The simulation of daylighting performance requires some knowledge about prevailing sky conditions, especially luminance. Commonly the CIE standard models (clear and overcast, and often some combination of both) are used to provide the necessary estimates to obtain the temporal and spatial variance of luminances and the consequent illuminance levels in buildings. We know from data analysis of measurements recorded at the Research Class measurement station located at the Sydney International Airport, Mascot, and from other data from the International Daylight Measurement Program (IDMP) that CIE models do not provide accurate indications of the sky conditions in some countries. Nor do combinations of the standard models provide a realistic estimate of intermediate sky conditions. An attempt is now Boeing made using recorded data to come to terms with the luminance distributions of actual sky conditions both in Australia and overseas. This paper presents some proposals which are aimed at defining a modelling strategy which will facilitate a much more realistic approach to the estimation of sky luminance. In particular we suggest that it is possible to define a digital form which enables the standard empirical models, as well as data from sky scanning devices, to be integrated into a common modelling framework. In addition it will be proposed that this standard form will provide a basis for the construction of model skies for particular locations given some local knowledge of the atmospheric conditions. These sky models are intended to be used directly by simulation packages or other software systems requiring information on the distribution of luminance across the sky dome.