A major obstacle to delivering the increasingly complex software systems that society demands is the resource drain from maintaining existing systems. The high expense of maintenance is related to the tendency of software quality to decline over time. Maintenance is often performed under tight resource constraints, with the minimal amount of effort required. Typically, there is a gap between this minimal amount of work and the amount required to maintain the software's quality. This gap can be viewed as a type of debt, which brings a short-term benefit (usually shorter release time) but which might have to be paid back, with ?interest? (decreased productivity), later. Many practitioners find this metaphor intuitively appealing and it is already transforming the way that long-term software maintenance is viewed. But its lack of a sound theoretical basis, empirically-based models, and practical implementation hinder its ability to transform how maintenance is done. Thus the contribution of this work is to provide empirically based models describing, and validated mechanisms for managing, technical debt. This project also supports the PI's activities in mentoring a diverse population of students, as well as UMBC's nation-wide prominence in the advancement of women and minorities in science and technology.