Calling Linux desktops elusive seems silly since there are more than 36 million references to Linux desktops on Google and plenty of places to buy one easily, including Dell and eBay.
But while we write about Linux desktops often and assume that they will eventually make a difference in the composition of the worldwide market, we dont see much difference here in the United States.
We simply could say that the main market for Linux desktops is outside the United States, where there is much less penetration of both PCs and Microsofts ubiquitous Windows and Office—95 percent of U.S. offices are believed to use this combination. And that may be true—certainly for now and perhaps for the foreseeable future.
Office and Windows are too expensive (in dollars and in their need for big, modern computers) for the BRIC (Brazil, Russia, India and China) countries, where the need for computers is vast and the growth rate of computer sales is amazing.
Of course, the main market implementation model in those areas may not involve a PC at all, but rather some kind of small device such as a multifunction phone.
In the U.S. market, we can see many Linux desktops, but we dont tend to find them on desktops in the United States, unless we look at the specific markets where Linux desktops have begun to make inroads.
Some good hunting grounds include university campuses (both students and academic computing), government (especially federal agencies) and cost-conscious businesses whose business partners have decided to support Linux (to a lesser degree).
So what is holding back Linux desktops from general market acceptance?
Not Linux. The operating system works just fine, and you are no longer forced to order it from an arcane site as a series of separate downloads, requiring a bearded guru to integrate even a single desktop. Mainline vendors will sell you a new PC preloaded with Linux and ready to go.
Its not support. Novells SUSE Linux Enterprise Desktop costs $50 per seat with a year of support. And its not familiarity with the Linux interface. Current Linux GUIs are designed to look as much like the current version of Windows as possible, and current versions of popular Linux personal productivity applications such as Open-Office.org are designed to have Microsoft Office-like interfaces.
More to the point, Linux desktops can easily run software that permits them to also run Windows applications so that users can continue to use any critical Windows application that does not have a Linux twin. They do this by employing Wine software. (Windows licenses for any application that will be employed are, of course, required.)
And its not the infamous device driver issue. Linux distribution vendors such as Novell and Red Hat have been cleaning this up over time.
So what holds us back? Its the tyranny of the installed base.
The ability of a large installed base to present a substantial barrier to change is well-known. A messy period of supporting users on multiple systems would have to be endured. Moving documents from one system to another would have to work. Support staff would have to be trained on the new system while continuing to support the current one.
A big area of concern is skill sets. It takes years to build up a large enough pool of skilled resources to provide software support on any and every level.
These are the people you can hire as administrators, support staff and application developers. We have those people trained, identified and certified for Windows and Office. Weve begun to build a pool of people with Linux skills, but its still relatively small by comparison.
Increasing that pool—or building Linux systems that are much easier to support—could be key to boosting the U.S. market for Linux.
Amy Wohl is a technology analyst and consultant with a focus in the commercialization of new technologies and concepts. She can be reached at amy@wohl.com.