Fundamental Principles of Good System Design

These design principles come from the experience of hundreds of engineers and managers. The particular references cited in the following paragraphs are not meant to be the authority—they are merely examples that have references in the literature. Use models to design systems: System design can be requirements based, function based, or model based. Modelbased system engineering and design has an advantage of executable models that improve efficiency and rigor. One of the earliest developments of this technique was Wymore’s (1993) book entitled Model-based System Engineering, although the phrase “model-based system design” was in the title and topics of Rosenblit’s (1985) PhD dissertation. Model-based systems engineering depends on having and using well-structured models that are appropriate for the given problem domain (Bahill and Szidarovszky, 2008). Bahill’s models start with the use cases. Use hierarchical, top-down design: Early on, translate the customer’s needs into goals, capabilities, and functions; these provide guidance for all future development. Work on highlevel functions first because, although high-level functions are less likely to change, when they do change, they force changes in many other functions. Decompose systems into subsystems, subsystems into sub-subsystems, etc. (Chapman and Bahill, 1992). In software, this decomposition is called layered architecture (Evans, 2004). Implementation is simpler if the dependencies and action initiations between these layers are unidirectional. Work on high-risk items first: Work on high-risk items first in order to reduce risk; in addition, high-risk items are more likely to change, thereby producing changes in other entities, so working the high-risk items first will reduce the rework due to changing requirements. Furthermore, if it was impossible to satisfy the high-risk capabilities and the project was cancelled, you will have saved the money that otherwise would have been squandered satisfying low-risk requirements (Jacobson, Booch, and Rumbaugh, 1999). The original spiral model of Boehm (1988) advocated risk-driven development.