It's almost the end of the first month of 2016, and you may have noticed that we still haven't released anything. We are getting closer, though! At the end of our last post we mentioned that a crash bug which came up before Christmas was preventing us from releasing R.A.B.B.i.T., and that the release of an advertising plug-in meant we could finally get Defendy Rocket out. Well! We spoke too soon; we bought an Alcatel Pixi, which was one of the devices being affected by the crash bug, and discovered that it was also affecting Defendy Rocket. So, with our plans to release anything at all hinging on a fix for this bug, we decided to focus on R.A.B.B.i.T. first.
The main thing we've achieved over this month was to port almost the entire game over to native code, and run with a version of Unreal Engine that we built ourselves. The latter means that if Epic have a fix for our issue we can grab the code from Github and build a new version of the engine to test it immediately, without waiting for a new release, which usually takes a couple of weeks at best. The former means that the game should run considerably faster, without the processing overheads that come from systems created entirely in Blueprint. It should also result in a smaller file size, as the code versions of our classes have turned out to be a lot more compact than the Blueprint versions.
We also got the chance to implement and fix a number of things that we had had to drop when we were trying to push the game out before Christmas. The game now has title screen music, new sound effects and jingles for various parts of the game and front-end, and generally improved buttons and graphics throughout the game's menu. We also finalised the game's icon, as the one I'd made before Christmas wasn't really cutting it. It now looks like this:
The tutorial system and game flow have also recently come under scrutiny, as we didn't feel like people were getting the hang of how to play the game quickly enough when they were stuck playing tutorial levels over and over. Watching them play Survival mode, however, which generates different level layouts every time, we noticed that they seemed to learn how to deal with different situations much quicker. To this end, we've made the tutorial considerably shorter, just three levels instead of eight, and completing the tutorial now unlocks only Survival mode. Getting a certain score in Survival mode will unlock Sprint mode; getting a certain score in Sprint will unlock Challenges mode (Which is the new name for Missions); completing a number of missions will then unlock Custom mode. We feel that staggering the unlocks of the various modes in this way will encourage players to develop the required skills a lot quicker than if they are banging their heads against the same levels repeatedly.
Finally, while discussing ways of making the tutorial work, we decided that adding stars to collect would offer a light challenge, and also encourage exploration of different ways of moving around the levels. At this point we figured, 'Why stop there?', and went ahead and added the stars to Sprint mode and the Challenge levels. In Sprint mode, stars will appear when you are low on time (one of the things we disliked about this mode was that it was possible to start a level with insufficient time with which to complete it, meaning you have to just let it run out) and give you a 2.5 second bonus when collected. The position of these stars needed to be procedurally generated, so we had to develop rules for placing them. Challenge levels will have five stars each, this time placed by hand. We're hoping this will add an optional layer of challenge to these levels, for when just completing them becomes trivial for you!
To finish, here's a GIF of what the stars look like in-game: