Old proposal about position:

To be able to control the position in 3D either in cartesian coordinates or in spherical coordinates we need two different messages (/xyz (x-axis y-axis z-axis) and /aed (azimut elevation distance) to
define a position. I would propose normalized coordiante system like it is already implemented in the ICTS Ambisonic modules (see e.g jmod.sur.speaker.setup.help), but maybe this is also an issue for the unit lib if you want to work in meter like "Moore's Space Unit Generator" [1]? is doing.

cartesian:   /SpatDIF?/Src/3/xyz -0.5 0.5 0.0   for Position of Sound Source Nr.3 in cartesian coordinates              
spherical:   /SpatDIF?/Src/3/aed 23.0 0.0 1.0 for Position of Sound Source Nr.3 in spherical coordinates

[1]?  http://crca.ucsd.edu/~syadegar/space.html

Trond comments:

Ideally we should be able to describe position as a namespace class independently of its use in spatialisation. That way it could also be used for e.g. OpenGL?.

Relative values don't work well with doppler, air filter or reverb that depend on distance

Trond recommended to have an active and passive flag with every message. Usually you only send a new position message only when a sound is supposed to be heared. That's why I think to send an active message wouldn't provide new informations. So that's why I think the active message should either be send separately by

/SpatDIF?/Src/2/active 1 tells that source 2 is active

or this could be another optional arguement

/SpatDIF?/Src/3/aed 23.0 0.0 1.0 active 1 tells that position of source 3 and that it is active

Comment: Put on hold for the time being, not sure if it is needed.