As my colleague Alex Wolfe noted, Linux hasn't made a dent in the desktop after years in the wild. The climb looks all the steeper now that Windows 7 and new versions of Mac OS X have arrived. I can think of a few other reasons why Linux hasn't achieved more than a fractional marketshare with end users, and they aren't pretty. (I've already donned my asbestos suit.)

Linux's best features are not things that matters to end users. More than anything else, I think, this it. The things that make Linux so, well, Linux-y -- its malleable kernel, its open source development -- don't automatically present anything to an end user that makes them sit up and beg.

Here and there a couple of people are realizing this and working hard to make Linux the start of something -- ChromeOS, Moblin, Android, etc. -- but what's come to the public so far are adjuncts to the desktop and not replacements for it. (And if they're good adjuncts, who's complaining?)

Most people already have what they need. Given a choice between the well-known and broadly-used Windows and Linux, most people will still choose Windows, no contest. Linux will be a choice amongst people who consider the choice in the first place, but for most people it's not worth the effort of thought. They have work to do, the computer is nothing more than a way to get there, and Windows already offers them a solution.

This is also why the "no viruses" line doesn't draw people to Linux in droves: sure, you can have all that, but at the (not widely advertised) cost of not running any of the programs you've become familiar with. And following that up with lines like "you can can still run your Windows apps in Wine" further shows how removed the die-hard advocates are from the reality of the average person's use of a computer.

It's harder to get this stuff right then it might seem. Deploying a general-purpose OS on a panopoly of system types is tough. Microsoft still has trouble with it, although they get it more right than wrong at this point. One of the reasons Apple sells OS X with its own hardware (apart from the obvious profitability of such a scheme) is because it cuts down on the number of variables.

But Linux's road is still rocky. A widely-reported poll on Ubuntu's own support forums shows that out of people who either installed clean or upgraded from a previous version of Ubuntu, a whopping 40% had problems they categorized as "unfixable". Even if you put aside the biased nature of such self-reporting, that's still bad news for Canonical. Even when Windows 7 wasn't in beta, it didn't regress that badly.

Free should not be the only feature. Windows comes "free" with most PCs anyway, or is nominally considered to be part of the cost of running the computer -- a peripheral of sorts. Linux is still struggling to demonstrate that it has some attraction apart from its lack of a pricetag. The things it has in strength are not consumer-oriented features.

Also, most people, especially most businesspeople, will pay for a program of demonstrable value. Free is handy if you want to get started with something for no initial outlay, but you can't bank on it as the strategy to keep people committed.

I'm not going to argue that nobody is switching to Linux. I'm also not arguing that Linux has no place in the world, period; that's lunacy. I am arguing that by and large Linux is not shaping up to be the Windows-killer that's been promised for years now.

It's also high time to stop expecting it to be. Obsessing about using Linux to unseat Windows from the top-dog slot is only going to inject the wrong set of motives into the development of Linux. I fear it might already have.

InformationWeek Analytics outlines the 10 questions you need to ask to see where netbooks fit within your organization. Download the report here (registration required).

Twitter: Me | InformationWeek
Facebook: InformationWeek