If you really want to make something simple, just to get you started, maybe I have a nice plan for you:
First, select something small,little,tiny like Space Invaders, Frogger, Pacman, Pong, Bomber Man, etc.
For example, I selected the old arcade game Wizard Of Wor.
Second, get it started with all standard ways that jME, the SDK and other tools can offer you.
For example, I actually want to make a remake, with sprites or 3d models - depending on player choice.
I will use the same model for all characters, and maybe some free models from blendswap.
Third phase, get the first version finished.
Many projects get started, only few projects get finished.
Show the world, that you’ve got the guts to do even boring stuff, until version 1.0 is stable.
Last phase - progressive refinement!
Add some new characters and items that are not part of the original, add nouveau gameplay mechanics.
Add a physics simulation for explosions, add the A* for the a.i. guys, add a fancy shader, add…
For example, I will allow the player to switch top down view and first person view during game.
I have designed new gameplay based on new items, observed my own ideas from all possible angles.
If you are satisfied with the state you’ve reached - move to your next project.
Think of crowd funding, invite old buddies from several channels to join your fleet.
I think, the whole open source movement can be regarded as ‘collaboration’ - world wide collaboration…
Does this sound like a good plan to success?