I don't know if this is the best-written article around. First of all, the author is using a very small sample size of operating systems. Namely, his household. I'm running two systems with Linux, one OS X, and a dool boot Linux/Windows. So I can come up with similar results, as I'm sure almost any member of this forum can. We are not the norm. I'm the webmaster at the company I work for, and I routinely look through webstats for compatability reasons. 90% Windows, 5% Mac, and 5% Linux is pretty accurate for my company's users.
Secondly, he seems to be blurring the line between desktop systems and server systems. Only 8.6% of servers are running Linux? I wouldn't doubt that. Most of them are still using Unix.
He does raise an interesting point, however, in how machines are sold. I don't think there exists an accurate account of how many machines are sold with Windows pre-installed that is later (or immediately) wiped out in favor of Linux. Though operating systems sold is doubtfully the measure of how many users are using what. I alluded earlier to checking end-users' preferences through web use. This seems more accurate.
All in all, a pretty interesting article. Not very scientific, though.
Last edited by fragglehorn; 04-08-2004 at 10:42 PM.