Post-mortem – Leave me alone

Posted by (twitter: @daredevildave)
December 21st, 2011 5:29 am

Leave me alone screenshot

Here’s the actual game page.

What went well?

1. Preparation

I spend the two weeks leading up to LD22 learning how to make simple models in Blender. I’ve never had much more than a passing knowledge of Maya or Max ( I could operate them, but not build anything ). Despite making games professionally for 10 years there had always been an air of mystery about how artists actually realised things in these packages. Turns out there is no mystery at all. It’s really straightforward, that’s not to say I’m a good artist now :-). Anyway, having that experience meant, I knew what I could build and how to build it. So there was very little learning in the 48hrs, I could just get on with it.

2. Knowing the engine

I wrote a significant chunk of the PlayCanvas Engine, so I know it pretty well inside out. Not only that, but the engine is written in javascript, so I could step into it and debug from game-code into engine-code with no added steps. It’s incredibly helpful not to be using a ‘black box’ engine, you can always see exactly what is going on.

3. Tools

I’m very big on tooling for game development, and I think it’s critical to have great tools. The PlayCanvas tools did well despite being early in development. Having simple features like Undo/Redo is such a time-saver.

4. Rapid deployment

I could deploy my game in a couple of seconds and have it ready-to-play by anyone. If there had been an extended packing and publishing stage for me to complete I probably wouldn’t have got it out by the deadline.

What didn’t go well?

1. Physics and Collision Detection

Collision detection and physics is hard, not only is it hard to comprehend, but it takes long time to debug and get stable and working. Trying to do 3D collision detection and dynamics in a 48 game jam isn’t very sensible. Still I got it mostly working. I had to remove some bits where you could jump gaps with the ball because the physics wasn’t stable enough. Next time I’ll either do something without complicated physics, or hope that there is an existing stable implementation. If you’re interested I used verlet integration.

2. Tools

Tools are so important that when they have issues, it can really slow you up. Some features of the tools were slow, and it cost me a lot of time to make small changes when it should have been quick. But on the plus side, I now have a big list of bugs to fix.

3. Audio

No time, no skills, no audio. :-( I really wanted to go out and record some grinding stone noises to add some flavour to the movement, but I just never had the time.

4. Timelapse

I wanted to record a timelapse, I did the first day, and them somehow failed to start recording on the second day, dammit!


This was my first Ludum Dare, though I followed the last one with interest. It was a fantastic experience, I’ll certainly be doing it again, and next time I’ll be even better. Sitting in on the irc channel was also really fun. This is a great community of people sharing, helping and really pushing forward games development.

Also this was the first real test of our HTML5 engine PlayCanvas, and I think it did pretty well. If you’re interested in trying it out I’d love to help you, drop me a mail

Leave a Reply

You must be logged in to post a comment.

[cache: storing page]