It's easy to fall into a "Linux vs. Windows" mindset, as if operating system choice is a zero-sum game, and at the end the victor gets a trophy while the loser leaves the ring with slumping shoulders. I'm about as firm a Linux advocate as you'll ever find, but I don't really care if Linux dominates the world.
What I want is a stable, low-cost operating system I can use for normal home and office computing tasks. I happen to like Linux better than other operating systems I've tried. I know and like many people who won't own anything but a Mac. Some of my friends love Windows. I have other friends who swear by FreeBSD.
Linux users are in the minority today, and probably will be for many years to come, possibly even forever. Long before Linux can possibly "take over" from Windows as the world's dominant OS, chances are something completely new and different will come along and make all of today's operating systems -- possibly even the concept of operating systems -- obsolete.
So let's stop with the "world domination" talk. I don't personally care whether Linux r00lz the world or not. All I want, and probably all you want when you come right down to it, is a Linux-friendly world. Or, if you're inclined that way, a Mac-friendly world. Or a BSD-friendly world.
What I think what we all really want is a friendly world, where interoperability is the rule, not the exception.
I want to be able to use any operating system and standards-compliant browser to do things like online banking. I want to be able to listen to music and watch video over the Internet without worrying about whether this format or that one is compatible with this operating system or that one. And I want to be able to view any reasonably mainstream Web site without having to worry about whether my computer is compatible with some sort of strange server extension.
I positively love the idea of XML; that we can have a single, universal file format we can all read and manipulate in a wide variety of text editors and word processors, no matter what operating system we choose, and that we can share XML files with others whose operating system and software choices aren't the same as ours.
Sometimes I have a vision of an automobile industry where Ford cars can only drive correctly on Ford roads, and there is a universal road scheme, controlled by Ford, that requires every driver, no matter what make of automobile he or she owns, to pay Ford a small sum for every ride on the .HWY system. Perhaps Ford wouldn't charge individual drivers directly, but would require businesses that sell to those drivers to cough up fees for their right to sell goods and services to .HWY users. Those fees wouldn't need to be huge. I'm sure Ford managers and shareholders would be happy to get five cents per gallon of gas pumped into every car in the world. Perhaps Ford would even give a two-cent break to refiners who agree to sell nothing but "Ford-optimized" gasoline.
This is a frightening fantasy. I'm glad it could never happen in real life.
I don't really care what kind of car you want to drive. As long as I can drive my Jeep Cherokee on the public roads and get gas and tires it can use almost anywhere, why should it matter to me that you don't like Jeeps and would rather have a Subaru or Hyundai? Or a Mercedes or a Honda or an Austin Cooper?
My real, secret reason for wanting Linux to become more popular is selfish: I want to be able to buy cheap scanners and other peripherals I can plug into my Linux computer and run without worry, and the only way hardware manufacturers are going to make most of their most popular products Linux-compatible is if enough people use Linux that it's worth their while.
The same applies to software and Internet-based services. Right now software authors and online providers can say, "There is no good reason to support Linux. Hardly anyone uses it." I would like to see Linux have enough market share that it is not financially prudent for large companies to ignore Linux users.
What will that take? 5% of the user-level OS market? 10%?
I've seen statistics -- the same ones you've seen -- that put Linux desktop or "end user" market share everywhere between 0.24% and 5%. The truth is probably somewhere in the middle, which means we have a good way to go before Linux is a significant factor in the user-level OS marketplace.
But once Linux becomes popular enough among "ordinary computer users" to command attention and respect from hardware manufacturers, software publishers, and online service providers, does it really matter how much it spreads beyond that point?