The documentation mentioned two types of outputs: Mesh and Parametric results. What are the implications and use cases of them? Guess Mesh is much lighter and carries textures, but can't be edited? Parametric models are heavier, but can be pushed and pulled?

The documentation mentioned that "Alternatively, your app can display custom graphics during the scanning process by creating and using a scan session object (RoomCaptureSession) directly". What would be the custom graphics? Any examples would be highly appreciated.

What are the criteria for RoomPlan to distinguish doors, windows, and openings? In the end, they are all openings, right? In addition, what is the probability of categorization errors? Thanks a lot.

Does RoomPlan recognize curved edges and surfaces? So far the videos and images showed only blocks and straight lines:)

Dear all, the video wwdc2022-10127 recommend scanning time under 5 minutes. Would it be possible to resume an unfinished scan across multiple sessions? 5 minutes may not be enough to capture a big space.

Just curious, would RoomPlan be able to tell the floor and its direction? The goal is to drop other objects onto the floor and make them stand correctly. Thank you!

Let's say a 10-feet wall. Would RoomPlan return 11, 10.1, or 10.01 feet? Thanks a lot.