Having designed one and worked with another, it's all about constraints. I'd start by asking some questions. What kinds of games? Text games? 2D games? Ani

Having designed one and worked with another, it's all about constraints. I'd start by asking some questions. What kinds of games? Text games? 2D games? Animated games? 3D games? Existing games? New games? Twitch games? Slow games? Multiplayer games? If existing games: What platform are they on? Windows? Do they use DirectX? Which parts? Which versions? OpenGL? What do they use for music? Mac? Are they PPC or x86? What version of OS X? Linux? I don't even know what questions to ask for that. If new-but-existing games: Are you expecting them to be recompiled or ported to your platform? What's involved in that? What constraints does your platform exhibit? If new-exclusive games: Why on earth would anyone build for your platform? What's your penetration? That both informs and places constraints on your hardware and software infrastructure, which in turn places constraints on your ongoing costs to keep the lights on, as well as on how end users will experience your games. Of course, all of this assumes you've done all the customer development and market research that would justify building one after both of those essentially failed in the market.