User:KimDabelsteinPetersen/LinuxWeight

Lets use search to get us a reasonable overview over how often Linux is used versus GNU/Linux. To do this we have to ensure that we remove bias where the kernel is involved, so we remove the keyword "kernel", as an afterthought i've removed the keyword "driver" as well. We search all documents that include the word "linux", and either contains the wording "GNU/Linux" or doesn't contain GNU/Linux anywhere in the document.

For science papers - we only check computer science and engineering, and for literary searches we check with Amazon only on specific publishers, with a high degree of technical books.

A search on websites looks like this for +GNU/Linux and like this for -GNU/Linux. A search on Amazon looks like this for +GNU/Linux and like this for -GNU/Linux. A search on Scholar looks like this for +GNU/Linux and like this for -GNU/Linux

Of course this should only be used as a guideline, and to get a feel for how often GNU/Linux is used vs. Linux, when we aren't talking about the kernel. (ie. the OS). Feel free to suggest further improvements for bias removal, i tried to be simple but reasonably accurate.

Caveat: for websites there can be a significant bias by support letters or sub-sites.

Results

edit
Search key stype +GNU/Linux -GNU/Linux Percentage
Science papers Scholar 2,240 111,000 2.01%
Sun.com Website 13,800 410,000 3.37%
HP.com Website 1,860 369,000 0.50%
IBM.com Website 863 750,000 0.12%
(all google) Website 1,140,000 73,300,000 1.53%
Prentice Hall Amazon 254 8,660 2.93%
Addison Wesley Amazon 148 6,840 2.05%
O'Reilly Amazon 329 13,200 2.49%