The objective is to derive, from experimental data, a physically sound model for the spatial distribution of sky luminance. If realized, this model will predict, for all insolation conditions, the spatial variation of the sky luminance using only solar position, direct and global irradiance as input.