This Small Business Innovation Research Phase I project will explore the feasibility of building a practical tool for software maintenance based on theory. Maintenance presently consumes most of the lifecycle costs of most software, yet current methods used for this work are at best ad hoc. Building on a formal theory of Transformational Maintenance by Design Reuse 0, Semantic Designs (SD) intends to demonstrate, by construction and validation, the utility of a semantics-based Design Maintenance System (MS) in installing incremental changes to software for functionality, performance and portability. DMS would enable a nearly continuous software modification process, unifying both software design and software maintenance activities. While the theory requres software generated transformationally from a specification, DMS would be augmented with (transformational) design recovery mechanisms to enable its application to existing code. During the SBIR Phase I research effort, SD will establish the feasibiity of a practical DMS by designing `nonlinear` transformation dependency networks, choosing suitable internal representations for code schemas and translation methods for converting code transformations to/from those schemas, defining an initial ar chiteture for the DMS, and evaluating it by manual simulation on sample real codes.
