Residual dipolar couplings (RDCs) are NMR observables that probe the orientation of inter-spin (e.g. bond) vectors in molecules that are partially aligned in a magnetic field. Each measured coupling is an average over all orientations of a given conformation of the molecule and all conformations sampled by the molecule: RDCs reflect both the overall shape of the molecule and the local dynamics of the inter-spin vector. This, coupled to the relatively long (up to millisecond) time-scales over which RDCs are measured, means that RDCs provide complementary information to other types of NMR observables commonly used to study molecular structure and dynamics. In addition to their increasing use in protein structure determination, RDCs can be used to bias or validate molecular dynamics simulations. Their interpretation in terms of structure relies upon two major assumptions, however: first, that the molecule is rigid, so that its preferential orientation can be described by an alignment tensor, and second, if internal motion is allowed, that it doesn’t affect the overall alignment, allowing the same alignment tensor to be used for all conformations. While these assumptions are reasonable in the case of well-folded proteins, they are questionable for proteins that are partially or wholly disordered, or undergo significant functional motion –the same proteins for which NMR and simulation are most beneficial. I will discuss alternative ways of interpreting RDCs that do not rely on either assumption, and how they can help to visualise alignment and aid the use of RDCs to validate or bias simulations.