Cooking spline components/actors into our own system, looking for tips

Hey there, I’ve been implementing a traversal system and was thinking about how to handle splines and their representation in the world. Correct me if I’m wrong, but it isn’t possible to query a collision directly against a spline, so I have to use a “parent” physical object that will let me look for a spline on the hit actor.

However, I would like to decouple the connections between the traversable object and the ledge itself. I’ve been thinking about a possible solution in which you would use the editor to place special splines into the level, but then export them into a different representation. Would it be possible to use splines only for the cooking process, store this information in a standalone system, and then get rid of these splines (or actors) altogether?

The system itself would handle the storage and capsule-tracing against these splines. Most likely, storing only the ledge coordinates (with some simple condition for traversal) in a grid that would be loaded based on the world partition.

I’m still in a brainstorming stage and thinking about how to handle this kind of interaction with splines. What do you think about this approach? Would it have any performance/memory advantage over splines with their box colliders living in the world? Is it even worth it, or would there be a better way to handle searching for static ledges?

Hello!

Can you share more details on the type of traversal you are thinking about? Could the navigation system (nav meshes) be used to fulfill the need?

Regards,

Martin