ok, so maybe I’m a bit late to the party, but here’s a quick post mortem for my ld10 entry. I had limited internet access for this one, and missed being able to hang out on IRC and see what everyone was up to as things progressed, but it’s been great reading the blog posts retrospectively.
:: Design
The compo started at 4pm Saturday for me (New Zealand). I got home around 6pm, completely not intending to enter, but when I saw the theme was chain reactions I caved. I spent most of Saturday evening thinking about game ideas. The main contenders were:
- a 2 player board game with tiles which keep changing colour until the board reaches a stable state.
- a dominos bowling game with many many pins standing all over the place.
- a platform game with lots of crates full of fireworks.
The last one was by far the easiest to visualise, and to know that somehow it would turn out to be fun without oo much messing around. And when I saw that Dr. Petter didn’t have platformers covered for once, it was a no-brainer. :p
:: Tools
This is my 4th entry, and I’ve used the same tools each time:
sdl, sdl-mixer, bcc55, mspaint, gimp. Also this time, sfxr.
:: Implementation
The first step is always to get a black screen which responds to keypresses. I hate that part. I cheated a little this time, by copying snippets of code out of a past entry (sprout) to get the sdl initialisation and keypress code working. Knowing I did this is a tiny lead weight which my soul will carry forever.
Next up was running around a test map. I love this part. It doesn’t take much to get a guy running around 2d blocky landscape, but it’s instantly satisfying to start leaping around and exploring the limits of the world. I spent a lot of time tweaking the physics until it had a good feel to it.
The whole time I was doing this, I was imagining how the fireworks might work. I knew how the sparklers and rockets would be, and they took very little time to draw and code. The roman candles I wasn’t so sure about, and I ended up spending 5x as much time on them. Writing code and drawing gfx doesn’t seem to be the bottleneck — it’s just deciding exactly how things should look and behave.
I had just moved back from overseas when I did this compo, and in the midst of daylight savings confusion and lack of internet connectivity, thought I was finishing at 3pm instead of 4. I realised my mistake quite close to finishing, and was very happy to have the extra hour. I didn’t leave enough time for the vital task of designing maps, but was still able to try out what I imagined to be the defining moment of the game - being chased down a corridor of exploding crates.
I did the sound in about 5 minutes with sfxr. I would formally like to donate my audio points to Dr. Petter for that one. I just grabbed the first 5 randomised explosion sounds that came out.
:: What Went Right
- Not using placeholder graphics. They have a weird way of finding their way into the finished game.
- Being familiar with the tools. I could concentrate on designing stuff rather than fighting with technical issues.
- Making sure of the design before getting deep into implementation. I didn’t end up spending much time reworking code or graphics.
:: What Went Wrong
- Bad time distribution. Even 30 more minutes working on level design would have improved the game a lot.
- Not budgeting time for details. Little things like a nice txt file to go with the game, windowed mode support, hardly take any time but would have been worthwhile.
:: The Fucha
The problem with making 48h games is that I alway want to spend more time brushing up my entry and re-releasing it. Splode is no exception — I’m thinking to make a nicer 10-level arcadey version with online scores and recording at some stage. I’ll post here when it happens.