We've heard countless times that “this year” will be the year of the Linux desktop. The Linux user base has certainly grown, but not at the ratio that a Linux vendor like Red Hat would want. It's for that reason that Jim Whitehurst, CEO of Red Hat, has expressed doubt that desktop Linux has mainstream relevance today. Addressing questions of interoperability, adoption rates, cost, support and more, Whitehurst claims that they can't make money on Linux desktops and may not even plan to. Further he suggests that while Linux has been successful on servers, even trying to push Linux on the desktop is a foolish move at this point, because what people need now will be radically different in just a few years.
It's an interesting commentary that does bring up a good point. You can break down today's desktop usage into office work, gaming, multimedia, social networking and numerous other facets. The majority of these did not exist when the desktop was spawned, and who knows what new facets will come to light as the desktop continues to evolve with the advent of cloud-based computing and smartphones. If Linux tries to compete with modern-day Windows desktop standards, but does it on a game plan that lasts five years or so, it seems that it would be a losing proposition.
Perhaps the stance of competing with Windows is the wrong one to take. Rather than overcome existing barriers, Linux vendors need to predict what sort of obstacles the desktop will face in five or ten years. It seems Red Hat may be on to that notion, and could offer something completely different in the future.