There are three reasons why Linux isn't succeeding on the desktop, and none of them are to do with missing functionality, using the command line or the politics of free software.
The first is that there's too much momentum behind Microsoft Windows and too many preconceptions about the alternatives. Linux is perceived as having too much of a learning curve for relatively few advantages and an unknown heritage.
Migrating big business to a Linux desktop is akin to turning a T1-class supertanker around mid-Atlantic. The opposite direction may look brighter, but it's easier to chug onwards into the storm. You only have to look at the number of people clinging to Microsoft's venerable Office suite to see this point clearly.
For the vast majority, most of its functional fecundity is wasted. Many people could arguably be just as (un)productive with Notepad, Calculator and Paint, let alone using an open-source alternative such as OpenOffice.org. Its use seems to have more to do with keeping face when attaching files to an email than a genuine operational advantage.
Most people will only consider an alternative when there are bigger issues, larger icebergs or uncertain territories on the horizon, Away from the desktop, Linux is faring better.
Smaller, more agile businesses quickly quantify the cost advantages to produce cheaper and more competitive products. This is why embedded Linux has been such a success on everything from Chinese mobile phones to almost every NAS box around. This may mean that success on the desktop is only a matter of time, or it may mean that the Linux desktop is too far removed from the Linux kernel.
The second reason for failure is that Linux lacks centralised marketing. This is because there's no real Linux Central. It's just a trademark owned by its creator, Linus, and a term normally reserved for just the kernel of the operating system – hardly the easiest product to sell.
There are plenty of people advertising their own Linux endeavours, all keen to push their own angle on its advantages. This divided effort compounds the problem. With the likes of Red Hat, Novel and Canonical all fighting for their own slice of the pie, there's no one left to push Linux as a distinctive brand. That's something Apple and Microsoft do extremely well, and something Linux leaves to Tux the penguin.
Many would argue that standards are the answer to this conundrum, and that would mean a single base distribution. This could then be the only distribution called 'Linux' - everything else would become 'Linux based'.
Mozilla manages this well with the use of the Firefox brand. It's freely distributable and modifiable, but it can only be called 'Firefox' in its untouched incarnation. Change anything and you need to change the name.
For example, Debian calls its Firefox build 'IceMonkey' because it needs to reserve the right to make modifications, thus breaking Mozilla's standards. This may cause confusion if you look for Firefox on your Debian desktop, but it also sets a precedent for the kind of experience that Mozilla expects its users to have, and Debian hackers still have the code to mess around with if they need to. It's a compromise, but it might work in a world with hundreds of Linux distros.
The third reason is easy to see but harder to solve. It's the reason why you're not using Linux now. The solution would make all other problems redundant. The reason why you're not using Linux now is because there isn't a good enough reason to.
Sober advantages such as better security, improved performance, rock solid stability and low cost aren't going to win converts. These advantages aren't exciting enough; they're the equivalent of a spreadsheet of mortgage repayments. What we really want is a significant upgrade, something you'd normally pay for.
Perhaps we should focus on value. Recent analysis of the kernel by Jon Corbet showed that 75 per cent of the 2.8 million lines of code in recent contributions were written by paid-for developers. That puts Linux freedom in context.
But the biggest challenge is sexiness. There's very little of it in Linux unless you're an antisocial geek, and products like the Apple's iPad illustrate this massive divide painfully. As Jim Zemlin, Executive Director of the Linux Foundation, puts it, "Linux can compete with the iPad on price, but where's the magic?"
Linux has the programmers, the managers, the community, the innovation, the time and the skill. But to succeed in 2010 and the coming decade, what it really needs is a magician or two.
Article continues below