Again the chosen theme wasn't the one I would have liked the most. Instead the winning theme was "Light & Darkness", which I felt as dull and not really restrictive enough to give extra thought for creativity.
It didn't take me too long to pick up a game idea to work with. I set out to make game in style of shooting gallery or the classic Operation Wolf, where player is supposed to shoot out all the lights. There wasn't much technical challenge in this, so I figured out that this time I will try to think about gameplay before doing finishing graphical touches. I decided to use tile-based layout for the graphics, and I used tool called Mappy for laying out the levels.
When building levels it turned out that for making the game even remotely interesting enough, there has to be sufficient amount of different content. So eventually I end up using quite much of the time to work with the graphics. This led to the fact that I didn't have time to implement more rich gameplay features such as moving or animated targets. However, I still tried to stuck with the decision of focusing on gameplay first. So, for example, I created the level state logic before doing sound and graphical touches like particle systems.
Still most of the levels and graphics in final version were done as the last thing, without enough energy or time to do proper playtesting to tune difficulty curve. While it generally gets harder in pretty reasonable way for the first level, turns out that at least a few players complain that e.g. level 8 seems to be a bit hard compared to levels before and after it.
In retrospect the chosen idea was bad in the sense that it required a moderate amount of work with graphics. Some theme with less such work with assets would have been better choice for such quick game project.
Working with the sound and music was a nice diversion from the other work. I had fun time making sound effects for the breaking lights -- I borrowed a kettle and handful of spoons from kitchen, and kept on dropping the spoons to the kettle to create a few sounds resembling shattering glass. :) I created a few small pieces of music in one module with Impulse Tracker, only to find out that with SDL_mixer (and its embedded MikMod) I couldn't properly get the module loop inside it with the standard looping commands. This time I also tried to pick better quality samples but the sound quality was still not that good in the game, which I figured out being a fault of either SDL sound, SDL_mixer or MikMod. I decided to drop those and switched over to FMOD, which solved both the looping and sound quality problems.
Sticking to libraries and tools you already know is good in projects like this. Again I used SDL libraries and the Turska UI framework. When switching the audio engine from SDL_mixer I first thought about using Audiere, but I haven't used it before so I decided to use FMOD instead which I was already familiar with. I'll try Audiere later with some other project.
I estimate that I used about 22 hours for actual development of the game, not counting sleeping, eating and other pauses.
What went right: