Zalusithix
Member
Going through this month's Jump Point, I found this bit on the zones interesting..
JP: (From Oberscht) What would you say are the biggest features or quirks the engine is going to have that will be unique among engines once done?
Christopher: The supported scale, and most likely how we store positions.
Regular engines have ObjectsInWorldSpace -> Renderer -> Transformed into ScreenSpace on GPU.
We have: Positions in ZonePartitionSpace, which is sort of a regular grid inside each zone (to allow using 32-bit SIMD even at universe scale). Those are transformed into ZoneSpace, which is dependent on all the parent zones of this zone, which is then converted into camera space (to allow the renderer to use 32-bit operations), which then goes into the GPU for the final ScreenSpace transformation.