Happycake Development Notes:
(11 November 2005)
The project was finished back in April sometime. Because of requests for nondisclosure from our business partners, we didn't make any more log entries.
Overall the project went well. The networking came along fine (though
there were a few surprises), and the concept was proved. Here's a video
made in early 2005, showing the first time we got the giant robot into the game,
fully locomoting with destroyable joints. This was before we attached some
of the turrets, and you can see them fall off the robot at the very beginning of
the video. Later on all turrets were firing, separately destroyable and
(22 megabytes, WMV format)
(In a fully polished game that you were selling, though, you'd spend a lot of effort on making the robots look cool, making them fall down well, and balancing them better. For example, we wouldn't want this robot to fall down so easily -- in the demo video you take out one leg and it's immobilized. Rather, you'd want it to limp along. With the programmer-generated hexapod from the early development log entries, this was actually somewhat implemented. But it's harder to do when you have this thing with a weird weight distribution that an artist gives you. Given time we could have done it, but this was a short project.)
At the end of the project, through extensive toil, Atman made a much longer video showing the final version (fighting in the city setting that you see in previous updates, smoothed out animation for the character, etc etc). However this video is absolutely huge since it's 1280x720, 60 frames per second, high compression quality. (The goal being to show the demo off, it seemed like a good idea to produce high-quality video at HDTV resolution). Making this video was very painful, since there's apparently no hardware that will record such a video feed from a PC live (none that we could find out about, anyway), and frame-grabbing programs like FRAPS have basically no hope of capturing 1280x720x60Hz in real time. So we had to do it by journaling the input during a gameplay session, then playing it back in a special "video making mode" that dumps every frame to disk and thus runs very slowly. I hate writing input journaling systems for some reason, and that really showed this time -- the system was buggy and I failed to get all the bugs out within the time limit, so the playback would often diverge significantly from the recorded session. This gave Atman a lot of pain when he was trying to set up cool scenes for the video. I owe Atman a lot of beers for that. Sorry dude!
Because Atman's video is so big, I'm not hosting it here.
One of the nice things about working on this demo was that we have the (nonexclusive) right to re-use the code on arbitrary future projects. Parts of the engine from Happycake are already hard at work in my next project, which is entering the tail end of development now. (But you wouldn't know it from looking -- it's a 2D game with no physics and no shaders!)
Thanks to everyone who helped get the demo done, the whole Happycake team as well as the team from our business partner. This was a very interesting project to work on and I got a lot of good experience.
[Back to Happycake development log]