hi for any body who is intrested i hav made a survay about the popularity of diffrent OS's on google click on the link below to reed the full artical

http://berrancenetworks.co.uk/lstuff/lnx-vs-os-ggl.php

also for any body wjo has writen an artical or review about linux and would like it to be published on a website for free you can either PM on this forum ore he one on my website me or send an E-mail to me at berrance@berrance.com, i would preffer e-mail as i dont come on hear a lot. any reviews will be published as they are on topic and have good content. if you do send me a review please dont forget to put your name on it.

thanks

berrance

I took a read of the article. However, much as I love Linux (I am an avid supporter of Linux and use it as my desktop OS), I have to unfortunately point out that your research is flawed.

I notice that for some of the items you omit the spaces, such as "Windows95", "Windows98" and "Mac OSX", which you say have 1,100,000; 862,000; and 951,000 results respectively. However, the full and proper name for each of these includes spaces - therefore, for "Windows 95", "Windows 98" and "Mac OS X", the correct results are 7,560,000; 6,890,000; and 7,680,000 results respectively as of this time of writing. I notice that you mention "WindowsNT" in the article but I'm assuming that you searched for "Windows NT" with the space, as the number seems high enough (a search for "WindowsNT" here produces 515,000 results).

Another problem is the way the search for Linux only included its generic name. It's no surprise you got such a large number of results for it as lots of people refer to their favourite distribution of GNU/Linux as just "Linux" (as I did in my first paragraph, for example ). Similarly, lots of people refer to all versions of Windows as just "Windows" (for example, "It's a Windows program."). Of course, the trouble with this is twofold; firstly, it's an overlapping search (it would also find the "Windows" in "Windows XP", etc, so you'd need to remove those results from the search), and secondly, it's also a generic name for the things through the sun is probably shining through into the room you're in at the moment. Because of the latter problem, any search for "Windows" by itself will yield false positives, but because of the former, any search excluding "Windows" by itself will yield false negatives.

Of course, "Linux" has no such problems - it's not a generic term, so any use of it is almost certain to refer to the kernel - and so you end up with a much higher tally of results.

You also say in the article:


No doubt, because you'd be falling into the trap of using an overlapping search. Searching for "Linux" on its own automatically includes all the various distros that mention "Linux" in the name, such as "Mandrake Linux", "Debian Linux", "Red Hat Linux", etc. By including those searches again you'd be artificially inflating the result. A better approach would be to include distros that don't include the word "Linux" again, such as The Fedora Project, Knoppix, etc.

That said, you made a good effort, and I do agree that Linux is gaining popularity (as well it should! ). But all the above being said, counting results on a search engine is still not a very scientific way of measuring usage, or popularity.