GBGames

As I noted in my last post on Oracle’s Eye development, I wanted to transfer functionality from PlayState to the GameWorld class.

Short story: I did it. Next step is to try to make the Player object obey the WallTile boundaries.

Long story: There were multiple times when I felt like I was spinning my wheels. Maybe it was just that I haven’t been exercising my brain by designing as much as I should, but I was starting to despair that I wouldn’t be able to figure out what to do. After all, I already had a huge chunk of temporary code to see how the program would work, and here I was trying to completely change the design of it. Where do I begin? Will I ruin what I have and waste time before I figure it out?

After some persistence, I created a number of new classes, modified a few existing ones, and I molded the GameWorld class into exactly what I needed. It just all came together. For hours I was struggling with thinking about how to handle the changes I needed. At one point I even considered reducing the scope of the project so that the stick figure just had to reach an exit in a maze since it seemed simpler.

When I started this last development session, I found it difficult to imagine that I would be able to implement the complete game. Even after a few hours of “work” (in which I might have spent a big portion of it not actually concentrating on the task at hand since it was easier), I still had doubts.

Then everything just worked out, and now I can see that my original plans should be possible. Next time, even if I feel overwhelmed with my project, I’ll just have to remember that going the extra mile and being persistent will do more for me than worrying or massive redesigns would do.

I don’t have anything really newsworthy to share since last time, but I thought I should update the blog on my Oracle’s Eye progress.

I had a stick figure that would glide around, and three others that would move at different speeds and in different directions relative to the one you control. The Player objects were hardcoded into the PlayState as global variables, and I named the variables gTemp_player to remind myself that these are meant to be temporary.

I decided that I needed to create a GameWorld class to handle all of the entities and their interactions. I set out to design the class, and I found that it would have to own the Room and the Player. It makes it very easy to have a GameWorld class that determines how they interact because otherwise the PlayState would know too much and it would add way too much complexity. The cool thing is that keypresses can be handled by delegating to the GameWorld object. For instance, if I press the left arrow key, PlayState will simply tell the GameWorld to moveLeft(). The GameWorld can decide what it means for it to move to the left, but it will likely just move the Player object in that direction.

Great. Now I just need a Room class…and a Room class needs a Tile class. I already know that I want different types of tiles, so I created FloorTile and WallTile classes that derive from a base Tile class.

In the end, I created all of these classes, but there wasn’t enough time in the development session to actually get them all to work together. I’m still hardcoding PlayState, which is fine for now. Since I can’t just create a Room with Tiles easily and I want to see the tile sprites, I just used a Player object that won’t move and used the tile sprites as the sprite for the image.

It actually didn’t look too bad:

Technically all of the extra code didn’t actually do much to change what someone would see when you run the program, but I think the foundation is ready for me to make significant progress. Tonight is another scheduled night to work on it, so it will still be fresh in my mind.

My next step will be to transfer functionality from PlayState to the GameWorld class. I will then attempt to force the Player to walk only within the boundaries of a Room, probably by just having GameWorld check if Player’s position would intersect with WallTiles. FloorTiles will be in the background, and Kyra makes it easy to check only the sprites I need.

In response to my post on Creative Writing Prompts and my search for a similar tool for programmers, scott posted a link to Code Kata at Dave Thomas’ (The Pragmatic Programmer) blog, PragDave.

There are a few posts describing Code Kata, but one of them mentions the Dreyfus model for skills acquisition.

The Dreyfus model suggests that there are five stages in the acquisition of mastery. We start at novice: unsure and with no experience. We don’t particularly want to know the “why,” we just want to be shown what to do. We need to know the rules, because we just want to achieve some goal.

While Code Kata is for learning how to program in general, I think the skills acquisition description was very insightful and can apply to learning game development. I think I had posted previously about how beginners should start by making simple games because doing so will teach them a lot about what it takes to make more complex games later. Of course, I was mainly talking about myself as the beginner game developer.

I already knew that I needed more game development experience, but I need to realize that it is not just OK to make a clone for my first game or four. It might be necessary. As PragDave mentions, it is like learning karate. If you’re a beginner game developer, you should be able to emulate existing game developers before you start trying to work on your own plans. You can’t start by being the Master. You have to start as a beginner by emulating the Master.

People on GameDev.net periodically find that they need to inform idealistic newbies that they won’t be able to make the best MMO game ever. “Try something simpler, like a Tetris clone” or “You need to learn how to program” or “/me slaps you with a trout” might be common responses. The newbie is a student of game development who thinks he/she can skip the beginner and intermediate phases and start making games like a master. The newbie needs to learn how development works first. He/she needs to learn the rules. It’s learning the “How” before the “Why”.

I have a number of ideas that I think are somewhat unique, and I would love to make games out of those ideas. At the same time, I know that if I try to make those games today they won’t turn out nearly as good as I would want. I can, however, make simple clones, and I’ve already made a Pac-man clone in QBasic years before, complete with bouncing cherries and multiple levels, so I know I can do it. Perhaps a Space Invaders or Asteroids clone wouldn’t be such a bad idea after all. For a long time I thought that making clones of existing classics would be wrong in some way. I could do better and make something unique for my first few projects, right? I mean, does the world need another Pong clone?

Well, perhaps it does.

I’ll still try to finish Oracle’s Eye first since I think that finishing a game project is a good habit to gain. My next projects will likely be some simple clones which will be easier to complete in a month.

Paul Taylor wrote about working on an indie game called Determinance, and he is exactly the kind of musician I would want to have working on any of my own games:

Each track, even the menu music and closing theme, is based around a core melody or chorus. This is something I believe in passionately – if you can’t whistle the tune of a song then it’s 99.9% likely to be worthless.

Wouldn’t you want a person who is passionate about what he does making your games?