The broadest definition of dendritic cells (DCs) is white blood cells that can take up antigen, process it and then present antigen-derived peptides to activate cognate naive T cells. Although this definition is by no means perfect, it is nevertheless now textbook. The source of frustration more recently has focused on other issues, including the distinction of the DC subtypes, their differential roles in the immune system, their lineage relationship to each other (and other leukocytes) and whether the mouse and human DC findings overlap. Here, I condense the classification of DCs in both the steady state versus infection, with primary focus in the mouse. Emphasis is then given to debates surrounding the in vivo pathways of DC differentiation in different conditions, which culture models best represent these processes (fms-like tyrosine kinase 3 ligand versus granulocyte-macrophage colony-stimulating factor), and what the human and mouse DC subtype equivalents might be. In addition, a model termed 'graded' commitment is proposed that, as a departure from the classic binary models of hematopoiesis, attempts to explain the recent clonal data where subtype-specific DC precursors branch from this pathway.