Comment 2 for bug 536542

The main question is whether to implement node regions or triangle regions. It depends on which resolution is wanted. Triangle regions would have twice the resolution of node regions, and require twice as many region IDs stored in memory. Terrain is a property of triangles, so for triangle regions it is possible to say for example that a particular region has 12 wiese1, 8 steppe_kahl, 3 bergwiese and 1 berg1 (or let the editor tool select an area based on terrain and create a triangle region from it). It also depends on what the regions should be used for. Vision, Military Influence and Ownership are properties of nodes, so if regions should interact with these, they should be for nodes. But what if we find out in the future that ownership should be a triangle property in the future? The issue whether ownership should be on nodes or triangles has not been investigated thoroughly.

(But it would actually be possible to implement both independently.) I am currently leaning towards node regions. Then I would make some events and triggers capable of operating on regions, as an alternative to the point/radius that they support now. The events in question are conquer_area, move_view (to the region's ID-point) and unhide_area. The triggers in question are building, military_influence, ownership and vision.