A disordered quasi-liquid layer of water is thought to cover the ice surface, but many issues, such as its onset temperature, its thickness, or its actual relation to bulk liquid water, have been a matter of unsettled controversy for more than a century. In this perspective article, current computer simulations and experimental results are discussed under the light of a suitable theoretical framework. It is found that using a combination of wetting physics, the theory of intermolecular forces, statistical mechanics, and out-of-equilibrium physics, a large number of conflicting results can be reconciled and collected into a consistent description of the ice surface. This helps understand the crucial role of surface properties in a range of important applications, from the enigmatic structure of snow crystals to the slipperiness of ice.