Dodgeball

3D Dodgeball Game

NOTE: Check the wiki for the latest up-to-date information. This readme will only be updated when there is a major release.

In dodgeball there are two teams of six players on opposing sides of 10x10 m court. For the sake of simplicity, this game will not include all the rules and regulations of dodgeball as laid out here (such as substitutes or "The Rush"). Ultimately, this game will allow the users to:

• Form two teams of six via a computer network i.e. a networked game.
• The two teams are "blue" or "red"
• Players obtain balls by bumping into "inactive" (balls on floor, balls that have bounced, etc...) balls (if the player possesses 2 or fewer balls)
• Players can possess up to three balls at any given time
• Players can be hit by an active ball up to three times before they are "out" (removed from the match) "lives"
• A single match lasts five minutes with 12 balls on each side. Balls are conservative i.e. there will always be 24 balls total in any point in time during a match.
• A single gameplay will consist of a single match. (Once the match ends, the program exits with a message of who won etc...)
• The winning team is the one with the most players at the end of the match.
• Matches will end if one team lost all its players.
• The game will only initialize once all twelve players have joined the session.
• A central server will manage every match.
• Super-balls that have unnaturally advantageous characteristics (like a "missile" ball, a multi-bounce ball, etc...)

Unfortuantely, due to real-life time and resource constraints, there is a large chance that most of these features will not make be complete. In fact, 3D computer games are notorious for ruining a student's project due to the involving nature of games. It is very likely that over the course of the ten weeks, there will be crippling problems that surface at extremely inconvenient times. But it is this uncertainty that makes game projects so thrilling :)

Plan

In order to address the issues above, I'll be taking a two-phase approach: The first phase is a proof-of-concept app that demonstrates the feasability of the game. If the proof-of-concept app yields negative results, then the project will convert to implementing a simple path tracer. If the proof-of-concept app yields positive results, then the project is officially greenlighted (second phase).

Unfortunately, even though the proof-of-concept shows that such game is possible under given constraints, it is most certainly does NOT indicate a successful project. Chances are, there will be major roadblocks that cripple the project. To account for that, my plan will have "plan Bs" laid out beforehand in case things go awry. Ultimately, if everything crashes and burns, my project will default to the Google AppEngine CMS.

The main objective of the project is to learn how games are put together and how to bring together various different components in a cooperative environment. The actual game's usability or user base is second to learning. Therefore, I predict that the end result of this project will lie between two "states":

Worst-Case

Many roadblocks are hit, causing major parts of the game's features (but not functionality) to be omitted from the result. This is the state where the resulting game doesn't differ much from the original proof-of-concept.

• Players can throw a ball
• Balls interacts with the scene
• No network management
• No super-balls
• No other players
• No ball possession capabilities
• No "lives"
• No "matches"; the game just runs forever
• No ball-player collision detection

Best-Case

Where all features listed above are met. If I get here, something is wrong with the world (or I just got extremely lucky).

So somewhere between best-case and worst case is where this project will most likely end up if the proof-of-concept passes. In other words, I really don't know what will happen :) The project is a success if it ends up at least the worst-case state.

Timeline

Each week there is a goal that must be met or else the project will immediately revert to the app engine or path tracer "mode" depending on how far along the project has already reached.

Prototype (Week 1)

A proof of concept app must demonstrate:

• An FPS-style interface
• Ball throwing
• Reasonable texturing of models within the game
• Basic court setup
• Physics integration (balls bouncing, collisions, etc...)
• Bouncing/collision detection (ball hit floor/player, player hit by active ball, etc...)
• Everything is as simple as possible: players are just cubes, balls are just spheres, the environment is just a box, etc...
• Devise an architecture for the game: Classes, methods, modules, etc... Everything is gonna be a state machine!!! If these conditions can be met by the end of the first week, then the game is greenlit. If not, then my project reverts to a pathtracer app.

Multiplayer (Week 2)

This stage must demonstrate multiple players in the game. Players can be controlled with keyboard strokes. Only one screen/view is required, no split-screen needed.

• More than one player is displayed in the game; the model itself must be a rudimentary character
• The players do not need articulation capabilities
• The players can move
• Only two players required
• No network support required (no split screen)
• Collision support; players don't have the capability to walk through walls, etc...
• All players can throw an arbitrary number of balls

If these requirements cannot be met, then the project is converted to a pathtracer.

Multiplayer II (Week 2)

The multiplayer-ness is enhanced, allowing 6 players on either sides of the court.

• Six players on either sides of the court.
• Players cannot travel to the opposite side.
• Players are articulated with rudimentary animation (running, idle, throwing).
• Rudimentary gaze following (player faces the right direction)
• The game stops once a player gets hit by a ball (no "active/inactive" capability needed)

If these requirements are not met, then the project is converted to a pathtracer.

FPS Enhancement (Week 3)

Dodgeball requires a wide field of view. Therefore, the camera projection matrix will need to be modified.

• Experiment with the camera's projection matrix to allow for wider field of view
• Add a crosshair :)
• Add HUD elements; can be rudimentary text/images
• Fine-tune the user interface/interaction with the FPS style.
• Possibly rewrite the camera FPS system to avoid using Irrlicht's createFPSCamera...(...) function.

If these requirements are not met, then the project is converted to a pathtracer.

FPS II (Week 4)

I'm anticipating issues with the HUD and FPS-ness of the game. The object for this week is to freeze further development and assess the project's state. There may be requirements that have been pushed around, and this week would be the time to fix them.

• Enhance HUD; display basic indicators like numbers and text
• Begin adding state machines to the balls and players.
  • Balls have active/inactive status so they cant "kill" a person once it has bounced on the floor etc...
• Start adding lives and ball possesion statistics.
• Limit gameplay time.

This is the "point of no return" for the project: Past this point, there's really no turning back. All "failures" after this point would cause the project to convert to the AppEngine state. So I'll be assessing the situation at this point to see if I should consider switching projects to a more manageable one. This is also the point at which if I think the game will not make its goals, then I'll switch my project to the pathtracer. This is the scary part ;)

Networking (Week 5)

Networking will be tricky. Irrlicht is a C/C++ library yet I really want to use python's twisted API. So this week will be to check the feasability of networking. I'm thinking of running a Twisted server that tells each client when to start/stop the game and the ip addresses of all the other players. So the question I want answered by the the end of this week is: Will this game be networked?

• Now that I've come to think of it, netowrking will be EXTREMELY tricky (synchronizing bullet simulations is a nightmare). Depending on how this week goes, if it looks as if making the game networked is not going to happen, then the multiplayerness will come from splitscreen/multiwindow/monitors. The number of players will be reduce to 6 total per match (3-on-3). Controls will then be transferred to joysticks. UPDATE: I found out about (ENET)[http://enet.bespin.org/], which is a networking API suitable for games. This might be key for a successful networked game?

Networking II (Week 6)

I'm anticipating that if I deem networking the game feasible, it will take multiple weeks to get it working. Therefore, this week is devoted to planning and designing the roadmap for networking the game. By the end of the week, I want to determine:

• TCP vs. UDP?
• What communications model to use? Client-server? Peer-to-peer?
• What is the protocol? How will clients and servers create, join and start a game?
  • How will I deal with corner cases? What happens when a player suddnely leaves a game? etc...
• What is the packet structure? How do I serialize the game's state?
• What happens when the network latency turns out to be prohibitively slow?
• How will I test the game if it requires networked computers?

Networking III (week 7)

This week is dedicated to implementing last week's network specifications. Specifically, by the end of this week:

• Computers must be able to communicate with each other under a sound protocol.
• The game state must be serializable.
• A central server has the capability to start and stop games.
• Scorekeeping functionality.
  • Who wins/loses
• Player movement is synchronized. But the physics simulation state need not be synced.

Super-Balls (Week 8)

Start experimenting with cheats in the system such as missile balls that track a specific player. Other ideas below:

• Missile balls: They "lock-on" to other players and once thrown, will follow the target until it becomes inactive.
  • It should be rather simple, just apply a force vector on the ball that is aimed at the player (difference between player vector and ball vector). However, the ball will continue to fall, so at some point it will still hit the floor and become inactive.
• Multi-throw: Throw all possessed balls at once.
• Fireball: An impossibly large impulse is applied to the ball upon throwing, essentially becoming a bullet. This is kinda fun because with the bounciness of the ball, the results could be amusing...
• Aside from experimenting with superballs, if the network game idea is still sound, then this week will also be dedicated to
  • Synchronizing game states. Bullet has a serialization solution for transferring game states, I think that's the way to go.
  • I'm thinking that the server will run its own bullet simulation of the game, and stream it to the clients.
  • All clients send their moves to the server, while running the entire game themselves too. They just sync with the game server's version.
  • Another way to do it is to have the server just relay each player's actions to everyone else and maintain its own official game simulation.
  • The server would broadcast its official game state to all clients every few milliseconds to sync up Worst-comes-to-worst, I'll just focus on making the fireball. Its the balance between fun and ease of implementation.
• If the networked game system doesn't look like its gonna work, then just setup the split-screen view...
  • maybe we can limit the number of players on the network, while still having a networked game just with two computers only.
  • Kinda peer-to-peer ish. That way we won't need a single computer to have to handle 6 seperate views...

Finalization (Week 9)

Home stretch! This is the time for polishing the game's functionality, clearing bugs and finishing TODOs.

• Texture the models and build a generic gymnasium as the envrionment.
• Players should be articulated and have some form of movement/texture. They should have colors too.
• The HUD should look presentable. The mouse should be hidden from view and constrained to the window.
• Basic keybidnings for quitting the game should be implemented.

ROLLOUT! (Week 10)

Further development is now frozen and all effort goes towards fixing bugs and testing the corner cases. Further cleaning and refining of the code and user interface is good, but no new features should be pursued. Possibly employ some "beta testers" to help check the game?

• Possibly start porting the game to windows/mac if the game will be played during class time?

Yay! (Week 10)

Hopefully by now the game is in a good, working state. :) All there is left to do is to play it!

NOTE: If for a given week that most but not all requirements are met, then those unmet requirement(s) may be moved to the next week.

Final Thoughts

This is an ambitious project, and not every part of it will work. There is a good chance of the whole thing crashing and burning but there's an equally good chance that it will succeed. Its like landing the rover on Mars; risky, but fun nonetheless.

Sammy's Project

Right now, I have two ideas that I have been bouncing around in my head:

• A Google AppEngine app that behaves as a CMS...
• A 3D dodgeball game using Irrlicht+Bullet+network (hopefully)
• A pathtracer...
• Mini operating system (???)

Right now I'm debating between the game and the appengine projects; the major battling point is between practicality/awesomeness :) Specifically, the game will take gobs more time and effort to complete, BUT its a little funner. On the other hand, the appengine project is very likely to work, BUT it is mostly just reading the API docs and calling functions.

For now, I'm leaning towards the game...