For #19. I'm not completely happy with the snowballs algorithm (primarily due to it's tendency to sometimes generate deep 1x1 holes), but it seems to function well enough that we can merge it. Perhaps a second opinion is just what we need.
On a separate note, I would like to implement a 3d algorithm at some point too. Perhaps a wind-based algorithm perhaps?
Maybe we could have a system where a given node type is considered an emitter, and we emit wind particles that float around which we simulate instead of snowballs. Each wind particle would be given a random initial starting position (that's inside a random emitter node) and 3D velocity vector, and we'd move from there. I'm unsure as to how it would interact with surfaces though.
Would it be redirected away from nodes?
What would happen to the particle's speed?
We probably would want to just delete it, because we'd then be essentially simulating something like a TNT blast for example.
We'd probably want it to continue somehow, how the question of how it would do this in a fashion that it could continue to erode nodes effectively is a difficult one.
As much as we may dislike doing it, we'll probably need to calculate normals (i.e. a measure of the slope direction of a surface) for that to work. Unlike the snowballs algorithm, we'll want to do the calculations dynamically here for the algorithm to be effective.
I've thought of using the player as a source for wind particles instead, but I'd prefer it if the command wasn't tied to the player to enable more effective use in scripts etc.
A constant angle direction could work too, but I'm unsure.
For #19. I'm not completely happy with the snowballs algorithm (primarily due to it's tendency to sometimes generate deep 1x1 holes), but it seems to function well enough that we can merge it. Perhaps a second opinion is just what we need.
On a separate note, I would like to implement a 3d algorithm at some point too. Perhaps a wind-based algorithm perhaps?
Maybe we could have a system where a given node type is considered an emitter, and we emit wind particles that float around which we simulate instead of snowballs. Each wind particle would be given a random initial starting position (that's inside a random emitter node) and 3D velocity vector, and we'd move from there. I'm unsure as to how it would interact with surfaces though.
snowballsalgorithm, we'll want to do the calculations dynamically here for the algorithm to be effective.Thoughts are welcome :-)