GSoC 2021: Porting Popular Modifiers To Geometry Nodes: Feedback

Hello everyone! This is the feedback thread for my GSOC project: “Porting Popular Modifiers To Geometry Nodes.”

More coming soon.

7 Likes

Hey coming from your latest weekly report! I’ve seen solidify working differently as expected and wanted to know if this was intended, why does “rim” hide the actual mesh?

afaik only rim was to fake thickness without the extra polygons, or at least i use it for that reason

I doubt that this has anything to do with this GSOC, if anything it seems to be regression from implementing the simple/complex modes for solidify modifier. (Although I didn’t bother checking how it worked before the two modes were split) Complex mode gives proper result of “Only rim” isolating the rim loops while simple gives odd result of removing one of the shells.
T89050 notes that support for simple mode for the GEO node implementation will be dropped so this should solve itself.
Personally, on the notion of nomenclature for T89050, I think “Fill” should be replaced with “Shell”. Solidify is essentially an operation that creates two shells from a single surface and “fills” an “rims” between them. Referring to the shells as “Fill” doesn’t make much sense, albeit it is already much better than Rim Fill was…

One fairly large design incoherence/point of confusion with Generative Modifiers has been that some have Vertex Group selection that affects created geometry (e.g. bevel, decimate…) while some only control the influence while creating duplicate geometry in 0.00 influence zones (solidify, wireframe)
This might not really be an issue anymore in geometry nodes since you have more control over which geometry you feed into the solidify (compared to only option before being splitting into different objects) but I figured I’d point it out.

2 Likes

Saw the voxel remesh node prepared commit. This type needs (the modifier needs it too) a soft limit when sliding the value. This prevents crashing or letting blender load for ages.

2 Likes

Will the weld modifier work nice when physics nodes come into play ?

Like if you weld 2 meshes together to create 1 big one, unlike with the current weld modifier that does not play well with physics