Water is an important substance. It is part of us, of our environment, and is a fundamental prerequisite for the existence of life as we know it. The structure of water is still, after over 100 years of research on the subject, however under debate. In this thesis x-ray absorption spectroscopy (XAS) and reverse Monte Carlo (RMC) modeling are used to search for structural solutions of water consistent with many different experimental data sets, with emphasis on the combination of different experimental techniques for a reliable structure determination. Neutron and x-ray diffraction are analyzed in combination with the more recent synchrotron radiation based XAS. Geometrical criteria for H-bonding are implemented in RMC to drive the fits and allow to evaluate differently H-bonded structure models against the data. It is shown that the available diffraction data put little constraints on the type of H-bond topology or O-O-O tetrahedrality for the structure of liquid water. It is also demonstrated that classical MD simulations, using some of the most common interaction potentials for water, give rise to O-O and O-H pair-correlation functions with too sharp first peaks at too short distances to be in agreement with diffraction, and furthermore that requiring a large fraction of broken H-bonds is not in itself enough for a structure model to reproduce the experimental XAS. A contribution to the theoretical description of XAS is made by an in-depth investigation of important technical aspects of the TP-DFT spectrum calculations. A novel approach to RMC, applicable also to data that require a significant amount of computer time to evaluate, is developed which makes use of pre-computed properties from a large set of local geometries allowing RMC simulations directly on data from core-level spectroscopies such as XAS. / At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 4, 5 and 6: Submitted