Recent progress in nanoscale manufacturing allowed to experimentally investigate quantum dots coupled to two superconducting leads in controlled and tunable setups. The equilibrium Josephson current was measured in on-chip SQUID devices and subgap states were investigated using weakly coupled metallic leads for spectroscopy. This put back two "classic" problems also on the agenda of theoretical condensed matter physics: the Josephson effect and quantum spins in superconductors. The relevance of the former is obvious as the barrier separating the two superconductors in a standard Josephson junction is merely replaced by the quantum dot with well separated energy levels. For odd filling of the dot it acts as a quantum mechanical spin-1/2 and the relevance of the latter becomes apparent as well. For normal conducting leads and at odd dot filling the Kondo effect strongly modifies the transport properties as can, e.g., be studied within the Anderson model. One can expect that also for superconducting leads and in certain parameter regimes remnants of Kondo physics, i.e. strong electronic correlations, will affect the Josephson current.In this topical review we discuss the status of the theoretical understanding of the Anderson-Josephson quantum dot in equilibrium mainly focusing on the Josephson current. We introduce a minimal model consisting of a dot which can only host one spin-up and one spin-down electron repelling each other by a local Coulomb interaction. The dot is tunnel-coupled to two superconducting leads described by the BCS Hamiltonian. This model was investigated using a variety of methods, some capturing aspects of Kondo physics others failing in this respect. We briefly review this. The model shows a first order level-crossing quantum phase transition when varying any parameter provided the others are within appropriate ranges. At vanishing temperature it leads to a jump of the Josephson current. When being interested in the qualitative behavior of the phase diagram or the Josephson current several of the methods can be used. However, for a quantitative description elaborate quantum many-body methods must be employed.We show that a quantitative agreement between accurate results obtained for the simple model and measurements of the current can be reached. This confirms that the experiments reveal the finite temperature signatures of the zero temperature transition.In addition, we consider two examples of more complex dot geometries which might be experimentally realized in the near future. The first is characterized by the interplay of the above level-crossing physics and the Fano effect, the second by the interplay of superconductivity and almost degenerate singlet and triplet two-body states.