Basically you're right for crystals -- the planes of fragmentation will be self-similar to the molecular structure. But most of the crust is not diamond and beryl, so this study is looking at the average structure of much larger conglomerates of stone and ice and mud -- not as much long-range symmetry there!
Quoting the paper to give an idea of the formation being modeled:
These patterns have been reproduced in experiments of mud and corn-starch cracks, model 2D fragmentation systems, where the following have been observed: Fast drying produces strong tension that drives the formation of primary (global) cracks that criss-cross the sample and make “X” junctions; slow drying allows the formation of secondary cracks that terminate at “T” junctions; and “T” junctions rearrange into “Y” junctions to either maximize energy release as cracks penetrate the bulk or during reopening–healing cycles from wetting/drying
I studied a metallurgy subject that focussed on crystals as a bit of fillin in my electrical engineering degree some 35 years ago. The thing is that large scale macro features like crystalline surfaces and grains are driven by atomic structure that was first deduced through x-ray imaging and then like.
Quoting the paper to give an idea of the formation being modeled:
These patterns have been reproduced in experiments of mud and corn-starch cracks, model 2D fragmentation systems, where the following have been observed: Fast drying produces strong tension that drives the formation of primary (global) cracks that criss-cross the sample and make “X” junctions; slow drying allows the formation of secondary cracks that terminate at “T” junctions; and “T” junctions rearrange into “Y” junctions to either maximize energy release as cracks penetrate the bulk or during reopening–healing cycles from wetting/drying