In recent work we developed a first-principles computational approach capable of addressing questions of large-scale nuclear architecture in higher eukaryotes. Our work emphasised the importance of non-equilibrium or "active" processes arising from chromatin remodelling and transcription-linked activity. A number of results were shown to follow from simulations of this model, including the robust territorial organization of chromosomes and their differential positioning by gene density within the nucleus, both features of a large number of experiments over the past decade or so. In addition to summarising these results, I will describe how (a) other positioning rules i.e radial positioning according to chromosome size and even a combination of size and gene density-dependent positioning emerge naturally from our model, (b) how the two homologues of the X-chromosome (active and inactive) can be positioned differentially and, finally, (c) how comparing such computational approaches to experimental data might make it possible to infer a "positioning code" for chromosomes.