Linux has made major inroads on servers and in data centres running both open-source and proprietary applications on millions of computers worldwide. We've recently seen the rise of Linux on mobile devices. But the Linux desktop remains elusive. We know it's out there, but it only now seems to be approaching the tipping point.

Where is desktop Linux at the moment? Right now we see end-user Linux in fixed function and transactional systems and technical workstations. Transactional systems tend to be next-generation Unix replacements that connect to Linux and mainframe back-end systems. Technical workstations are largely in the developer arena.

Over the past 18 months, we have seen growth in desktop Linux usage, spurred by new more user-friendly distros like Ubuntu and new versions of Firefox and OpenOffice. But the holy grail -- business users -- are still few and far between. At what point will ISVs port applications to Linux, and is there enough demand?

Market interest is there. If the results of Open Source Development Lab's recent desktop Linux survey are any indication, the end-user market may be more ready for Linux on the desktop than most industry watchers were expecting. In fact, employee demand and matching competitors' usage were cited as the top two reasons for deploying Linux on the desktop, disproving conventional wisdom that desktop Linux is chosen for security and low cost.

To a large extent, the value of Linux on the desktop and associated open-source desktop applications is to limit Microsoft lock-in. It's in the technology world's best interest for there to be an alternative to Microsoft's monopoly on desktops. Linux is the obvious choice for its openness, security, flexibility and low cost. But there are still technical and functional barriers that must be addressed before desktop Linux can go mainstream.

Getting comfortable with Linux

Linux tends to be perceived as Unix's geeky, more in-your-face younger brother. Bolder and brasher, but the user interface is still immature compared to Mac OS or Windows.

One of the biggest desktop Linux inconveniences is the lack of support for existing proprietary applications from big vendors such as Adobe, Autodesk and Intuit. The same can be said for the lack of drivers for plug-and-play functionality related to Wi-Fi, PDAs and digital cameras. These may all be a burden right now, but they are solvable problems.

While it's highly unlikely that we'll see Microsoft applications running on Linux in the near future, credible replacements for applications that are Windows-oriented are needed.

When the OSDL last month gathered more than 70 Linux desktop architects, including developers from Intel, Mozilla, Novell and X.org, the group launched the Portland Project with the goal to make it less complicated for ISVs to run their applications on multiple Linux desktop environments. Much like we've seen with ISVs porting applications to Linux on the server side, so we will see on the desktop, now that there is a unified effort to tackle the inhibitors -- and to do so as a community.

Building on the momentum of Firefox

Firefox is the first mainstream open-source desktop application to reach critical mass. This proved a number of things to the market: Open-source applications not only compete with, but are often better than, status-quo proprietary applications; open-source community development works; and an application distributed primarily over the Internet, not bundled with an operating system, can achieve significant market share.

In conjunction with solving the technical problems related to things like device drivers and networking, the next step to garner mainstream business users is development of a complete office suite. In the best of worlds, this suite will be enterprise-ready, multiplatform, open-source and entirely free. It's at that stage that not only do open-source applications take over the desktop, but Linux on the desktop changes the economics in the favour of the user.

Will 2006 be the year of the Linux desktop? Maybe. Until Linux desktop applications meet business-user requirements and functionality is truly on par or better than proprietary apps, market share will tick only slightly upward as adoption continues in the same technical and fixed-function market segments. It will take a unified community effort to address the issues and educate ISVs into making desktop Linux a reality for business users worldwide. The Portland Project has started the ball rolling, but there's more work to be done.

Rosenberg is principal analyst at Open Source Development Labs, a non-profit, industry-supported organisation dedicated to increasing the adoption and growth of Linux.