Software, Process Control, Real-Time System, Software Engineering, Application Development Software, Programming, Prototype, Requirements Analysis, and System Development
Large-scale software requires effective management for production. Such large-scale software consists of application software, a utility subsystem, and an operating system. Individual software factories require levels of abstraction in a design process which uses prototyping, reusing, and program generating systems. The first level is the requirements level which defines the external devices with which the software communicates. A capsulated form of a requirements description is shown. The data-function or design level is the transition, the definition of a user's needs and the establishment of the model. Program models are defined and implemented in the program level. Prototyping is done throughout the entire process for the first operational versions of software interfaces. Productivity and reliability are the most crucial factors in management of a software factory. In addition to the encapsulated format examples, numerous block diagrams illustrate software production and the rolling mill software production example.
Communications of the ACM. Feb 1990, Vol. 33 Issue 2, p225, 8 p. table Fault detection process.
Requirements Analysis, New Technique, Technology, Software Metrics, Software Engineering, Software engineering -- Planning, and Software
N-fold inspection, an new technique for software requirements analysis, is discussed and compared with other development methodologies. An investigation into the User Requirements Document (URD) is the traditional first step in the software life cycle. The accuracy of the URD is crucial to its success or failure, because detecting and fixing a software problem during the requirements and early design phases is far less costly than doing so after delivery. Formal specification, design and implementation, prototyping, and formal inspection of the URD are common requirements-analysis methods. N-fold inspection replicates formal inspections using N independent teams. The results of a pilot study observing 10 independent software development teams are presented. It is concluded that N-fold inspection can provide reliability at the URD level and can also be useful in other development phases.