Progress update for 5/8/2012 - 5/11/2012

This is the first of a series of posts to encourage me to keep up with this blog and my ongoing game education. In each, I'll catalog what I've done on the project for the period beginning on Saturday and ending on Friday (or in this case, beginning when I created the blog on the 8th).

This week:

I started this blog. I almost don't want to include this since I view it as more procrastination than getting things done, but I'm including it here under the assumption that it'll help me keep on top of things down the road and thus be productive.

Added bullets and shooting to the game.

Started an article on game design patterns.

Thought about how to handle different enemy types and what kinds of design patterns I could apply to make the code more readable.

Created a quadtree tech demo to see how practical it would be to use in a platformer game. Really, I was looking into spatial hashing in general for application in platformers and bullet hell games. I wrote a quadtree implementation in C# and dropped it into a windows form, then drew a variable number of squares and had them move up and down inside a panel to see how feasable it would be to do a large number of objects at once and update them all each frame. Pictures are below. All in all, I think that it's worth trying out, especially for games where there might not be as much stuff on screen at once. Here's a picture of the quadtree program running with 10 objects. Finding an object given a bounding rectangle is trivial in this example. Here's another picture with 10000 objects. Finding all objects colliding with a 1x1 rectangle under a mouse click took about 22 comparisons in the most subdivided areas and 12 in some of the bigger regions. The highest I ever saw the comparisons take was a little over 1 millisecond, returning with as many as 8 objects that intersected.

Amid all this pondering of design patterns, I decided to look up what XNA would be able to do for me. After trying a few different game engines/frameworks, I was wary, but XNA looks like it may actually be helpful...so I decided to scrap what I had and move as much of it as I could over to XNA 4.0 using the XNA 3.1 Platformer Starter Kit as a guide. I'll write a post on this process later, but I did manage to get back to where I was in terms of functionality (not far) within a few hours.

So what did I learn this week? Well, I studied up on design patterns and how I can apply them to my game, I learned that quadtrees are a powerful tool with an expensive insert time, and I changed from using a completely homebrewed engine living in .NET's Windows Forms to using XNA as a base to help build my engine. I feel kind of bad about jumping ship from my DIY approach and going to XNA, but frankly there's a lot of good reasons to go to XNA, not the least of which is that it provides that much more that I don't have to code myself.

See you next week.