Games have always been a special case for Free Software because, as RMS freely admits, they combine elements of books and music which do not apply to many of the GNU general rules about keeping software Free.
"You can't make money off free!" the pundits say, confusing "free" with price. Still, people are going to want to make money off it. We need to ask ourselves just how we can still make money off software when it has all been liberated, as the FSF hopes.
In reality, there are many ways to make money off of Free Software. Donations, selling extras (T-Shirts, posters, etc.), charging for updates, and so on. All these ideas have problems of one form or another, especially when applied to a for-profit business, although lots of non-profits have been doing similar things for a long time.
What about games? They are a special case, remember? You might try the "selling extras" path (who wants an action figure of their favorite Quake/UT Warrior?), or perhaps giving away the game and charging for new levels (which isn't a very purist way of going about the problem). What else is there?
TransGaming is showing what is needed: Imagination. Perhaps many people suspected it all along, but I think now their claims can be solidified. Their plan, as I understand it, works like this:
People want games to run on GNU/Linux. Getting individual companies to port each individual game is a lot of effort. However, if we could get the WINE project to run those games reliably and without a noticable difference in speed than the Windows alternative, then we would have the whole bundle. What is needed to run most Windows games? DirectX. Recreate a DirectX API for GNU/Linux and you've got your games. So, to make money, we will ask people to subscribe to our service for x dollars a month. The benefit of subscribing is that you can vote on which games should be focused on first. With time, the DirectX API on GNU/Linux will be more and more complete, eventualy hitting 100% of the available Windows games. All the code can (eventualy) be released as Free Software, since we don't make money on it directly.
The only hole I see in this plan is that once you get 100% of the games and all your code is released, where do you go from there? This is a long term problem, and one I'm sure the company will have to deal with eventually. At this point, they will need to come up with some other way of making money. However, in the short term, I believe they will do quite well.
I always laugh when I hear some pundit say "All the good ideas in the 'New Economy' are taken! All you can do now is copy what everybody else is doing!" Then I hear about some other company doing some brilliant new idea on the Internet, at which point the pundits say "Now all the good ideas are taken!"
No, all the good ideas are not taken. It's just that no one has discovered the next Big Thing yet.