Alicornutopia

Barriers are perfectly straight, and almost never perfectly vertical. Okay, attempting to explain the distribution of siles without resorting to telling you to google "3d voronoi cells"...

Imagine that you have a bunch of squishable foam balls in varying sizes, and you pack them into a huge container as hard as possible, so that they're all crushed together and flattened against each other. Then refine/simplify them all into convex irregular polyhedra, so that each pair of adjacent ones shares exactly one side. Now fill them all with air, and add dirt/rock/water to most of them. That's approximately what siles are like before life starts getting involved. Except the "huge container" is actually infinitely large.

Oh, ugh, sorry again - you said voronoi previously in this thread and I think I looked it up then but completely forgot. x_x

So you could dig down from one, go through the barrier at the bottom, and fall out of the sky into the one below? And you could possibly get a tetrahedral cell?

Yes and yes, although tetrahedral cells are probably kind of rare because it would have to be unusually low on neighbours and I'm not sure how that comes about.

And going up is probably even more dangerous than going down because you might find yourself underground.

I think sile is starting to seem like a good word for this.

Do you have any idea how the seed points for the Voronoi fractures are chosen? Are they just arranged uniformly and independently at random with each cubic mile or something having an expected one point?

Probably tetrahedral happens to siles that are much smaller than the others around?

Yeah. Tetrahedral siles are likely to be filled, I think, for size reasons.

I think they are less uniform than you're describing, DanielH, and they might also be less random, but you probably wouldn't go too far wrong imagining it like so.

Just because something is made uniformly at random does not mean it looks uniform or random. Both very small and very large siles would still happen. I expect you know that, but it is worth pointing out that the terms can be intuitively misleading.

If they are less uniform and random, would they be more clustered (to make more filled siles)? Less? More aligned in planes so that siles have mostly-vertical walls?

Perhaps I wasn't clear what you meant by "uniformly at random", then.

Among other things, I believe that there are sometimes irregularly shaped 'walls' of contiguous filled siles.

I'm dissatisfied with the term "filled sile". Filled with what? I keep accidentally picturing one filled with rock or dirt.

I would prefer, say, "solid sile".

"Solid sile" works too.