With its recently launched assault on Microsoft's Windows, Red Hat is staring into the eyes of the dragon, hoping that, with its nimble actions and sharper sword, it can wound if not slay the foe. More specifically, it's pinning its hopes on better products and alternative methodologies of both code development and code distribution.

But let's look at the facts. Red Hat has just launched Linux as an end user desktop service for enterprises - a low-key, demo-free affair contrasting with its brasher, more commercial competitors. Red Hat Desktop and OpenOffice are to be delivered to a thin client via Web browser Mozilla, with the aim of usurping Windows and MS Office. The combination offers, said RH CEO Matthew Szulik at the London launch, a better, more efficient service, more secure and more manageable. Szulik said that customers were asking for more.

Analyst James Governor concurred, highlighting that many of Microsoft's customers have paid for upgrades but are failing to actually install them. Governor concluded that Microsoft had lost the faith of its customers.

Szulik agreed it was going to be tough. He pointed out that others who have taken Microsoft head-on have all disappeared or morphed into something else. The list is long and includes Novell/WordPerfect, Borland, Lotus, IBMÂ… but this time it's different, said Szulik because of that disconnect between Microsoft and its customers.

However, some of his claims for the advantage that RH has over Microsoft could suffer somewhat from subsidence. For example, following a series of tit-for-tat surveys and reports on total cost of ownership, one might conclude that the cost of running desktop Linux is at least as high as for Windows. This could change though, as end user familiarity and the number of certified engineers grows and skills shortages ease.

Better security is another claim made for Red Hat Enterprise Linux. Yet here too, Linux has suffered at least as high a number of vulnerabilities as Windows, if not higher, even though the latter's problems have been far more widely publicised - quite rightly of course, given the sheer number of users affected by those vulnerabilities. Szulik countered that the diversity of Linux desktops means the likelihood of a worm gaining traction in the Linux environment is lower. What's more, we're not Microsoft and are less of a target, he said.

It's also not going to happen in a hurry. At the launch of the RH Linux Desktop, Szulik was vague about timescales, and conspicuously failed to involve major hardware vendors at the launch, as Microsoft traditionally does.

But haven't we been here before? Those who have previously mounted attempts to dislodge Microsoft from the desktop - Sun and IBM among others - have deployed a similar argument: launch a service using the undoubted benefits of thin clients to crowbar their way into the enterprise. It's worked in many circumstances, especially for task-specific environments where the desktops can be locked down.

This is where Szulik sees the low-hanging fruit, as RH is aiming at government and academic as well as enterprise environments. Implementing security in these areas is much easier because end users can only access a limited feature set. And he's right too that, following the massive panic spend by enterprises following the Y2K debacle, only now are the purse-strings starting to loosen. At hardware refresh time, the OS market is, theoretically, up for grabs.

So what conclusions can we draw? It bears repeating that, in spite of the arguments among cognoscenti, Windows XP is robust and offers comfortable familiarity to most end users, even though the case for Office is probably more permeable.

The technical merits of Linux are many, though possibly not as overwhelming as the religious wing of its army of proponents claims. What it does offer is a price advantage and the arguable benefit of thousands of open source community eyeballs surveying the code - even if this does generate its own set of problems. From a market perspective, Red Hat has it all to do, especially since it's not sidestepping but exploiting the traditional pre-load market, led by hardware vendors HP, IBM and Dell.

Arguably, Dell is key. Big enough to make or break the Linux desktop market, it is rarely first, tends instead to adopt technologies halfway up the market adoption bell-curve.

So the Texan PC maker is a good bellwether: Dell enters markets when they are big enough and when the cost of support can be amortised over large volumes. Linux isn't there yet but, from his eyrie in North Carolina, Szulik is probably keeping his antennae tuned to signals from the south-west.