It's better to think about something really interests me and brought me to computer science's world--the sims. So it's better for me to develop an indie sim game rather than a non-sense zombie shooting game because as a simmer I know what should I do with the game.
Here are some thoughts about the indie game I'm going to make:
1. Has first person mode and third person mode.
2. Supports mods. Using pipeline to import all the stuff into the game.
3. No stupid AI.
4. Multiplayer. The main mechanism is to set up a server by one player and to ask other players as clients to join the server. Will consider a global server.
5. More interactive world. For example, players can drive cars literally, and play careers, which can be made from several mini games. For instance, if you want to play as a dancer, you need to hit certain keyboard to finish certain animation and gain experience. No rabbithole.
6. Just like a normal sims game, you have CAS system, need system, relationship system, etc.
Feasible Ideas about how the game core works and how it supports modding.
1. CAS system. It is a pain because I'm using UMA now, which currently works on built-in sliders and meshes. Meshes are imported into game linked by a pre-made asset file. Maybe I'm wrong since I haven't finished reading all the code. But I'll keep working on it and make the code as mod-oriented as possible.
2. Object system. All the objects are loaded at runtime and are attached to the script. I'm thinking to choose FBX format since the it supports animations natively. And here's the link to the importer.
3. Navigation System. A* Navigation