Network Trouble, Gameplay

I just received my second comment, only this one's legit. It wasn't long or detailed, but it was a complement, and that always counts for something, especially when it's all you've got. Anyway, it was from Zygote. He has a blog, Ziggyware XNA News and Tutorials, and a website, www.ziggyware.com. Thanks, Zygote.

On to the update:

I made some minor improvements and fixes to my input library. Nothing important enough to mention, especially since I'd have to explain how the whole library works first.

I also finally tested the network. It didn't work. I googled and googled, finding nothing, so I decided to test a sample game from the web, namely the Net Rumble Starter Kit. It didn't work either. Conclusion? Something's up with my network.

I've somehow completed four years toward an IT degree and still don't know a whole lot about networks. It's probably because I hated IT (I'm now CS) and blocked it all out of my mind. Anyway, I'll just have to figure it out.

I made some changes to the lobby and join screens. In the process, I think I've found some bugs in the beta. When I end a session, the SessionEnded event is never raised. This will cause problems on the client end. Nevertheless, I went into this knowing it was beta. They say they'll be releasing fairly soon, so I won't give up.

I've gotten as far as I can for now on the main menu. Now I'm just beginning to work on the gameplay (once again). Before I get too far, however, I'm brainstorming better ways to implement it. I already have a few ideas. I have to keep in mind the fact that it will be running over a network.

I'm also keeping scripting in mind. I said before that I'll implement it last, but I've had a few ideas for an interesting single-player with an AI that is purely made from scripts. The opponent would expand in interesting patterns from his base and the player would have to eliminate him within a time limit. If I'm going to have scripting, I may as well prepare for it now.

I'm going to redo my camera class. My previous camera class supported rotation, which was never used in Sever and probably never will be (as long as it's 2D; who knows what's in store?). Less math in the calculations means better performance.

I'm pretty sure I'm going to make each little person a Vector2 as opposed to a special class. The X value will be the person's progress through the route section and the Y value will be his current speed. Structs are faster than objects, so the performance should be improved.

Also, I'm thinking of hiding opponents' people. Each player will only be able to see his own people. This sounds arbitrary, but it means that I won't have to send data about each little person over the network. Much much less traffic. It also means less sprites to display. Can you go wrong with better performance all around?

I'm also going to look into geometric collision detection and fog-of-war. Geometric collision detection is important because it will allow me to places obstacles in a map. If I didn't have those, then what would distinguish one map from another aside from tiling?

Fog-of-war is important for two reasons. First, I think it will add to the fun. It will allow players to scheme in private. It will also allow players to move around the map without the opponent noticing, allowing for more complex strategies, such as sneaking around and attacking from the side. Without fog-of-war, the game would be brute-force. Second, I think I would enjoy implementing it. It would be a real challenge, but I think it would be worth it.

I'm planning to post some gameplay videos from my last Sever implementation. I don't know why I didn't do that before.

Suggestions? Questions?

clevceo