Ludum Dare 27 has been my first ever Ludum Dare competition.
I was not sure of what was going to happen, if I were capable of finishing a game in under 48 hours, if I would have felt stressed or relaxed, et cetera.
For the competition, I decided to use C++11, SFML, and my own framework, SSV, which is free, open-source, and always looking for contributions/critique.
My development environment was Arch Linux x64, using QTCreator as my IDE, and Sublime Text 3 as my text editor.
The development machine uses an Intel Core i7 processor, NVidia GTX275 and 10GB of DDR3 RAM.
My goal was producing a game that was worth playing in under 48 hours, with native Win32 and native GNU/Linux x86 binaries.
I’m very happy to have reached that goal, and I’d like to share my thoughts about the whole development process.
I worked on the game for about 30-32 hours. I slept, worked on a video for a friend’s birthday, and relaxed for 1-2 hours (played some Spelunky and browsed the internet).
The first thing that surprised me is that I felt constantly stressed. I do not know if everyone feels like this, but I couldn’t stop thinking about the deadline, about the end result.
I have to say that, as far as personal feelings go, I didn’t dislike the 48-hour deadline development process, but I didn’t find it fun either.
However, after finishing, I felt a great sense of satisfaction and reward, which kind of made up for the stressful coding hours.
The second thing that surprised me is that my framework, the SSV framework, was up for the task of creating a game from scratch.
It literally took less than an hour to get a prototype where I could walk around.
A big effort in SSV was put into development of SSVSCollision, a header-only pseudo-physics library intended for retro-style games.
It handles collisions very differently from all other engines out there, and, while not suitable for realistic physics simulations, it is great for retro-style games, where physical bodies do not interact much with each other, but have infinite stability and very precise collision resolution. Here’s a video of it’s performance.
It also lacks all sort of issues that would arise with a realistic physics engine, such as the common error where bodies get stuck between tiles in tile-based worlds.
Anyway, I also created a player sprite, which I had to divide in two parts (arms and body) to avoid repeating unnecessary frames. I used Pinta for the task, a Paint.NET clone for GNU/Linux.
I’m not an artist, and that is obvious by looking at the poor end result of the player sprite. I used the same tool to create all other graphics in-game.
I dealt with sounds by using sfxr, the free, open-source sound generator advertised on the LD website itself.
For music, I used LMMS, a GNU/Linux production software with an UI similar to Fruity Loops. I’m not a musician either, so the end result was poor here too.
The game concept was actually created after the prototype version. I had no idea what I was going to make. I just made stuff and tested stuff.
Then I had the idea of this cool throwing mechanic, where suddenly turning your character would increase the force of the throw.
This is where stuff started getting interesting. I had to deal with my peculiar physics engine in order to allow the player to grab/throw/release blocks.
It went pretty smoothly.
This is what the first grabbing prototype looked like. I also had added the number on the crates but had no idea how to use those yet.
I also had no idea how to use the crates yet.
Then I combined the throwing concept/mechanic with a time-based constraint (10 seconds theme), and had the idea to make the game into a reflex-based, time-based puzzle platformer.
I designed some game elements and threw some test levels together. But I didn’t have time to create a level editor, or to write a JSON level specification. So what did I do?
Tab based in-code level editing. Dear god.
Yep, I used tabs, newlines and spacing to re-create the structure of the level in the IDE itself, so that I could have a rough idea of where I was placing elements.
After making some levels, I created a menu screen, which was very easy thanks to the SSVMenuSystem module of the SSV framework. And that’s pretty much it.
There is a problem with level 5, which is almost impossible because I forgot a game element. But it is actually possible, even if insanely hard.
I’ll judge the game myself, now:
Innovation: I’d say the game is not unoriginal. The turn-based/jump-based throwing mechanic is pretty fun to use, and the game elements, while simple on their own, can be combined to create some interesting puzzles.
Fun: This is a very subjective point. The game is not easy, and can be very frustrating at times. Honestly, I find hardcore games pretty fun – I enjoyed playing my game, even if trying the fifth level for one-hundred times got frustrating quickly.
Theme: My interpretation of the theme is not very original, but I think the 10-seconds constraint that resets works well here.
Graphics: I’m not an artist, and it really shows. The sprites are of poor quality. I tried to redeem myself by creating variations of tiles that appear randomly and maintaining a simple flat look for the game.
Audio: Sfxr is a godsend. I love retro sound effects, and they work well here, I think. Music, on the other hand, is not catchy or memorable, and it’s just a simple loop. It was my first time ever producing music. Here I tried to redeem myself by adding a no-sound and a no-music option to the main menu.
Mood: I tried to create a simple story/world around the game. Basically, you’re working for this company, 10corp, in a futuristic (I guess) setting where getting a job is very hard. In order to survive, you have to work for this company, even if they terminate slow workers to maximize their profits. I used in-game messages (broadcasts from 10corp) to give the feeling of the player being observed and judged during its tasks.
Overall: Overall, I am satisfied with the end result. I’m still not sure if the game is worth improving, but as a less-than-48-hour product, I’m happy with how it came out.
I really hope you enjoyed my entry and this postmortem. Thanks for reading!