In our phone conversation Jump readily admitted that this study was only one-way, and made no attempt to account for PCs sold with Windows that are later converted to Linux.
She also noted another interesting post-purchase operating system factoid not mentioned in this report: That an uncounted but significant percentage of corporate IT users are buying new PCs and laptops with Windows XP installed and "downgrading" them to Windows 2000.
Another point she made was that Gartner studies show the worldwide percentage of computers sold with Linux and converted to pirated copies of Windows is only around 40%, not 80%. The 80% figure only applies to Asian countries. She also said the percentage of Linux-loaded computers sold to home and small business users in Asian countries that soon ran pirated Windows was even higher, like, "probably 90% to 95% of consumer machines."
Jump specifically told me this report was not sponsored by Microsoft; that Gartner "does not do 'sponsored studies.'"
Indeed, it was not particularly pro-Microsoft. In it, she said point-blank that companies that migrate from Windows to Linux for kiosks, call centers, and other single-use or limited-use desktop environments "can achieve a positive return on investment in three years or less," and also wrote, "Gartner believes that this is a key consideration in the decision to migrate from Windows to Linux."
Jump also said by phone that she expects most PCs shipped with Microsoft's new Windows XP Starter Edition to run pirated copies of Windows -- or possibly even Linux -- soon after they leave the store.
Starter Edition can run only three applications at a time, has no networking capability, and is limited to 800x600 maximum monitor resolution. Jump -- along with almost everyone else who has looked at it -- believes Starter Edition is a mistake; that it may turn more people off to Windows than it attracts.
How many Linux desktops are there, anyway?
Gartner believes that, worldwide, about 5% of all new PCs are now being shipped with Linux, and about 1.3% of all currently installed PCs run Linux, with 2.5% running Mac OS, while the rest -- over 96% -- run Windows.
IDC and other analyst firms tend to peg desktop Linux usage higher than Gartner, generally in the 2.5% to 3.5% range. It look like no one knows how many people really use Linux. A count of my own office would show four computers currently active, three of which were originally sold with Windows installed, but where I now have only one Windows partition on one computer. The single Windows partition is used only rarely; I don't think I've booted it for at least a month.
After consulting my favorite numerical analysis tool, I've decided to peg general desktop Linux use at 2%, with enough kiosks, single-application terminals, and server/terminal systems running Linux to bring the total up to 3%.
Jump's study has a chart on page four that predicts steady increases in shipments of computers preloaded with Linux, with an expected worldwide Linux market share of 7.5% in 2008. If her expectations prove correct, the U.S. will be the world's laggard, with only 2.9% of all computers sold here running Linux, while Eastern Europe will be the leader, with 13.9%, and the Asia/Pacific region at number two with 13.0%.
Of course, in the Jump/Gartner world, 80% of those Linux computers may convert to pirated Windows, and the areas where software piracy is most common are most likely to be the ones where people buy new computers sans Windows -- as long as they get enough of a price break to make the extra work of loading a pirated copy of Windows worthwhile.
But things might change
Jump doesn't believe her numbers are cast in stone. She's extrapolating current trends based on a limited dataset. She knows that applications for Linux are rapidly increasng in both quality and quantity, and that there are still a lot of corporate Windows 98 and even 95 desktops in the world. And she told me that a migration from Windows 95 or 98 to Linux "is where you get a return on your investment quickly."
She also knows that once a company starts using Linux successfully on limited-purpose desktops, it might not be a big stretch for that company to move a majority or even all of its desktops to Linux, especially if and when it is faced with a choice between Linux and Microsoft's Longhorn Windows, which may hit the market in 2006. (Or maybe 2007.)
"How complicated the migration ... that is the biggest issue," Jump said. She was talking about migration not only to Linux but also to Longhorn. If Longhorn migration is perceived to be hard, and migration to Linux keeps getting easier, perhaps by the time Longhorn is a reality instead of a pipe dream the Linux choice won't be as hard to make as it is today.
Another possiblity Jump noted is that there might be enterprise-level desktop Linux pilot projects and test installs we don't know about that may result in sudden (to us) announcements of major Linux migrations that could skew everyone's projections.
She said continued Linux application development can -- and probably will -- make a difference in Linux acceptance among consumers and small businesses; that as soon as "missing capabilities" that currently make Linux less useful than Windows for many desktop purposes are added, Linux will become a more attractive choice than it is now.
In other words, Jump is saying what most analysts have been saying for the last five years: that Linux is a fine operating system but needs more applications and more hardware drivers (i.e. more cooperation from device manufacturers) before it can be considered a true Windows alternative by a majority of the world's computer users.
And until those applications and drivers are available, in regions where software piracy laws are typically ignored, we are going to see many people purchase Linux-loaded computers because they are cheaper than ones that come with Windows -- and put illegal Windows copies on those computers as soon as they get them home.
In light of these factors, Annette Jump's -- and Gartner's -- conclusions are plausible.
Note, too, that Jump expects Linux use to increase steadily over the next few years, and is open to the possibility that the rate of Linux usage increase may end up higher than she expects if more and better Linux applications and migration tools are developed.
Indeed, the only real disagreement between Annette Jump and Linux optimists is over two numbers: Is that provocative "80%" figure correct? And is desktop Linux usage today really as low as she believes?
But these disagreements are minor. When it comes to predicting the major trend in desktop Linux adoption -- steady growth -- Jump paints a sad picture for Microsoft and a happy one for Linux, and no rational Linux advocate can possibly argue with this conclusion.