The prediction and interpretation of structural properties are the starting points for a deep understanding of thermochemistry, kinetics, and spectroscopic signatures of molecular systems. To give an example, detailed knowledge of the conformational behavior of the main building blocks of biomolecules in the gas phase (i.e., without the perturbing effect of the environment) is a mandatory prerequisite toward the understanding of the role played by different interactions in determining the biological activity in terms of structure-activity relationships. The first step to take is an unambiguous definition of molecular structure. We address the so-called Born-Oppenheimer equilibrium structure, which is defined in a rigorous manner and isotopically independent, and the target accuracy. For the latter, we aim at so-called "spectroscopic" accuracy, which implies uncertainties of a few milliangstroms for bond lengths and smaller than a tenth of degree for angles. If on one side the continuous enhancements of the experimental techniques give access to new and unprecedented spectroscopic determinations, on the other side they require increasing efforts for an unbiased interpretation and analysis. Among the pieces of information, accurate molecular structures play a particularly important role. Indeed, there is a strong relationship between the experimental outcome and the electronic structure of the system. Spectroscopic techniques, in particular those exploited in the gas phase, are therefore accurate and reliable sources for structural information. However, it is seldom straightforward to derive molecular structures directly from the experimental information. Indeed, even in the favorable case of investigations in the gas phase, vibrational effects are always present, and disentangling their contributions requires collection of information for all vibrational modes, a nearly impossible task. To overcome these limitations, joint theory-spectroscopy strategies can be identified, which are referred to as "top-down" and "bottom-up". The first approach, denoted as the semiexperimental approach, relies on extracting from experimental outcomes the equilibrium structure by using quantum-chemical computations to recover vibrational effects. The bottom-up approach consists in verifying the computed equilibrium geometry by means of a comparison between calculated and experimental spectroscopic parameters that probe structural characteristics. In this contribution, we try to review the most important challenges in accurate molecular structure determinations, with particular emphasis on the "solution" provided by a joint theoretical-experimental approach and on the current state of the art. Starting from the illustration of different strategies, we proceed by addressing the increasing complexity in the derivation of equilibrium geometries: we start from the construction of a database of accurate structures, we then face the problem of extending the dimension of the systems amenable to accurate structural determinations, and finally...