Year: 2004
During the first year of undergraduate study at Epita we have to develop a project in either Delphi or OCaml within a team of four to validate the year. Since this is the only year we are allowed to make a video-game, we chose to create a Diablo-like in Delphi as our school project in 2004.
For most of the student, and for my team, this is the first programming project we do ourselves. As such, we were lacking best practices and we learned along the way. The development was quite hard, because at the time we had no programming skill. Moreover, our team quickly fell out, with one member quitting the school after a few months and another one struggling with programming. It ended up being a two-person project.
Nevertheless, this was a truly rewarding experience since I learned the basics of video game creation and DirectX programming.
We had the choice between DirectX and OpenGL for developing our graphic engine. Most of the students chose OpenGL but we thought DirectX was more suitable for a Windows game because we had no intention to port the project to Linux.
We used the Jedi Project headers as a means to use DirectX with Delphi. The first contact with DirectX was harsh but after carefully reading the SDK documentation I was able to grasp the basics. I would like to thank my fellow student Buckman for helping me figure the logic of DirectX out.
The 3D models should have been generated with Maya but we were not able to produce any good 3D assets by ourselves. So we searched graphism websites for simple models and we modified them with Maya. For the texturing part and the conversion of the Maya file format to the DirectX one, we used Deep Exploration (by Right Hemisphere).
In order to make things dynamic, we just added a game loop with mutable variables (by pressing keys for moving the camera for instance). Unfortunately, the game loop was not set on the time but on the cpu cycles. As a result, it runs as fast as your cpu can, making things lagging.
Finally, I added a 2D display for game informations such as health. It was done using DirectSprites and DirectFonts. A bug with the proportions forced me to tweak the sprites size manually in order to make the game look as expected on screen.
For the pathfinding, I used the classic A*. Each tile of the map is enabled or not, make a boolean map on which I applied the pathfinding.
“Sound engine” might sound a bit fancy for what was done. At first I thought I would use the DirectSound library since our graphic engine was done with Direct3D but I quickly changed my mind.
DirectSound allows you to use incredible 3D sound effects like reverberation but it was a nightmare to use it. I spent an entire night just for trying to play a simple .wav file. Without success.
Then I tried Fmod that everybody was using at the moment and I was able to play a .mp3 file in 10 minutes and 3 lines of code.
For the needs of the game, I included a background music and I added step sound effect when the player is walking.
The AI is just some nested conditions (“if .. then .. else”). The principle was the following: an enemy is making his rounds and it attacks of the hero is close enough. The enemy goes back to making his rounds if he is almost dead or if the hero is getting away (the hero is of course faster than any enemy).
My friend created a truly rich map editor that was used for making the game. He also created the elements of the world. He was then able to define position and rounds for the enemies directly from his editor. Our game engine only had to get his data from the editor in order to generate the levels.
He was also in charge for the management of the hero: life, equipment, etc.
Our game was quite pretty but it was very slow due to the too detailled models used for the hero and the enemies. It lacks animation too. I really wanted to incorporate animation but I wasn’t able to.
