The notion of ``lifecycle'' is central to the AKT project in which our group is involved. We envisage a high level of automation in our lifecycles, which means that there must at least be some basic, uniform level of formality to support automation. Getting this right is essential not only in providing the technical glue to link different techniques but also to provide a broader, theoretical framework within which we can understand methods for individual lifecycle stages. I shall present some preliminary ideas on a calculus for describing lifecycles and, if time allows, sketch some links between this and the work of the IBROW consortium on reusable architectures for problem solving methods.