The parameter is more accurately the L-infinity norm of the state-space, which then rescales the representation accordingly. The name radius would suggest that (1) the space is a sphere (no, it is a cube), and (2) the decoded state variable x is unchaged (no, it is rescaled).
On second thought, even though this name is misleading, it is a useful / intuitive descriptor that connects with known concepts in nengo (albeit imperfectly). Documentation should be enough.
The parameter is more accurately the L-infinity norm of the state-space, which then rescales the representation accordingly. The name radius would suggest that (1) the space is a sphere (no, it is a cube), and (2) the decoded state variable x is unchaged (no, it is rescaled).
Possible new names:
inf_norm
linf