It's a long-standing joke in the free software world that this will be the year when we see GNU/Linux make its breakthrough on the desktop - just like last year, and the year before that. What's really funny is that all the key GNU/Linux desktop apps are already being widely deployed, but not in the way that people have long assumed.

The indestructible optimism about GNU/Linux appearing on the desktop seems to be the result of a misguided view that since it grew from insignificance in the server sector to become a serious challenger to Microsoft Windows there, the same is bound to happen to GNU/Linux on the client side.

But there are fundamental differences between the server and desktop markets. Free software has traditionally been written by geeks for geeks: this means that it suits well those whose job is tending a company's servers. By contrast, all the things that non-technical users demand – ease-of-use, an engaging user interface and above all continuity with what they are used to – have typically been absent.

Similarly, although it is difficult enough to impose a new server solution on an unwilling IT department, it is trivially easy compared to trying to get end-users to switch to a new desktop. After all, IT staff are paid to deal with IT problems (including their own), whereas general users just want to get their non-IT jobs done; a completely new IT solution is seen as an unnecessary distraction from their “real