With the atomic clocks and VLBI catalogues of extragalactic sources, the theory-independent references for both space and time will be quasi ideal in their underlying principles (non-rotation of very distant objects, constancy of the frequencies associated to atomic transitions). The paper is an attempt to clarify the role of these natural references with respect to those which are implicitly defined by the dynamical theories.
The example of time is discussed, since it already gave rise to serious difficulties, but similar problems may occur about the space references. Such complexities are suprising: one could have expected simplifications from the use of references which are not entangled with theory. Some suggestions are made for the use of these natural references.