All Linux vendors need to do to break Microsoft's death grip on the desktop is unravel a few chicken-or-the-egg type mysteries.

Which comes first? Widespread demand for Linux operating systems preloaded on PCs, or hardware manufacturers anxious to do the loading? The latter are as scarce as hen's teeth today.

Which comes first? Confidence among would-be converts that Linux support for the most important Linux applications will approach that of Windows any time soon? Or a developer community that sees enough demand to make that support happen? Too many developers put all their apps in one basket as things stand.

None of which means desktop Linux isn't creating a buzz: witness the major product development and marketing commitments from IBM, Novell, Sun Microsystems and Red Hat. The penguin recently waddled past Apple's Mac OS into second place in the desktop operating system market, according to IDC. However, second place means a 3 percent share, which IDC only sees rising to 6 percent come 2007. But progress is progress, the Linux advocates say.

Besides, those figures don't paint a true picture, says Bill Weinberg, open source architecture specialist for Open Source Development Labs (OSDL), an industry consortium dedicated to advancing the cause. Market surveys account only for commercial desktop Linux and not the myriad free downloads and installs, leading some to peg what IDC calls 3 percent at anywhere from three to 12 times higher. Weinberg says he believes that, "the truth is somewhere in the middle."

Where the floor is today and where the ceiling may be tomorrow are matters of great debate for those who follow Linux.

"It's unclear to what degree Linux will ever displace Windows so long as it's merely a cheaper but almost as good alternative, which is how it's generally positioned today," says Gordon Haff, a senior analyst at Illuminata. "The breakout would be to deliver a Linux-based desktop system that was compellingly better than Windows in function, look and feel, or some other dimension."

However, Haff adds, such a breakout advancement also could prove a challenge for Linux vendors because it would entail new training.

Linux desktop adoption really should be viewed as an outgrowth of Linux in the data centre, Weinberg says.

"For the corporate information worker, technical workstation, financial transaction terminals, data entry – anywhere the scope of activity is well understood and where installation and support are backed by IT staff – that's a very good place" for desktop Linux, he says. "Our members who have adopted Linux in the data centre are moving in that direction."

They're moving with good reason, says Jeremy White, founder of the Linux Desktop Consortium and CEO of Codeweavers, maker of the CrossOver Office program that lets users run applications such as Microsoft Office and Lotus Notes on Linux. Desktop Linux has made great strides from a technological standpoint, he says.

"There are challenges, there are warts – pretty serious challenges and warts – but technologically both the Gnome and KDE (user interfaces) are pretty comfortable. You can install them pretty reliably on a laptop now," White says. Besides the "cheaper/just as good" deployment issue, he spotlights limited applications support as one of desktop Linux's more serious challenges.

"There are some great applications for Linux – the Mozilla suite, the Firefox browser is fantastic, there are great e-mail clients… But the real issue is that you can't get the apps you want or the apps that you're familiar with. That's a nasty Catch-22," he says.

A number of IT consultancy reports issued over the past year have argued that long-term considerations, in particular applications support and training, will offset any initial savings that might be gained from Linux rather than Windows. The Linux evangelists give little ground on the point, however. "I absolutely think there is a return on investment, but I don't think your payoff is in the first year," says White of the Linux Desktop Consortium. "Your time to ROI is a three- or four-year horizon."

White's counterpart at the OSDL insists that Linux on the desktop passes the ROI test where it counts most: in real-world deployments.

"We don't have definitive data, but we know that our members' customers are not choosing Linux because it's more expensive or because they have an emotional or ideological commitment to it," Weinberg says. "They are finding that the (total cost of ownership) is better over the owned lifetime of the device. Because Linux doesn't require constant, aggressive upgrades to hardware, the deployed lifetime of your Linux workstation is longer so you have a better chance of amortising the costs."

Still, the bottom line is that few see anything but a marathon sales job ahead for desktop Linux.

"It's just a long, hard slog," White says. "I think what everyone is looking for is sexy, sudden, magical, transformative spike, and I think this instead is what we do: We stay in the trenches and keep fighting."