This proposal is for work to formally specify and then implement a modular, extensible backend engineering database supporting data translation between an open-ended set of applications. The unique features of the data model and its prototype implementation are: (1) dynamic extensibility, allowing new applications to be added throughout the process of design, and later throughout the product life cycle, with expectations that new applications interfaces can be implemented by users, not database administrators, (2) support for the iterative nature of design, in the form of explicit monitoring and management of partial integrity, (3) a richer data model tailored to the needs of engineering design, reducing the impedance mismatch between the current functionality of advanced CAD systems and current databases. (4) a new model of translation between applications, that does not rely on a neutral file approach, but rather treats translation as a set of transformations over a type network. The concepts supporting these capabilities will be developed and formalized. The prototype implementation will be general, supporting a variety of engineering design areas. Applications to be tested on top of the database will be drawn from civil and mechanical engineering design research literature. The prototype implementation will be developed to be easily portable, and made available to other research groups.
