Knowledge that Transforms

This article describes the use of information technology to leverage knowledge—knowledge in the form of pioneering research and development work in the electric utility industry, managed by the Electric Power Research Institute (EPRI). EPRI is the R&D consortium for the U.S. electric utility industry. In 1984, EPRI embarked on building EPRINET, a state-of-the-art electronic communication and delivery system. This effort coincided with EPRI expanding its mission to encompass not only management of R&D projects but also delivery of its information and expertise. This article describes the lessons that EPRI has learned about managing such a large systems effort while the organization is in the midst of making cultural changes.

Management of the software development process requires a thorough understanding of the environment in which development takes place. Ability to estimate, plan, and manage resource consumption is limited by the central problem of determining the size of system specifications. To address this issue, a general strategy for measurement and evaluation of system development environments needs to be established. This article presents a research model that will help managers and researchers understand and establish the linkages between units of systems requirements specification, design, and source code. Initial validation of the model was performed by reverse engineering systems written in a fourth generation language from source code to design metrics. Results indicate that the model may provide reliable measures of system size in terms of both design metrics and lines of code.

This paper describes an approach to application software development (the Application Software Factory) that enables over 90 percent reuse of code, produces application code where quality is measured in defects per million lines of code, and generates productivity exceeding that of interpretative 4GL environments. The environment is built on top of a commercial CASE tool and does not rely on exotic technology. The delivered systems are COBOL transaction processing systems using relational database technology and are predominantly online. Two applications of approximately 200,000 lines of code have each been developed by teams at two different sites. A third application of over 2 million lines of code is nearing completion. The paper depicts the important relationships between technology, management, methods, and design approaches that comprise the Application Software Factory.

Growing competition in the investment banking industry has given rise to increasing demand for high functionality software applications that can be developed in a short period of time. Yet delivering such applications creates a bottleneck in software development activities. This dilemma can be addressed when firms shift to development methods that emphasize software reusability. This article examines the productivity implications of object and repository-based integrated computer-aided software engineering (ICASE) software development in the context of a major investment bank’s information systems strategy. The strategy emphasizes software reusability. Our empirical results, based on data from 20 projects that delivered software for the bank’s New Trades Processing Architecture (NTPA), indicate an order of magnitude gain in software development productivity and the importance of reuse as a driver in realizing this result. In addition, results are presented on the extent of the learning that occurred over a two-year period after ICASE was introduced, and on the influence of the link between application characteristics and the ICASE tool set in achieving development performance. This work demonstrates the viability of the firm’s IS strategy and offers new ideas for code reuse and software development productivity measurement that can be applied in development environments that emphasize reuse.