I review the basics of the disc instability model (DIM) for dwarf novae and soft-X-ray transients and its most recent developments, as well as the current limitations of the model, focusing on the dwarf nova case. Although the DIM uses the Shakura-Sunyaev prescription for angular momentum transport, which we know now to be at best inaccurate, it is surprisingly efficient in reproducing the outbursts of dwarf novae and soft X-ray transients, provided that some ingredients, such as irradiation of the accretion disc and of the donor star, mass transfer variations, truncation of the inner disc, etc., are added to the basic model. As recently realized, taking into account the existence of winds and outflows and of the torque they exert on the accretion disc may significantly impact the model. I also discuss the origin of the superoutbursts that are probably due to a combination of variations of the mass transfer rate and of a tidal instability. I finally mention a number of unsolved problems and caveats, among which the most embarrassing one is the modelling of the low state. Despite significant progresses in the past few years both on our understanding of angular momentum transport, the DIM is still needed for understanding transient systems.