I want to…
- Finish these stupid courses - I have to use Java again for the other one, eww
- Prototype a game idea I had while driving home today - but not spend over a week on it
- Learn to draw some vector art - because it should be easier for "programmer art" than pixels
- Get a new job for Autumn
- Get some sleep
Artillery Brawl: Prototyping
Lately I've read a lot about game design (Danc's blog has been especially enlightening). I'm now convinced I have to prototype a few "risky" aspects of Artillery Brawl's gameplay to make sure what I'm doing is actually fun before wasting effort on other parts of the game.
The core gameplay of each game has to be fun for the game on the whole to be fun. Platformers are about jumping, shooters are about shooting and Go is about placing stones on a board. Players are going to spend most of their time playing doing this core activity; everything else is basically extra to hook players at first or keep them interested in the long run. In Artillery Brawl, core gameplay is of course shooting, and before I dwelve into anything else, I have to make shooting fun. In other words, it has to be rewarding.
Artillery Brawl isn't fun yet. There are a couple of obvious problems I'm going to fix first and see what the results are, then decide what's to follow. First, there's absolutely no feedback or reward for shooting; there aren't any kind of effects (visual or aural) when you shoot or hit something, and units simply disappear when they are destroyed. Not very fun at all. So I'm going to add some particle explosions and sound effects to help with that. Explosions are always fun, right?
Second, there's the user interface. Currently shooting is controlled with the arrow keys and modifiers (alt and shift), but while I want to keep that, I also want to add mouse control that's dead simple to use. It should be immediately obvious (at least after pressing the Shoot! button once or twice) what to do in the game without any kind of documentation or tutorials.
Last, but certainly not least, there's the camera. Currently, other than automatically focusing on the current unit when a turn begins, the camera is completely manual. With the turn system I have in place; multiple tiny shells in the air, many things happening at once and many players in front of the same screen, each interested in different things, this simply won't work at all. I have to improve the automatic behaviour of the camera, and luckily I have a couple of simple to implement improvements in mind. At the beginning of a turn, the camera should focus so that the current unit and the place the last shot landed should be visible. Between turns, the camera should zoom out, showing as much as possible. When a shell is about to land, the camera should zoom in on that. Later, I can improve it so that if multiple shells are about to land at the same time, I can use a picture-in-picture to show both at once. Also, I'm going to improve the visibility of both units and shells, and add a minimap.
If, after these improvements are in place, the game still isn't beginning to approach fun, I'll have to either change something dramatically or start a new project. Simple as that.
Artillery Brawl: The concept
I'm starting to write about the game design for Artillery Brawl on this blog instead of keeping it in my head and my private wiki. As opposed to Red Nebula, whose concept is wide open, a "playground" of sorts, the vision I have for Artillery Brawl is very focused, and I have already dismissed several gameplay ideas that simply don't fit that vision. To start with, here's the concept for the game.
Artillery Brawl is a remake of my first real game, Arty, and a tribute to Artillery Duel, Scorched Earth and other games of the classic artillery genre. Two or more players, with one or more artillery units each, battle to gain victory by destroying each other or reaching some other victory condition depending on the rules.
Core features
- OpenGL-accelerated 2D graphics
- Realistic physics simulation
- Procedural levels
- Real-time gameplay paused during turns
- Damage model not based on hitpoints
- Multiple game modes with different rules
- Shopping system to buy and customize units
- Hotseat multiplayer on one computer
Everything else is extra, but the above features are central to the vision.
World
The game world is a 2D terrain always viewed from the side and bounded on each side. No unit or other entity may cross the bounds; for example, artillery shells crossing the bounds are silently eliminated. To help with this, there is a "buffer zone" on each side, where no units are placed at the beginning of the game. Scale is at the unit level, with no higher level organization or lower-level detail.
Gameplay
A game consists of a number of rounds. Before a game, the players agree on a set of rules, including ending conditions for each round and the game, and how money is gained. A round might end when only one player has units left, or after a time limit. A game might end after a set number of rounds, or when only one player has money left. Before each round, the players may buy new units and customize existing units at a shop screen with their money.
Each round has a continuous timeline. When it's not anyone's turn, time advances in real time. Time is paused during turns, and continues again when the turn is finished. During a turn, a player may give orders to their units, but units only act between turns. A player may start a turn any time they want by pausing the game, or automatically when a unit has finished its current orders.
Style
The style of the game is semi-realistic; not terribly serious, but not "casual" either.
- Units and other equipment will be loosely based on real equivalents
- Units and other entities are affected by physics simulation, but some gameplay elements might take liberties with this
- The world is larger than usual in the genre, though not realistic; we don't want units to be kilometers apart
- The damage model is not based on hitpoints as usual, but will be simple
Box of horrors
Today I adapted the nice setuptools-based plugin module from Spineless to more easily support different physics toolkits in my code, and played around a bit trying to write a Box2D plugin. Frankly, I don't like the Box2D API design at all, and I like the pyBox2D API even less, as it's a direct port of the C++ API and doesn't even try to be Pythonic. Eww. Seems I'll have to try fake sweeping collision detection and other tweaks to pymunk next to fix the collision problems I wrote about.
In other news, starting tomorrow I'll try to finish two university courses that are really long overdue and that I want out of the way before September. I have one short paper to write and one coding project to do, so game development progress will probably be a bit slower again for a while. But every finished course takes me closer to graduation (which I'm planning to do next year), so yay for that!
