There are two kinds of Linux users. Political and Technical.
The Political Linux user will have long abandoned any technology which hasn’t reached his or her standards of political extremism. They’ll have removed all Windows partitions and yet resent their bank for not catering to their minority needs and the iTunes store and themselves for wanting doohickeys like iPods. They’re the ones with the various shades of window manager and boasting about how power management works. Easy to spot. Easy to lose in a cafÃ© too (just close your MacBook and leave. They’ll take a minute or six to shut down and get packed up.
The Technical user will, of course, be expedient with his or her use of technology. They’ll likely use a MacBook of some shape or size (because, you know, if you don’t you’re some sort of weirdo) which may or may not dual-boot to Linux or Windows. The only reason they have Windows is for their bank or maybe so they can actually play some decent games.
Of course, neither of these definitions explains exactly why there are so few games for Linux. It could be the (entirely correct) perception that Linux users don’t pay money for software. That’ll be a big one right there. And while companies can make a buck selling support for Linux as an operating system, selling support for games isn’t going to go far as people just hacked off when a game doesn’t perform.
What I wonder, however, is why there hasn’t been some sort of “x86 gaming platform” invented. I mean, almost all the hardware out there runs on x86 based machines now. Why not engineer a solution not dissimilar to the PlayStation where the OS was loaded from the disk at the same time as the game? Why hasn’t Intel pulled their finger out? We’d end up with a system where we bought CDs and DVDs, maybe even USB keys, with a base Linux kernel that would autodetect the hardware, run the drivers and autoload the game. The entire game would almost be copied into RAM and there’s your solution. Reboot to play, takes a few seconds to boot and doesn’t require using Windows.
Right. That’s the hard bit thought of. I’ll leave the easy bits (the technical side, the code, hardware, distribution, licensing, advertising and sales) to others.