Experimenting with a custom surface tension algorithm. The solver is FLIP based, and the surface tension uses both implicit and explicit components of the data. It is not the most efficient way to simulate water drops, but I learn a lot (and had fun) while working on it.