Why do this? With whom would you share it, in return for what?
Sale: Do the NetDevil thing. Do it better and retire early.
Fame: Open Source, or at least binaries for free. Wow the Open Source world, list it on your resume.
Fun: Don't show it to anyone, that way you can't ever be embarrassed. Low pressure. The default starting position, presumably -- unless someone feels an immediate need for more incentive.
John Harding - How about all three? ;-) I think the natural progression is Fun, Fame and then Sale. Of course, I suspect we never get out the Fun stage - but hey! at least it's fun!
Jim Horn - How about world domination?
Frank Swierz - I think NetDevil did it right.
- Fun - We're building a game that WE want to play. Who cares is anyone else is interested? Not me.
- Fame - Well, eventually we will get bored just playing all by ourselves. If we have fun with the game, then others will too. So get a beta going and let lots of people play for free. Plus, let them help us evolve the game into something even cooler.
- Fortune - If all those people are having fun then sell-out to some monopolistic game distribution agency and retire early as Dan suggested!
DWM - But if you go through a "Fame" stage as I described it (Open Source), you've greatly diminished the possibility of selling your code for profit. That's not the only way to make money off a game, of course, but I expect it's the least complicated and quickest.
Istvan - Open Source has appeal, yet nothing gets accomplished in a timely fashion, and you very quickly lose consistency control. That can be fine in the circumstance of building a specific tool, or building a very generic sort of game (in which central design might be said to be irrelevant). However, it strikes me that the "Fun" to "Sale" progression makes the most sense. "Fun" works unless the core resources have trouble dedicating time to the project, or until the project reaches a presentable stage where venture capital might be obtainable (not likely in the current economic climate).
DWM - Consistency control can be achieved in Open Source projects, and often is. Open Source is not synonymous with 'open access to the primary source repository'. But someone has to take the time to review submitted changes and accept, reject, or amend them. Personally, I'd rather write code than review it. OTOH, Open Source might just attract some people that would be interesting to work with, too. You never know.
The biggest problem, as you point out (but regardless of Open vs. closed source, I think) is staying motivated to finish things on time or at all. Peer pressure works up to a point, but eventually it's nice to have the carrot of at least modest financial success dangling in front of you. :)