I didn’t make very much progress today, especially not as much as I’d hoped. I needed some offline time, so less development happened.
I’d like an overview of a much larger map to test terrain generation algorithms on, so I decided to make something that could show me much larger maps. Cocos2d is just too slow with this many sprites being displayed, so I decided to write a simpler version using PIL (actually Pillow).
Because I wanted to keep the chunk size the same, but have a much larger viewport, my chunk generation code doesn’t work properly anymore, it generates the 9 chunks centered in the middle and that’s it. Getting chunk generation to fill the a much larger took more work than I’d anticipated due to bugs in my existing chunk generation, which I’d have needed to fix if I end up wanting a full screen view, rather than a windowed view.
So I experimented with hex tiling, finding one that worked really nicely, but means that the center hex is no longer in the center of the chunk. If I decide to go forward with this, I might start indexing the hexes based off of their top left hex.
Today sort of felt like a bit of Yak Shaving, after making good progress on other areas. But I want the terrain to be more than a call to randint(), and getting the chunks nailed down will allow me to move on to proper procedural generation of the terrain.
At the end of the day, I hadn’t even got the drawing working for this properly. Chunk generation wasn’t working properly, so it was commented out. But at least I could draw ~56k hexes in under a second. Once I get this working properly, I can use it to refine my procedural generation algorithms for the terrain, which will include enemy spawns (enemy controlled cores) and varied terrain types. I’d like the terrain hex colours to actually mean something by using them to indicate different types of terrain.
I’ve committed the work in progress code to a branch.