Planets will be placed in their correct positions relative to the Sun as calculated from the time of day but will only proceed in their orbits during the end-game sequence. Planets will revolve on their axes at a rate of one revolution per minute. The interplanetary units will be millions of kilometers, i.e. Pluto will be up to 7,360 units from Sol. The units of planetary scale will be some multiple of kilometers (currently hundreds of thousands, rounded up to one).
The path of the arrows will be affected by gravity and density of source planet (indicated by speed and arc taken). Planet positions will be projected onto the 2D plane but their distances will be taken into account for the arc calculations. Must aim a bit high...
Currently implemented as cylindrical meshes from planet A to planet B. There is an undetermined error in the length calculation. New idea of using gradient shaded triangles suggested by Bob based on idea from Mark Subbarao to create a motion blur effect.
The player character will be displayed with a third-person view when on a planetary surface; at all other times, a first-person view will be used. The character wears a bulky device (inferred to be the beacon) and carries an object that looks like a bow. No quiver; the arrows are composed of energy drawn from the planet.
Planetary models have been created in Maya using surface textures acquired from various sources (e.g. NASA, Celestia, Adler Planetarium) on spherical primitives which have been manually distorted to create mountains and valleys. This creates very high polygon counts, so more distant planets will be approximated with spheres.
Ambient music will be played according to your location within the game; specific sounds will be played when walking, moving along the links, firing arrows, occupying core, jumping, etc.
Core: fixed position, free look.
Surface: chase character from behind, need to be able to lean back / look up / aim arrow
Orbit: fixed distance, move over surface, free look
Link: face along link, fixed perspective
The enormous world scale has proven difficult to implement. If the scale is too large or too small, stereo becomes complicated; if too large, Z-buffering fails, in both cases, the camera is hard to manage.