There's been some discussion lately about promoting Linux as a gaming platform in order to win the struggle against proprietary operating systems.  It's an interesting idea, but I don't think it meshes well with reality.  Here's why.

First, this concept presupposes that "Gamers are adventurous folks," and "Linux adopters often need to be adventurous in order to even install a new operating system."  First of all, I know a lot of gamers who are actually not particularly adventurous when it comes to computers.  Playing World of Warcraft may mean that you like having adventures in fictional worlds, but it doesn't mean you like your operating systems adventurous, too.

More importantly, the objective of desktop Linux, or at least of Ubuntu, is supposed to be the creation of an operating system that doesn't involve much adventure, because it 'just works'.  Adventuring is great if that's your thing, but most people just want to use computers to get work done, and Ubuntu should appeal to those users before billing itself as an operating system for the adventurous.

Those arguing for Linux gaming also seem to assume implicitly that quality games can only be created if developers are paid and adopt a closed-source development model.  This assumption is inherently at odds with the open-source spirit at the heart of Linux, and will upset a good portion of the free-software community if it's used to push Linux gaming.

The idea that only closed-source developers can produce viable games is also flawed.  Ten years ago, Linux-haters argued that no decent office suite could be produced by an open-source project.  Even Linus Torvalds, they pointed out, was using Microsoft Office.  Then came OpenOffice (admittedly, most of its developers are paid by Sun, but its code is open nonetheless). OO may have its flaws, but it's clearly a viable and valuable product, despite being developed along open-source lines.

Another obstacle that Linux-gaming enthusiasts seem to ignore is the difficulty of porting games to Linux.  Sure, it wouldn't be too hard for those based on OpenGL.  But a lot of Windows games use DirectX, and unless Microsoft can be persuaded to release a Linux version, rewriting those games to run on OpenGL is going to be more work than most vendors will be willing to commit to.

What about wine?

Rather than porting games to Linux and dealing with all of the philosophical, political and technical conundrums involved, perhaps Linux advocates should work to ensure that more Windows games will run on wine.  This approach solves several problems:
  1. greater consistency.  Rather than having to tweak a game to run on Ubuntu, Fedora, Slackware, and your neighbor's custom distribution, vendors only have to deal with wine, which behaves relatively consistently across most flavors of Linux and on different kernels.
  2. vendors can more easily keep the source closed if they prefer, avoiding philosophical conflicts with free-software stalwarts who want to abstract non-free code from their systems.
  3. DirectX games run just as well on wine as OpenGL ones.
In general, there are a lot of reasons to avoid wine and promote native Linux software instead.  But when it comes to gaming, I think that developers are best off relying on wine as a vehicle to the Linux desktop.  Vendors should make sure their games will run well on wine, rather than delving into the much muckier work of writing native Linux games.

Moreover, I'm hardly convinced that gaming is the key to desktop-Linux's success.  Rather, producing an efficient, free and attractive operating system with a variety of productivity applications that 'just works' should remain at the top of Ubuntu's agenda.