Every year a new crop of articles appear on the subject of the time when Linux, or another alternative operating system, will become a suitable competitor to Microsoft Windows. The problem is no group of people have gotten together to establish what the term suitable means, or what number should be affixed to the determination.

An article over on Geeks are Sexy speaks of an idea that had never really occurred to me, although parts of it I had contemplated many times. The article posits that Linux is already more used, and in the mainstream than we know, basically because the places doing the reporting are funded partially through the ads of Microsoft, and it simply would not look nice to say that a competitor that Microsoft had slammed so many times, and threatened with litigation, was actually larger in force than the powers that be in Redmond were willing to admit.

Caitlyn Martin at O’Reilly, examines the idea that Linux only has a 1% market share on desktops (including laptops), and finds that it’s lacking. This 1% number comes primarily from usage shares detected by web browsers.

I know that this is what is used, but why is simply inane, as there are many factors, some mentioned below, some not, which make that wrong, or wildly inaccurate. At the levels being reported, you would think that no  one ever used Linux, or any of the BSDs, or Solaris, for anything other than exploration. I know a few people who use a *ix system as a house file server, or gateway, so no accounting of a browser running on it will ever occur. Now, I’m not saying this grossly distorts the figures, but it should be thought about.

However, Martin says that the 1% number simply doesn’t make sense. For example, Linux had 32% of the netbook market in 2009, despite the fact that retail stores tended to only sell Windows notebooks. A third of Dell’s Netbook sales were preloaded with Ubuntu. And if you take the idea that netbooks were around 18% of desktop/laptop sales for 2009, you end up with the conclusion that around 6% of all computers were sold preloaded with Linux. This number doesn’t include people who bought a windows machine and then reformatted and installed Linux on it. This leads Martin to conclude that some observation bias is likely, as certain web sites (like Ars Technica) may attract certain operating system users.

Additionally, it’s possible that Linux users often use multiple operating systems, essentially “splitting their vote.”

This is the one that is so wrong, as many people I know do just this. It makes the voting come out in favor of Windows, of course, but it shows more than many would like to admit are using, or trying to use, Linux or one of the other choices.

Best estimates, according to Martin, is that Linux has a share roughly equal to that of MacOSX; which is certainly not a slouch on the desktop/laptop market.

Throughout the 1990s and 2000s, people decried the fact that Linux wasn’t mainstream – it’s clear that today, it certainly is. A minority, yes, but a mainstream minority – Linux is not in the same category as, say, IBM AIX. So if you wanted to know “when Linux would be mainstream on the desktop,” the answer is probably “around 2009.”

What are your thoughts?

It would be nice to get a real accounting, but it is very difficult. Part of the problem is that since there is no cost, and no restrictions on the use, it is very hard to track.


What happens if a big asteroid hits Earth? Judging from realistic simulations involving a sledge hammer and a common laboratory frog, we can assume it will be pretty bad.

Dave Barry

Opera, the fastest and most secure web browser