There is a story about the game of chess and how it was invented centuries ago. After the game was invented, the ruler of the country was so happy that he offered the inventor a reward for their efforts. The inventor requested a grain of rice on the first square, two grains on the second square, four grains on the third square, and so forth. The ruler, speaking quickly, said "Let it be done." The problem was that there was not enough grain in the entire kingdom.

Today we might think of this as "two to the 64th power," a number so large that it is fairly unimaginable. Even two to the 32nd is over four billion, and, if you start with a number greater than two, it gets very large very fast. It is this concept that I was thinking about when I asked Linus Torvalds to port Linux from the 32-bit Intel platform to the 64-bit DEC Alpha.

Today, I would like to apply this concept to a different subject, that of desktop Linux, one of the few places where Linux does not have the penetration that it should have. Few people would say that Linux does not have penetration in the high performance computing market, the server marketplace, the embedded system space, and the cell phone space (as Android), but we have been stymied in the desktop space to the point where "The Year of Desktop Linux" has been a running joke for at least two decades.

In the early days of Linux, there were some legitimate desktop issues. Linux was cited as being hard to install, and Microsoft Windows was "easy." People ignored the fact that all laptops and desktops that people would normally buy already had some version of Microsoft Windows running on it, because that is the way that distributors and retailers sold it.

Very few end users ever installed Microsoft Windows, they just re-installed it from backup images already created for the system. Periodically, people would install new devices, but these devices would come with binary device drivers created for various versions of Microsoft Windows by the device manufacturer.

Today, there are few issues of installing mainstream distributions of Linux. With recent announcements of GPU manufacturers restructuring their drivers to include more open source and to integrate that open source with the kernel, there will probably be even fewer issues.

Other complaints were a lack of applications, which over time have been addressed by more applications being ported and use of virtualization, simulation, and emulation. More than that, this is a catch-22 issue: If there were more people using Linux on the desktop, more application vendors would support Linux.

Lack of games was another famous issue, which has lost traction with more games being ported, Steam creating a Linux platform for games, and game consoles becoming less and less expensive and becoming a platform used by many gamers. Additionally, native porting of games would be increased by having a greater desktop presence.

One of the real issues why Linux does not have more desktop presence is simply marketing and sales. The number of advertisements, product placements, and other marketing events that can be produced by FOSS companies is dwarfed by the marketing that can be done by Microsoft.

When I worked for Digital Equipment Corporation (DEC), a rule of thumb in the software industry said that 36 percent of the retail price of software was spent in sales and marketing. In other words, more than one-third of what people paid had nothing to do with engineering or even manufacturing the distribution media. Today, with the software either pre-installed or downloaded for installation, sales and marketing would probably be more than 50 percent of the cost, and (as in the case of bundled software) sales and marketing of third-party software (known as "bloatware" to readers) may actually make money for the original equipment manufacturer (OEM).

FOSS software finds it hard to compete with that marketing model on the desktop.

However, FOSS can do marketing using the chessboard model.

Each of the existing Linux users could spend some time with two Windows users and help them install Linux on their desktop. Help them find beginning books on Linux, help them find the applications they need, help them find online help groups and really get them started, and then next year we would have three times the number of Linux desktop users.

It is even better if those Windows users are teachers, town council members, government officials, business people, etc.

Then the next year each of those three find two more Windows users. This might sound like too little and too late, but we are starting with millions. There are approximately two billion desktop users. It is estimated that three percent of these are Linux users (I am using very conservative numbers), so that is 60 million Linux desktop users.

I think it is time to play chess.