Looking at existing engine options at the moment for building MMO strategy game - the design is along the lines of the strategy element of Rome Total War(RTW) but with real time mechanics. A distant camera view of a large game world with relatively little action. However I was expecting a desirable extension for tactical battles to be resolved in a 3D environment (again such as in RTW) to have to be placed in a distant (maybe never) release due to the obvious performance difficulties.

Some conceptual questions then.

1)The initial requirement are for a game world maybe spanning 10,000 km squared represented by a heightmap, the game view port needs to cover around 50 km minimum and I want to be able to present this either as a 2.5D style shaded-relief style map (or similar) or preferably as a 3D rendered terrain with some ability to zoom in to a smaller area (though with a camera still fairly remote) and without the need for a huge amount of detail.

a) Can unity accommodate 2d backgrounds, sprites etc? b) could a region of the world heightmap be loaded into Unity procedurally and rendered (on the cheap) with minimal delay to the player after say he has selected an area on a small "jump map" of the world to load into the viewer?

2) On reading about the potential of Unity I am beginning to wonder if it might actually be possible on a realisable release schedule to design the game with tactical battles in 3D an expectation all be it if not on the initial release, rather than with a simple calculated outcome as will otherwise be required. I guess ideally the scenario in question 1b (if it is achievable) would be the start point from which to zoom in onto a battle perhaps covering a an area of 10 km square, though an artificial jump to a new scale migh be fine.

a) Would this sort of terrain size be achievable particularly if constructed on a procedural basis and with rendering on the fly (using data about relative heights from the heightmap and randomising terrain artifacts such as trees, ponds, cliffs etc)? b) Its difficult to talk about numbers of concurrent battles at this point but if resolving interactive 3D battles is practical I am wondering about what limits Unity today/tomorrow might put upon this ideal. This is both from the perspective of terrain rendering as mentioned and also from the perspective of scene management in which perhaps each battle might have up to 4000 3D models running amuck?