Among the various issues that must be addressed in the building of software systems, there are two issues that fall within the province of version and configuration management: component derivation and component consistency—that is, recording or determining the derivation of one component from another, and determining whether the components within a particular configuration are consistent with each other.

Much of the past and current work in version and configuration management has addressed the problem of keeping track of how components are derived. We have systems that manage version and configuration histories—for example SCCS [8], RCS [10], NSE [1], etc—by effectively keeping either a tree or a graph representing the derivation history of source versions. We have various tools that provide automated derivation of secondary objectsl—for example, various forms of Make [3], etc, or such opportunistic processors as Marvel [5]—to help us build executable versions of our systems. In general, we have a fairly deep understanding of the issues in managing the derivation relationship both for manually derived components as well as automatically y derived components. For example, see Borrison [2] for a-discussion of these latter issues.