When I saw the theme list on the final day of voting, I was rooting for one of the stealthy themes. They didn’t make it, but escape allowed me to work this idea I had for the past week into a game. I wanted to try and create a top-down game with true line of sight. The end result is Abstraball, a game with true line of sight, so I guess I succeeded with that. But it could’ve been so much better if I changed some things at the start.<\/p>\n
Day 1<\/strong><\/p>\n <\/strong>I did my usual routine for making an XNA game. Create a simple screen system, a world class, entities, and fix up drawing things. This time around I wanted to use polygons instead of using the spritebatch, and I wanted to make variable sized geometry rather than tiles. So I worked on that. The camera would be orthographic instead of perspective, and not being limited to square tiles would mean I could make simple big geometry for large portions, and resort to more complex (read: diagonal walls!) shapes when I wanted fancy things. After doing that, I worked on the shadow effect. First I made the area next to and behind you black. Then I added the effect on geometry. And that’s when I should’ve made a different decision. An orthographic view with ‘shadows’ like that is practically the same as a perspective view with pillars reaching to infinity, but without the texture support, and the without ability to easily prioritize what should be drawn. At first I didn’t realize that, but now I do, and I definitely would’ve done it differently if I had known at that point.<\/p>\n Anyway, I continued with that, thinking I could make it look better later on. I started working on the map. I had the ability to make any shape I want, to make any complex level I wanted. But I had no level editor, and placing things by hand was a pain. So I resorted to creating a bitmap, and importing\/parsing that. So at this point I had fancy, unstructured polygons shaped as square tiles, without the benefits of a tile grid. That was mistake #2. I should’ve made it tiled, then I could index things more easily and efficiently, and also easily make a bit more diversity with other shaped walls. But I didn’t. Pathing the world was tiled, but they didn’t support each other. I also had to make some optimizations so unnecessary geometry wouldn’t be added.<\/p>\n Day 2<\/strong><\/p>\n At this point I started to realize the mistakes I made during day 1. I lost some motivation, but kept on going anyway. I worked on the AI, and after a few hours I got a nice and robust AI system that got the job done very nicely. But again, I had optimization issues because of the way I built the world. So I had to cut on a few things to make it playable. Pathfinding was just bad, and occasionally it would take a second to calculate a path. Line of sight calculations were bad as well, and many other things were bad too. But I worked around most of them. Next to those things, I added the footstep sounds, and the positioning for it. So by now I sort of had a playable game. I named it abstraball because I knew I couldn’t fix the graphics, so maybe coining it as an abstract game would help.<\/p>\n With about 8 hours left I started talking to a friend. He playtested my game and gave a bunch of tips for features and gameplay changes. I added difficulty modes, tweaked the AI, added a compass, the energy bar, fixed a few glitches, added the plague bearer mode, and did some other minor things. He wants to be credited for being awesome, which he is, so his name is Final thoughts<\/strong><\/p>\n While I’m not that happy with my game (it could’ve been much better), I still greatly enjoyed creating and playing it<\/strong>. I feel my previous (and first) entry of LD20 is a lot better overall, but during the creation of this game I learned a lot of new things and tricks, which I am sure to use next time. I’m going to stay away from the spritebatch completely now, as I now have a relatively good understanding on how to work with primitives. For the next time I’m going to roll out my own engine\/framework\/toolset\/whatever to prevent having to search for and redo basic things, and to do it right. The current code is such a mess, it’s just plain terrible. I broke so many best practices because of laziness, and while it’s part of LD, it still feels terrible for doing so. \ud83d\ude00<\/p>\n In any case, I really enjoyed this LD and I’ll participate next time as well! I’m also going to attempt to do the october challenge as a motivator and deadline to finally do something with a project I’ve been working on and off for over the past year.<\/p>","time":"August 23rd, 2011 10:15 am","title":"Abstraball postmortem"}<\/del>savooipeerd.<\/strong> With about 4 hours left I added a few final touches. I composed the music with iNudge in about 10 minutes, added keyboard configuration options, and started to playtest some more for balancing changes. I fixed a few annoyances, and with an hour or so left I was done.<\/p>\n