That’s pretty accurate, yeah. The Linux crowd has an alarming fetish for irrelevancy. The rest of the computing world is interested in what their computers and software can do. Can you imagine if a bunch of typical Windows users got into a debate about which updates they applied? Or if that’s too abstract for some of our slower readers, imagine if a lumberjack went on and on about the special customized handle and decals on his chainsaw.

And it’s not just number of users. The fourth sentence on Debian’s website talks about the number of packages they support. Linux users are notorious for uptime talk. Meanwhile Windows and Mac users generally have no clue how long their system has been up, not that it’d be a genuine figure since most of them with large uptimes will hibernate their systems when not in use. And if it’s not uptimes it’s kernel versions. Think the average Mac user knows what his patch level is or is even aware that they exist other than part of the update blob that bounces in their dock every few weeks?

To be fair, you kinda have to know your kernel version in Linux land, at least these days, though prior to 2.6 it was like MacOS where you could say, “I run 2.4″ and people would know what you were talking about. Now it’s, “I run 2.6.33.4-git4-am-patch-4-ubuntu9″ only to be met with, “You’re still running that? 2.6.33.4-git4-am-patch-4-ubuntu9 is so mid-April 2010! Get with the times!”

Linux reaching (x) amount of users will not make it a better OS. If numbers were what made an OS great, then Windows would indeed be the best OS on Earth. But no, that’s unthinkable to the Linux crowd.

Of course, Linux *would* reach critical mass if it truly were such a great OS. And of course, the Linux crowd are arse-about-face; joining groups, writing positive reviews and handing out endless CDs won’t make Linux a great OS. It will only be great once its fan- and developer-base stop assuming it already is. And only then will the numbers follow.

OS Share in 1997:
Windows 95/98 69.4%, Windows NT Workstation 9.2%, Windows 3.x (with DOS) 7.7%, Mac OS 4.6%, Linux 2.4%, DOS (IBM, Digital Research, Microsoft) 2.3%, Unix 1.0%, OS/2 Warp 0.8%, other 2.7%
http://pctimeline.info/windows/win1997.htm

Germany has always had a higher than average Linux base but even they are down in past few years.
http://gs.statcounter.com/#os-DE-monthly-200807-201004

From what I have read Linux desktop share peaked in 1998 and has fallen by about half since then.

As for Lunduke he is not only trying to herd cats but also take them fox hunting. It will be funny to see his Linux sucks video next year when he has lost even more patience with them.

Thanks for expounding. The worst part about all of it is that every information gathering method uses the same faulty self-selected surveying. The German magazine is the strongest, though 260 responses is hardly statistically significant and would only tell the Linux base within the particular readership of that magazine. Furthermore, if the question was really posed as “Who uses Linux at home?” the data is further diluted since you’ll get people with tinker boxes and router firmware. Do owners of Linux WRT routers really qualify as “Linux users”? It’s like counting people who use ATMs as OS/2 users.

Finally, after sleeping on it, I recall registering at some counter site way back in the 90s. This is probably it. Am I still counted? I haven’t used Linux as a desktop system in years. What happens to people like me? The (hard to find) FAQ doesn’t really say:

http://counter.li.org/help/faq.php#person4

Rules have been defined what should be done with registrations which have not been verified for a certain amount of time.

What does that mean? Are they deleted or not? What’s the aging breakdown of the entire data set? What’s the confidence? Where’s any documentation whatsoever on the survey aggregation? This is quite important for a site that deals solely with data gathering! Instead we get useless graphs on CPUs, kernel versions, and uptime. Who the hell cares?

http://counter.li.org/help/faq.php#person1

The Great Deletion of November 2001 removed 90.000 entries.

Yet early 2002 saw asymptotic growth. What’s this about? Given the above info, it looks suspiciously like a compensation factor, as if the user base were artificially inflated to demonstrate growth.

http://counter.li.org/help/faq.php#common1

We are an HTML purists.

Yet the site uses HTML 4 Transitional, which is a compromise between “pure” HTML and Netscape-corrupted style elements, of which they take advantage by using tables for layout and the body tag for style. HTML purists indeed! More like, “We learned Netscape style web design in 1995, never bothered to keep up, and defend our stance by pretending our way is correct.”

This may seem inconsequential to some, but it fits in with the rest of the faulty assumptions the site makes. If they were really purists, they’d either upgrade the site to XHTML Strict and banish all style elements from the HTML files or drop back to ITEF HTML 2, which totally lacks styling with the exception of bold and italics*, which were later deprecated in favor of the realistically equivalent but more structure oriented strong and emphasis tags.

* I guess horizontal rule and preformatted text count as styling, though they’re pretty borderline compared to using tables to layout text.

lol’d.

Funny thing is, that couch is probably as comfortable to a humas as using ubuntu is.