Experience realising a meta‐model for wide system understanding: The global system model

Abstract

AbstractThe Esprit DOCKET project seeks to reduce the problems associated with software maintenance through enhancing traditional, reverse engineering tools with other sources of knowledge, enabling mainters to produce more accurate system models. To produce accurate system models, knowledge about an information system, obtained from different sources, must be interpreted and integrated into a single model (known as the global system model in DOCKET). The definition and implementation of such a model involves the resolution of many issues. These include multiple views of a system at different levels of abstraction, handling inconsistency, incompleteness and system description evolution. This paper reports on the implementation of the Global System Model using the language SML. SML enabled design and implementation problems to be surmounted at a high level of abstraction and facilitated the rapid implementation and testing of the global system model, greatly reducing the effort required to implement it compared to more conventional facilities.