Ok, so I've been trying to figure this out for a while with little success. Edge groups are everywhere in various different SOP nodes, and yet I have no clue really what they are? Are they point groups under the hood? I've seen that explanation offered, but that doesn't make much sense as:
1. They don't show up anywhere in the geometry spreadsheet as point groups would.
2. Theoretically I can have many different edges between two points (different primitives), so it would need to store more information.

Are there any staff folks around who could give a quick explanation of how edge groups are represented under the hood?



(My motivation here is that I'm debugging some of the AutoSeam/AutoUV GameDev tools)

Edge groups are defined by point pairs
You can use expandedgegroup() to get array of point pairs from existing edge group
Or setedgegroup() to add/remove point pair to a group, plus other edge functions

Since edges are technically not an element of a geometry they will not show up in spreadsheet don't have edgenums and can't hold attributes, therefore the point pair representation

Each point pair defines exactly one edge
what you mentioned in 2. houdini calls half-edges and you can also query them using vex

Thanks for the help, very much appreciated. The distinction of half-edges makes sense.

This might be a silly question, but do you know why they don't show up in the geometry inspector? It seems that edges are a just as much of a fundamental structure as point, vertices, primitives, etc.

Edges are a by-product of connecting points into prims via verticies. Hand wavey explaination here:

http://www.tokeru.com/cgwiki/index.php?title=Points_and_Verts_and_Prims#Edges_and_prims [www.tokeru.com]