One day, the grokalow crawled out of the swamp.
“It looks like a gargantuan alligator”, cried a bystander, as it approached.
“Nay, it is a brobdingnagian crocodile”, countered a second, as the grokalow licked its leathery lips.
“Without a clear definition, we cannot conclude this thing is a threat”, surmised the third, settling the matter.
The grokalow ate them all, with great satisfaction.
When I spoke about autotiling, I briefly touched on how it’s possible to use Wave Function Collapse (or other constraint based generators) as a form of autotiling, i.e. user-directed editing of tilemaps.
I’ve usually referred to this technique as “editable WFC“. It’s a combination of autotiling and WFC, and contains the best of both:
- Being an autotiler, it allows users to easily and interactively make changes to an existing level.
- Being constraint based, it automatically ensures that those changes are consistent with the predefined rules of the constraints, potentially making further changes to the level to make it fit
This is different from most other autotilers, which either require manual configuration of patterns used to enforce good behaviour, hidden layers, or come with more stringent requirements on what tiles are available.
Back in 2017, I described a method of random path generation called Chiseling. It gives very nice wiggly paths, but I was never satisifed with the performance. I later revisited it, and found a faster algorithm, but it was a bit complicated to implement.
I’m pleased to say that I think I’ve finally found a way of implementing it that is both fast and simple.
One issue with 2d Marching Cubes (i.e. Marching Squres) that I glossed over in my original tutorial is the issue of ambiguity. Later I talk about how dual contouring avoids this problem. But it turns out there’s actually a well known, but little documented way of resolving those ambiguities called the Asymptotic Decider, which I’ll explain below.