Has anybody checked out this article about connections to the BBC web site?
[currybet.net/articles/user_agents/2.php]

Linux is less than one half of one percent of the BBC connections. Can this be true?

Assuming it is true...

Is the BBC site an unfair sample of Linux popularity?
Does the site design discourage Linux?
I was unable to listen to music. (Put very little effort into it, but Shoutcast works just fine for me with no effort at all.)

Probably I have missed something obvious. Surely Linux is more popular than .4%

Here are some numbers from IDC-Australia...
[arnnet.com.au/index.php/id;1701652176;fp;2;fpid;1]
- 1.1 per cent of SMEs planned to use the operating system in the next whitebox PC purchase.
- 1.7 per cent of consumers said they used Linux on their branded notebook PC.

The 1.1% seems to be in synch with the BBC, as the BBC .4% is for all users and the IDC 1.1% is for new users. The 1.7% doesn't seem to fit here, but might fit with the IDC number below.

Sources often quote the 3% that IDC reported for desktops in January 2005.
[blogs.zdnet.com/ITFacts/?p=285]

IDC's 3% is "based on unit sales". "Unit sales" and use/popularity may be two entirely different things.

I am having trouble finding objective measurements of current desktop-Linux popularity (relative to other OSes). Counter.li.org estimates are not helpful.
Anyone have a link?

[Since I don't post much, and therefore received a message that this post will be blocked because it contained a URL, I have included only a portion of the links.]

So?

Thanks for your response, ctkroeker.

Sorry that my post was hard to understand. I edited it twice in response to messages from LinuxQuestions and lost the flow.

The guts of my questions are really simple...

It is surprising that such a low percentage of BBC website connections used Linux. I wonder why.

My own research results about Linux use/popularity are confusing. I hope that someone can help me find current objective information about the use of Linux.

Perhaps since 95% of desktops come preinstalled with Windows is the reason. I don't know many users or administrators that do their daily browsing from servers that are installed with Linux or the like..

The problem with getting accurate percentages of OS use is impossible. Most can claim the 95% of desktops come with Windows preinstalled but who knows how many wipe it to put their own OS of choice on it or dual boot it. Most Linux distributions are available for free so unless all the people in the world and took a survey of what OS is on what or how many machines they own/use, the actual %'s are always going to be guestimates and not real numbers. There's going to be that margin of error.

I believe that the report from the site is based on what the browser is reporting itself as. It could well be that Linux users aren't looking at the BBC website or that we are simply tweaking our browsers (Konqueror via the settings, Firefox via a plugin) to report themeselves as "Internet Explorer 6.0 on Windows XP".

Do browser stats really matter that much? I don't think so.

trickykid & XavierP

I see now that meaningful stats would be difficult.

Using my family as an example..
25% of our computers came with Linux installed.
50% use only Linux.
75% run Linux as the primary OS.
95% of use time is Linux.
100% of Internet use is Linux.

So, the use/popularity/market-share of Linux at my house could be 25%, 50%, 75%, 95% or 100%.

I have not tweaked my browser as it works fine with the great majority of sites. If there is a site that doesn't work well with stock standard Firefox, I avoid the site.

BBC gave me problems and Shoutcast did not. Depending on other factors that won't be explored here, BBC might be expected to report a lower percentage of Linux users than Shoutcast.

If many Linux users tweak their browsers to report themselves as IE/Windows (e.g. if that tweak allows them to use the BBC site), and leave them permanently tweaked, then internet measurements of OS use would not only be biased, but false.

So, measuring popularity is certainly difficult. I understand how some measurements would tend to understate Linux users.

Thanks for the thought-provoking responses.

PS. trickykid, thanks for the Intoxicated Shadows link. That's a cool site.