With the exception of quantum key distribution, experiments in quantum information science are far behind theoretical work. It is important to try to bridge this gap. I will begin this talk by briefly describing two fruitful collaborations between theorists and experimentalists. First, together with the Optics group at Glasgow University, we have recently been able to obtain violations of Bell inequalities in 11 x 11 dimensions using the orbital angular momentum of light. We also verified, through analysis of the experimental data, that it could not be explained by a state which is entangled in less than 11 x 11 dimensions. This demonstrates that light with orbital angular momentum can provide high-dimensional entangled states for quantum information. Using high-dimensional entanglement, it may also be feasible to close the detection loophole when testing Bell inequalities. The second topic is a realisation of a quantum walk using a compact and scalable fibre loop setup, in collaboration with the University of Erlangen. Quantum walks have been shown to speed up search problems and to play a role as a computational primitive. The fibre loop setup, due to its flexibility, provides a rich playground for investigations of the physics of quantum walks. Finally, I will describe some theoretical schemes to prepare entangled or nonclassical states of nanocantilevers e.g. using interactions mediated by a Bose-Einstein condensate. We have also investigated under what conditions decoherence may be helpful in driving a system towards a non-classical state. It turns out that it is crucial whether or not what is commonly referred to as the rotating-wave approximation can be made or not.