SPROUT—A Modeling Language for SEED

Abstract

An important capability of the SEED system is rapid generation of internal design representations, including design alternatives, that can be evaluated in terms of a broad range of criteria. It consists of modules that must communicate with each other and remain extensible. From a software engineering perspective, there exists no single programming language or database management system, commercial or research-based, that provides all the needed support. In this paper, we describe the SPROUT modeling language, the cornerstone of efforts to bridge this gap. It supports the specification of a system information model, which contains a building product model as a subset. Developments are discussed that lead to the current language design as well as the support for case-based design, persistent storage of objects, management of versions and alternatives, and communication between different SEED modules (i.e., agents). Emphasis is placed on the automatic generation of computer programs and database schemas from a SPROUT specification. Lastly, this effort is compared to other comparable efforts including PDES/STEP.