“Bunnies, Back Into Your Cage!” was my second Big Ludum Dare game, so I didn’t expect anything surprising here. Of course I was wrong.
WHAT WENT WRONG?
1) Idea and tools
I began my 43 hours of game development (I woke up ca. 5 hours after the theme was announced) without any idea. I already knew I wanted to use Unity3D though (because I use it at work and made my last LD game with it), so I tried to make my concept fitting a 3D engine. After some time I had the idea of a guy running faster and faster as long as he doesn’t hit anything concrete. If you run fast enough (e.g. after running in circles quite some time), you could break certain walls and destroy enemies. Objects would have different values to which speed they react.
So I started Unity3D and did a small prototype. One of the first problems here were the physics though. Smashing through walls only looks good with falling and tumbling pieces of trash, but physics aren’t much controllable. As soon as the player runs into something, he becomes slower. When there are many bricks (sized 2*1*1 meter or so) lying around, the player becomes slower constantly. Which is tedious and not much fun.
The lesson here: Don’t add physics unless they are part of the core gameplay.
2) Motivation
In the end, the prototype just didn’t make any fun even after some hours of work, so I started to distract myself with things on the web and IRC. Motivation was completely vanished – the fact that the theme “Escape” wasn’t really to my liking didn’t help much either. But as soon as I realized that I was more procrastinating than working on the game, I stopped. Grudgingly, I closed Unity3D.
Even though this was bad at least it led me to using Flash, with which I made my final entry: “Bunnies, Back Into Your Cage!” – a game about capturing bunnies that escaped.
3) Again, theme
My second idea didn’t fit the theme any time. I had the idea of a really minimalistic Minecraft which I would call “Pixelcraft”. From there, I began to write code, with more motivation, and at some point, the prototype was playable, and even a little bit fun.
Of course, it was missing the “Escape”. 
4) Controls
The player controls were wonky from the beginning, and still are wonky. “Tedious” is the adjective I hear most as feedback to my game.
This is because everything moves/works block-wise. A block was originally meant to be pixel-sized, so this is why I chose to do the movement this way. Now, to add a little bit detail, the blocks are 8*8 pixels (they really are 32*32 pixels, but you can’t see that, hehe), and the controls feel strange. It feels like you often can’t direct the protagonist where you want him to be. Of course, I have no problems with the control scheme, but I am the creator, so this is expected.
The lesson here: Take the time to let your game test by others. They will crush your ego by pointing out all the bugs, but it will be worth in the end. I think. (Of course, the *real* problem is that in the Ludum Dare IRC channel were 330 people who all needed game testers. Try another channel or your girlfriend then. ;))
5) Being not Notch
With ~600 entries, not being the creator of Minecraft can draw some of the attention off your entry. You have to compensate it with marketing. And like most people, I hate doing marketing.
Lesson learned: Be notch. Top notch!
6) inudge
I like inudge, but the music sounds just like the one from my last entry. Also, there still is no way to export to WAV or MP3, so I had to record the tune with Audacity. But whatever recording method I tried, it either didn’t record or it all sounded strange and noisy. Even other recording tools made it sound like some horrible sound experiment.
The lesson learned: Sometimes, just rebooting your computer can help. It’s some kind of magic.
WHAT WENT RIGHT?
1) Flash. And FlashPunk.
I just began to love this combination, because you get results so fast. I started my second try at 18 o’clock, so I only had 34 hours left – sleeping included. Yet I came around to make full game, with levels. Cool!
2) FlashDevelop as a level editor.
This is a screenshot of my “level edtor”.
As you can see, you can edit not only rows, but also columns with FlashDevelop. This made it much easier to edit my ASCII array level definitions. Also, I am happy that I decided not to dwell into level definition files, but just hardcode the whole thing. I got levels pretty late, about 8 hours before submission, and yet it didn’t feel like “too late”.
3) Pixel art.
I often hear “nice art” or “I love the graphics” for my entry. This baffles me – everything is just 8*8 pixels. It was a piece of cake to do the graphics. I don’t even have animations, or different frames for walking left/right. OK, color selection may be a factor, but that isn’t hard to get right, either. Just choose colors with nearly the same saturation. Bang, instant good-looking art!
The lesson here: Minimalism is for teh win!
…
That’s it for now. Thanks for reading, and don’t forget to play my game! 
Tags: flash, LD21, post-mortem, postmortem, screenshot
Thanks!
