Interestingly, there's even the mid division on some planar faces even with ngon export as can be seen here. Must be something I'm not understanding about how the algorithm is handling things.

Last one, I promise. Sorry for all the messages. Another interesting behaviour is that it doesn't do it consistently to "similar" faces. For instance, in this slightly more complex shape than the first: notice the blue surface is triangulated, but the red one isn't. I like the red one ;) Though it still has the midpoint subdivision as above.

Part of this feels like a sort of hidden "divide larger than" limit. At least the midpoint division thing...sorta...

Hi MK, you're getting that result because the front face cap is not actually a simple surface, it's a surface with an interior hole. That means that there is an inner trim boundary in addition to the outer trim boundary.

Most polygon mesh file formats (except SketchUp .skp) do not have this same concept of an interior "floating" hole inside the polygon not directly attached by an edge to the outer boundary. So when MoI encounters this type of situation where it ends up with an n-gon with an interior hole it will divide the n-gon in half right through the hole so that it doesn't have an internal floating hole anymore it will have a bridge between the outer boundary and the inner hole boundary making just one larger n-gon.

Your sweep case produces a different result because it doesn't have a surface with an interior hole in it anymore, it has all simple surfaces with only an outer trim boundary and no interior trim boundaries.

So if you don't want the original result you would want to configure your model to avoid creating surfaces with interior hole trimming boundaries.


> It really seems like this behaviour stems from the midpoint generation behaviour. As if CentroidTriangulation
> isn't fully observed from all angles, or there's some kind "divide larger than" limit somewhere.

It's the "hole breaking" that is giving this type of behavior.


- Michael

Thanks Michael. I really appreciate the response. Well that makes a lot more sense now. I wrongly assumed there was a bit more of a disconnect between how the Mesh generation "sees" the geometry and the underlying MoI modelling process. I guess there are a few MoI processes that would end up with holes in surfaces (like shells, insets, etc.). Though knowing that, it makes sense why the output is typically so great with complex things and now I have a way better way to look at things. Just gotta get out of my engineer mindset that keeps thinking of cutting holes when modelling hard surface things ;)