Research and setting candidates

Intake covered three areas from public design write-ups and technical articles; the full list is in sources. Reliability for each source is tagged there. Nothing below is a decision.

Geometry approaches (candidates)

• Per-cell solid/open with a smooth union
• Stateless hashed 3D feature points (the keyed sites used here)
• Exact power-diagram cells (a cell-by-cell library such as Voro++)
• Surface extraction: marching cubes with Transvoxel seams, surface nets, dual contouring
• Signed-distance-field edit list (the Dreams approach, read only via excerpts)
• Strict-grid voxels (Teardown) as a cost contrast

Base-building lessons

• Connectivity-graph support (Medieval Engineers, Valheim friction)
• Plain, forgiving rules (Notch: gravity is not realistic because it would wreck player work)
• Terrain as a connectivity graph (NOWHERE’s tetrahedra argument)
• Terrain deformation as a reward (Astroneer)
• Systemic threats that come from simulation, not scripts (RimWorld, Dwarf Fortress, Breath of the Wild)

Setting candidates (not survival)

Name	One line
Reef Wright	Grow and shape a living reef-like structure
Thaw the Geode	A Terra Nil-style restoration, underground
Tide of Foam	Build against a rising, spreading medium (Creeper World lesson)
Mineral Heresy	A faith that treats stone as scripture
Memory of the Mass	The ground remembers what was built in it
Symbiont Commune	Co-build with a colony of organisms

A setting is judged by whether mechanics deliver it (the MDA check), not by theme text. Terra Nil criticism (rules that contradict the fantasy cost the player friction) is the main warning.