 100%
This however needs to be adjusted for UserAgent spoofing. I don't have a reasonable data set for this, but given that more than 80% of the Linux Users I know spoof their UA string, I would guess that 10% would be a good conservative estimate. This would bring the total market share to about 2% all -users.
Bookmarks