Knowledge management, now a distinct domain of research and practice, has roots in many disciplines. As a result, a wide variety of philosophies, theories, and definitions of knowledge management are used in the literature, and in practice. This has led to many models and methodologies being used in developing knowledge management systems, but without sufficient cross‑pollination of ideas from the various influences and adopted philosophies. We argue that this has led to significant gaps in the understanding of what is needed for knowledge management systems and to divergent and inadequate models and methodologies. These problems are hindering both research and practice. Fieldwork in knowledge management systems development for organisations has been supplemented by an in‑depth analysis of the literature, which has revealed particular gaps in knowledge management systems research. The notions that should underpin knowledge management systems development are confused and incomplete. This paper summarises the most salient of these and challenges several of the published notions of knowledge, knowledge management, and models of knowledge management. In particular we challenge the apparently accepted dichotomies and propose how different facets can be considered within a matrix of KM models.