Edge modelling codes are currently used to understand and interpret the results ofpresent day tokamaks, and for predictive modelling in support of engineering design of existing, as well as future, tokamaks, e.g. ITER. It is therefore crucial to compare the existing codes, and to understand, document and resolve the differences between them, whether these arise from differences in underlying physics approximations or choices of numerical treatment. A programme for comparing edge modelling codes is underway, starting with D without drifts, D with drifts, D+C without drifts, and, finally, D+C with drifts. The codes currently being compared are SOLPS ("B2.5-Eirene"), EDGE2D-NIMBUS and EDGE2D-EIRENE, all with Monte-Carlo neutrals; and SOLPS ("B2.5") and UEDGE, with fluid neutrals. In order to remove one possible source of differences in the codes, all of the codes are using the same grid (the standard grid generated by the JET GRID2D or, for some test cases, an orthogonal grid generated by UEDGE). Significant effort has also gone into ensuring that the boundary conditions and transport coefficients are equivalent for all of the codes.