Boaradoor wrote:
Well, I'm not entirely sure if people would buy a GBA game...
I think a PC/ Maybe DC release would be better

I'd avoid anything Nintendo... getting devkits for anything like that is a huge pain in the ass and they're really picky about who they choose. Nor are they cheap.
Same goes with Playstation, though it's easier then the big N.
Microsoft on the other hand is really lenient and rather affordable.
Both the PC based DirectX and XNA SDK is free to use and develope with. And there is tons of resources and communities out there about them.
OpenGL (not MS related) is free, but the amount of documentation for it is skimp compared to directX.
X360 dev isn't so affordable, and creating an actual title isn't the easiest of deals... but XNA dev on the 360 can be had for pennies... you can even release them in the XNA indie marketplace for very little. I personally applaud XNA because it can be easily be designed on two platforms (PC and 360), and with projects like "Mono" (C# .NET for Linux and OSX) and "Mono.XNA" (XNA written for Mono) means a very soon possibility of easily creating projects across all the main OSs and X360 with very little porting effort (sometimes as little as just a recompile for the different machine).
The prototype I'm working on right now is actually in C# with XNA, and leaving it open enough to allow for simple porting with Mono.XNA when it gets off the ground more.
As for older out of date hardware, especially like Dreamcast comes with huge hurdles. The documentation is really at a loss here. Just finding the latest compilers is a pain in the arse. There are nice homebrew communitites based on some of these (like XBOX and DC), but even then the homebrew community sticks to some of the more underpowered features of Dreamcast, usually building it on top of some base kernel like Microsoft CE or a linux kernel compiled for the machine. And if you go back further you run into even more headaches... like a cart based console has tons of free dev kits floating, of which very few match each other, and require different paradigms. It's a restricted C of some sort, or C like language and take a performance hit... or it's reading the documentation of the processor and architecture and writing scaled assembly across the entire board...
not impossible, I'm just saying you're looking for a very nerdy crowd who are willing to dedicate some time here. Of whom I bet several would be hesitant to work in a community that has a majority of none or novice coders.
I don't mean to discourage... but some forward direction is necessary and a viable understanding of the community as a whole. Not to say any individual here doesn't have a specific skill necessary, it's about managing and combing said skills.
And worse more what happens when a group of individuals who put time in turn out to really just handicap the whole project in an entirity... a lot of broken hearts when the other half say "sorrz mate, but well... you kinda suck."
Just sayin'...