The results of a workshop on the structure and chemistry of CuInSe2, (CIS), held in Jackson Hole, WY (USA), are presented. The participants are the authors of this paper. Issues in CIS properties and device performance were divided into those dominating common CIS and those relevant to the best CIS. Common CIS is non-equilibrium material containing numerous lattice defects and second phases surrounding the grains or segregated to the surface. Common CIS is capable of producing solar cells yielding in excess of 10% conversion efficiency provided that there is no CuxSe present. The best physical vapor deposited (PVD) CIS is formed at high temperatures, where rapid segregation should minimize the incorporation of dissolved point defects. This CIS has relatively large, strongly (112)-oriented, nearly perfect grains. Growth of the majority of the film under strongly Cu-rich conditions provides the best results, although the final film must be free of CuxSe phases. The differences between common CIS and the best CIS are relatively process-independent (but process-variable-dependent) and tend to limit the open-circuit voltage and fill factor of devices. The workshop consensus recommendations were as follows. The selenization reaction must be studied to determine how desirable properties, reproducibly demonstrated in PVD CIS, can be reliably achieved. Controlled experiments should be carried out to determine the effect of extrinsic doping and grain boundaries on minority carrier lifetimes. Finally, a mechanism is needed for analysis of films to show how materials produced by different methods compare.