Planet Green (Our Smileys Game)

Currently, we have:

One version that is fully functional as a choppy one player game which uses Commands to fire and move.
One version that is fully functional as a choppy one player game which has a start screen and game over screen.
A partially implemented network design (not currently functional):

The Client listens for key presses and mouse clicks
When input is received, the Client sends a MessageBean to the Server which processes that input and updates three Vectors holding Bogeys, Players, and Bullets respectively. Then it sends a bean containing all three Vectors.
Client has a thread to listen for input from the Server. When it receives it, each vector is transversed and each object drawn to an offscreen buffered image and then blasted to the screen.

Finalize the Client and Server. We have both done(theoretically), but they needed to be work out just a little to finish and get them working together.
On the Client we would have worked on double buffering more and would want to get it done also. (Our programs images look a little choppy when it is running)

We felt so much of a time restraint that many things didn’t get planned.
Code written only to test certain elements of the game were tightly coupled with the test code.
Versions were not well managed and there were branched versions, each having some new element to it, but without the new elements the others had.
Plans were hashed out and potential problems solved but without good documentation, we had to find the solution to a single problem several times.

We would sacrifice some time to create a UML diagram so we all knew the general function of the game and what each module should do.
We would write down quick notes when we found solutions to a problem.
Everybody would be assigned some module of the diagram to be responsible for, and be expected to present it as a module when finished. We think this would make implementation would be much easier.
“Test code” would not be distributed